article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the life times of giant molecular clouds ( gmcs ) have been at the center of a major debate for at least the last 40 years @xcite .
gmcs that live for many free - fall times need a mechanism that prevents them from gravitational collapse . over the last couple of years
the consensus is growing that the life times of gmcs are likely a few free - fall times , or even less @xcite and the focus has shifted towards the challenge of explaining the low star formation efficiencies in gmcs .
the star formation efficiency per free fall time @xmath1 is defined as the ratio of free fall time @xmath2 to gas depletion time @xmath3 .
in other words : @xmath4 i.e. the instantaneous star formation rate ( sfr ) is proportional to the available amount of molecular hydrogen ( @xmath0 ) via the proportionality factor @xmath5 .
the observed value @xmath6 ( e.g. @xcite ) means that only 1 - 2% of the mass of a gmc is converted into stars over a free - fall time .
if star formation is supported by supersonic turbulence @xcite @xmath1 is expected to be only very weakly dependent on the mach number of the turbulent flow and thus approximatively constant , but this may be an oversimplification , see e.g. @xcite . on the other hand , if gmcs have life times of the order of a free fall time they do not need be supported by turbulence .
the star formation efficiency in such clouds may increase as the mach number in the flow decreases and the cloud collapses , see e.g. @xcite .
we note that a time varying @xmath1 should introduce additional scatter in the relation between star formation rate and @xmath0 surface density on small ( @xmath7 pc ) scales .
this scatter should propagate up to @xmath8kpc scales , see e.g. @xcite , and hence would contribute to the scatter in the kennicutt - schmidt relation .
this , at least in principle , could be used to test observationally the time dependence of @xmath1 . in section [
sect : pitfalls ] we will discuss two common misconceptions that could give rise to the impression that @xmath1 varies over the life times of gmcs even if it is a constant .
in section [ sect : model ] we present and analyze a toy model in order to exemplify and quantify our statements .
in equation ( [ eq : dotmstar ] ) we define the _ star formation efficiency per free fall time _ @xmath1 .
another commonly used efficiency is the _ star formation efficiency of the gmc _
@xmath9 , i.e. the fraction of @xmath0 mass of the cloud that is converted into stars over the life time of the cloud . in a picture where gmcs start with an initial reservoir of @xmath0 , which is used in the subsequent star formation process
, the final stellar mass @xmath10 is divided by the _ initial _
@xmath0 mass of the cloud .
if the cloud accretes a substantial amount of @xmath0 over its life time , the definition has to be generalized .
we will use : @xmath11 where @xmath12 is the _ maximal _ @xmath0 mass of the gmc . by definition @xmath9
is a non evolving quantity and it can be estimated , e.g. , by comparing luminosity distribution of ob associations in the milky way with the mass spectrum of molecular clouds @xcite .
it _ can not _ be directly measured on a cloud - to - cloud basis , because @xmath13 and @xmath14 must be known at two different times .
instead such observations , see e.g. @xcite , estimate the following quantity @xmath15 the latter , approximate equality is due to the fact that for most observed gmcs @xmath13 is smaller than @xmath14 .
obviously , @xmath16 increases over the life time of a cloud and should not be confused with either @xmath9 or @xmath1 . from ( [ eq : dotmstar ] ) we can estimate @xmath17 , hence @xmath18 , where @xmath19 is a constant fudge factor of order unity that depends on the actual time evolution of the sfrs and @xmath14 ( and @xmath1 if it is time dependent ) .
we will estimate @xmath19 for a simple toy model in section [ sect : model ] . combining this result with equation ( [ eq : gmceps ] ) and ( [ eq : gmceta ] ) we obtain : @xmath20 \left[\frac{m_*(t)}{m_*({\rm final})}\right]\left[\frac{\max(m_\h2)}{m_\h2(t)}\right].\ ] ]
there are several ways of creating large values of @xmath21 and they correspond to the various terms in this equation .
first , @xmath1 could be time dependent .
for instance , it could smoothly increase as the cloud collapse advances or , alternatively , vary stochastically about some average value .
a second possibility is that some clouds may live for many free fall times , i.e. @xmath22 is large in a subset of gmcs .
the third factor in the third bracket in equation ( [ eq : gmceps2 ] ) explain why @xmath21 can also be _ smaller _ than @xmath9 .
finally , @xmath21 can be boosted if the observed @xmath0 mass is significantly less than @xmath12 , i.e. if gmcs lose ( in one way or another ) a large fraction of their molecular hydrogen over their life time .
the latter scenario predicts that @xmath16 should roughly scale @xmath23 over the life time of _ individual _ gmcs .
an observational sample of an _ ensemble _ of gmcs shows this trend @xcite .
however , this trend can also be produced by a selection effect based on stellar mass , e.g. selecting gmcs with @xmath24 excludes values of @xmath21 that are smaller than @xmath25 , see equation ( [ eq : gmceta ] ) . in fact , @xcite is selecting clouds based on ionizing luminosities , which roughly corresponds to selecting clouds based the stellar mass formed within the last 4 myr .
such a selection effect explains why a different study of @xmath26 gmcs find much lower efficiencies @xcite .
the existence of the selection effect is _ not _ an argument against or in favor of an evolving @xmath1 , rather it shows that the gmcs with large values of @xmath21 in the sample of @xcite are likely a heavily biased subset . in the case that @xmath1 is , in fact , a non evolving quantity and the measured large values of @xmath21 are driven by changing molecular gas masses , we can make a rather generic prediction .
the similarity of the scaling with gmc mass vs. @xmath14 for the data presented in @xcite gives a slope of @xmath27 .
this is consistent with the prediction of our toy model ( slope @xmath28 , see section [ sect : model ] ) that takes into account that , in fact , not the total stellar mass has been measured , but only the stellar mass formed within the last @xmath8 4 myr . ]
( @xmath23 ) of @xmath21 , on the one hand , and the lower boundary of the region excluded by the discussed selection effect , on the other hand , implies that the observed gmcs with large values of @xmath21 should have rather similar _ maximal _
@xmath0 masses @xmath12 .
the toy model that we discuss in section [ sect : model ] predicts @xmath29 .
we note that this scenario explains rather naturally the absence of massive ( @xmath30 @xmath31 ) gmcs with high values of @xmath21 .
a different issue can arise if one compares star formation rates and @xmath0 masses in order to estimate @xmath32 via equation ( [ eq : dotmstar ] ) .
for example , let us assume that we measure sfrs and @xmath0 masses within small ( @xmath7 pc ) apertures around peaks of co emission ( tracing the @xmath0 mass ) and peaks of @xmath33 emission ( tracing star formation rates ) , see e.g. @xcite . if we observe that co peaks have lower sfrs at given @xmath0 mass compared with peaks of @xmath33 emission ,
does this imply a time - varying @xmath32 ?
the answer to that question depends on the way the sfrs are measured .
sfrs that are derived from @xmath33 emission are effectively averaged over the past 5 - 10 myr , which might well be a significant fraction of the life time of the molecular cloud . for sfrs that are based on @xmath33+@xmath34 emission
this averaging time span would be even longer .
the star formation efficiencies per free - fall time that are estimated from such a time averaged sfr will be small initially ( no stars have been formed over most of the time averaging interval simply because the gmc has only formed recently ) .
the measured sfrs will increase until the age of the gmc is similar to the averaging time span .
in addition , the @xmath0 mass of the cloud might evolve ( possibly decrease ) leading to an additional increase in the apparent value of @xmath32 with time .
if the following three conditions are satisfied , a difference in the measured sfr per measured @xmath0 mass can provide strong evidence for a time - varying star formation efficiency per free fall time .
first , the averaging times of the sfrs need to be small compared to ages of the observed clouds .
second , the observable @xmath0 reservoirs need to be close to @xmath12 , and , finally , the free fall times of the clouds need to be known . a recent study that measures sfrs with reasonably short averaging times ( 2 myr , @xcite ) estimates star formation efficiencies per free fall time of the order of @xmath35 for most clouds in the sample , with the scatter mostly driven by the mass of molecular gas of relatively low density ( @xmath36 @xmath37 ) that does not participate in the star formation
we will now discuss a toy model in order to both exemplify the points made in section [ sect : pitfalls ] , but also to provide a framework in which we can make some quantitative predictions .
we should stress that the statements made in the previous section are completely generic and do not depend on the specific assumption that go into the model that we are going to present .
our model is almost insultingly simple , and , given that , our aim is not to reproduce the full complexity in the evolution of gmcs or even , to be consistent with any available observation . on the other hand
the model offers a pragmatic approach to the mass evolution of gmcs and may be easily generalized to facilitate more complex scenarios .
the ansatz of the model is to supplement equation ( [ eq : dotmstar ] ) with an equivalent equation that describes the evolution of the @xmath0 mass : @xmath38 the extra term @xmath39 is motivated by assuming that stellar feedback is limiting the life time of molecular clouds , e.g. via photo - ionization , thermal pressure or radiation pressure @xcite .
this feedback should therefore couple to the formed stellar mass via some efficiency factor @xmath40 that sets the time scale for the destruction / removal of @xmath0 from the cloud should refer to the total stellar mass times a weight parameter that takes into account that feedback is provided by stars which have a limited life time .
for simplicity we will assume that @xmath13 is the total amount of stellar mass formed within the cloud . ] .
the term @xmath41 is the net `` accretion '' rate of @xmath0 , which includes all processes that create and destroy @xmath0 and are not directly coupled to either @xmath13 or @xmath14 .
both @xmath40 and @xmath41 could in principle be time dependent . for simplicity
we assume that they are constant .
our model is minimalistic ( compared with , e.g. , @xcite ) , but it has the advantage that we can parametrize our ignorance of the relevant physical processes that destroy and disperse the cloud into the parameters @xmath40 and @xmath41 .
together with appropriate initial conditions equations ( [ eq : dotmstar ] ) and ( [ eq : dotmgas ] ) fully determine the evolution of the masses of molecular hydrogen and the stellar component in a gmc .
we will also make the simplifying assumption that the free - fall time does not evolve strongly over the history of the gmc , i.e. both the star formation efficiency per free fall time and the star formation time scale are now fixed .
this assumption is not crucial for the model , but we will use it for the following reasons .
first , there is no clear systematic trend of free fall time with mass over the range of gmcs that we are comparing to , see e.g. table 2 of @xcite .
second , assuming a non - evolving free - fall time allows for a convenient analytical solution of the problem .
third , we find that even with this assumption our model describes the observed data reasonably well .
we stress that our main aim is to show that a simple model can produce an observational signal that could be misinterpreted as evidence for evolution of the star formation efficiencies .
we do not try to model the precise properties of the ensemble of gmcs in the galaxy . with @xmath2 fixed ( and , of course , we assume that the star formation efficiency per free fall time is a constant , too ) we can insert ( [ eq : dotmstar ] ) into ( [ eq : dotmgas ] ) and obtain a linear 2nd order differential equation for @xmath14 , i.e. the equation of a damped harmonic oscillator . * _ no accretion scenario : _ assumes @xmath47 , @xmath48 , and @xmath49 .
it follows @xmath50 , and @xmath51 . *
_ pure accretion scenario : _ assumes that all @xmath0 is `` accreted '' , i.e. @xmath52 , @xmath49 and @xmath53 . in this case phase and amplitude
are given by @xmath54 , @xmath55 .
we adopt the parameters @xmath56 and @xmath57 myr , which are consistent with observations of @xmath1 over a range of density scales @xcite , and with the free fall times @xmath58 myr measured in the sample of @xcite , respectively .
we note that only the ratio @xmath59 myr@xmath60 enters our model .
the @xmath40 parameter is chosen such that the life time of the cloud , i.e. the time @xmath61 at which @xmath62 , is @xmath63 myr @xcite . hence , we use @xmath64 in the no accretion scenario and @xmath65 in the pure accretion scenario , respectively .
assuming @xmath66 the life time of a gmc is given by @xmath67 the left ( right ) expression refers to the no accretion ( pure accretion ) scenario .
we note that in both considered scenarios the life time does not depend on the initial cloud mass or the accretion rate , respectively .
the evolution of @xmath14 , @xmath13 and @xmath68 normalized to @xmath12 is shown in fig .
[ fig : tm ] . in fig .
[ fig : eps ] we show the predictions for @xmath21 and @xmath69 of the two scenarios of our model , together with @xmath1 and @xmath9 , and the observational data from @xcite . to be consistent with observations only the stellar mass that formed within the last 4 myr is included in the definition @xmath21 , see ( [ eq : gmceta ] ) , and @xmath69 is estimated as @xmath70 .
our model reproduces the trends of @xmath21 and @xmath69 with gmc mass , suggesting that these are maybe not solely due to selection effects . in both scenarios @xmath21 and @xmath69 roughly scale as @xmath71 over the mass range @xmath72 - @xmath73 . with the chosen parameters our model predicts that @xmath21 is only significantly larger than @xmath9 for the last @xmath74 myr in the life of a gmc , this includes most of the gmcs with masses less than @xmath75 in the sample of @xcite .
we note that the precise time does depend on the assumed life time of the cloud .
clouds with shorter life times spend _ more _ time in a state in which @xmath76 .
our model exemplifies that it is difficult to prove the existence of a time - varying star formation efficiency based on observational quantities such as @xmath21 or @xmath69 .
this is not to say that such a time - dependence does not exist , we merely conclude that current observational evidence for its existence is insufficient .
|
a star formation efficiency per free fall time that evolves over the life time of giant molecular clouds ( gmcs ) may have important implications for models of supersonic turbulence in molecular clouds or for the relation between star formation rate and @xmath0 surface density .
we discuss observational data that could be interpreted as evidence of such a time variability .
in particular , we investigate a recent claim based on measurements of @xmath0 and stellar masses in individual gmcs .
we show that this claim depends crucially on the assumption that @xmath0 masses do not evolve over the life times of gmcs .
we exemplify our findings with a simple toy model that uses a constant star formation efficiency and , yet , is able to explain the observational data .
|
the experiments have been performed on a 1.9 @xmath35 thick degenerately doped @xmath1-gan : si film with electron concentration ( @xmath3610@xmath37)@xmath38 and carrier mobility , @xmath39 , grown by metal organic vapour phase epitaxy ( movpe ) on @xmath40-@xmath41 . a 10 nm permalloy (
py = @xmath42 ) film is used as fm layer and source of spins .
the py film is passivated with a 6 nm @xmath43 film which protects the fm from oxidation in order to avoid the detrimental effects of oxidised py on the spin pumping efficiency @xcite .
the py/@xmath1-gan : si bilayer is driven to resonance conditions at room temperature under an x - band microwave excitation of ( @xmath44 ) ghz with an external magnetic field in a bruker elexsys e580 electron paramagnetic resonance spectrometer . the permalloy being a soft fm has a small magnetocrystalline anisotropy , so that @xmath23 in py is aligned along the film plane when an in - plane magnetic field @xmath20 is applied .
ohmic contacts to @xmath1-gan : si are fabricated by e - beam evaporation of ti / au / al / ti / au as a metallic stack . the quartz sample holder for the fmr measurements is provided with two high conducting copper wires and connected to a keithley 2700 dmm for measuring the generated dc voltage .
the dc voltage due to inverse spin hall- , thermal- and galvanomagnetic - effects is measured and the value of the component due to inverse spin hall effect is employed to calculate the spin hall angle in @xmath1-gan : si . control experiments are also performed on contacted @xmath1-gan : si without py , on py/@xmath40-@xmath41 , on bare @xmath40-@xmath41 substrates and on the wired sample holder solely .
however , no measurable voltage or fmr absorption have been detected from these test samples and control experiments , ruling out experimental artefacts affecting the observed results .
a schematic representation of the spin pumping mechanism and of the geometry employed for the detection of the generated spin current in the nm are shown in figs.[fig : fig1](a ) and [ fig : fig1](b ) , respectively .
the sketches depict the magnetization precession in the fm for an applied magnetic field @xmath20 at resonance condition under microwave excitation , leading to the pumping of spin angular momentum and subsequent generation of spin and charge currents in the nm through the she and ishe .
measurements are carried out with the magnetic field applied in the in - plane and out - of - plane configuration , respectively .
spin pumping is quenched for a perpendicular magnetic field and the angle dependent measurement of the emf is essential to rule out secondary effects or experimental artefacts .
the direction of the applied magnetic field w.r.t the sample plane is indicated in fig.[fig : fig1](c ) , while the first derivative of the fmr signal for py/@xmath40-al@xmath45o@xmath46 and py/@xmath1-gan : si/@xmath40-al@xmath45o@xmath46 is provided in fig.[fig : fig1](d ) .
the dotted and solid lines represent the fmr line - shapes for py/@xmath40-al@xmath45o@xmath46 and for the py/@xmath1-gan : si bilayer , respectively .
the broadening of the fmr signal for the py / gan : si bilayer w.r.t .
the reference py/@xmath40-al@xmath45o@xmath46 layer is the evidence of adiabatic spin pumping from the ferromagnetic py into the rashba semiconductor @xmath1-gan : si under resonance conditions @xcite .
the fmr signal and the electric potential difference between the electrodes attached to the @xmath1-gan : si layer are measured to detect the ishe .
the dynamics of the magnetization @xmath23 in py under an effective magnetic field @xmath18 is described by the llg eq .. in fmr regime , the spin pumping driven by dynamical exchange interaction , pumps into @xmath1-gan : si pure spin current , which gets converted into charge current @xmath0 ishe , according to the relation @xmath47 , as discussed earlier . the charge current in the rashba semiconductor leads to an emf proportional to the generated spin current and whose amplitude is proportional to the microwave absorption and maximized at the resonance field @xmath48 . :
gan : si layer for in - plane magnetic field orientations at ( a ) @xmath49 and ( b ) : @xmath50 for @xmath51 .
dots : experimental data ; solid lines : fitting according to eq .. [ fig : fig2],width=9 ] the emf @xmath52 generated in the nm @xmath1-gan : si is detected simultaneously with the fmr and at the resonance field a peak is observed in the measured voltage .
the experimental emf is plotted as a function of the applied magnetic field @xmath20 and reported in figs.[fig : fig2](a ) and [ fig : fig2](b ) respectively for @xmath53 and @xmath54 .
the angles @xmath55 and @xmath56 define the directions of the applied in - plane magnetic field in accordance to the geometry provided in fig.[fig : fig1](c ) .
since the rashba - edelstein effect is observed in 2degs and in materials with giant rashba soc @xcite , and given that the system treated here belongs to neither of these two classes , the observed emf is attributed to the interplay between she and ishe .
the experimental emf @xmath52 is a superposition of the voltage due to ishe in the @xmath1-gan : si layer @xcite and of spurious voltages originating from galvanomagnetic effects like ordinary hall effect ( ohe ) in @xmath1-gan : si , anomalous hall effect ( ahe ) , planar hall effect ( phe ) in the py layer and thermal heating effects due to microwave irradiation @xcite .
the voltage originating from these spurious mechanisms can be separated from the one due to ishe by fitting the experimental data with the function @xcite@xmath14 @xmath57 where @xmath58 is the half line width of the fmr line - shape and @xmath48 is the resonance field .
the symmetric part of the total voltage function , @xmath59 $ ] with an absorption line - shape is the contribution of the ishe voltage , @xmath60 , developed in @xmath1-gan : si due to the adiabatic spin pumping and to heating effects . on the other hand ,
the asymmetric part of the function , @xmath61/\left[(h - h_\mathrm{fmr})^2+\vargamma^2\right]$ ] with a dysonian dispersion line - shape is a consequence of the contributions from ahe and ohe , as discussed earlier .
the hall voltages @xmath62 change sign across @xmath48 , while @xmath60being proportional to the integrated microwave absorption intensity is symmetric across @xmath48 , as expected from the fundamental spin pumping model @xcite .
a fitting of the measured emf at a microwave power of 200 mw with eq . for @xmath63 yields @xmath64 and @xmath65 .
the ratio @xmath66/@xmath62@xmath76 indicates that the major contribution to the measured voltage is provided by the efficient conversion of spin - to - charge current due to an interplay of spin pumping and direct and inverse spin hall effects .
the spurious heating effects from the microwave irradiation can be further eliminated by averaging the symmetric voltage for parallel and antiparallel orientation of the applied field @xmath67 according to@xmath14
@xmath68 this approach to the treatment of the data is justified , since @xmath60 changes sign upon reversal of the magnetic field owing to a change of sign of the spin polarization vector @xmath69 , not occurring for the voltage due to the microwave heating effects .
thus , for spin pumping experiments , the measured emf @xmath52 includes contributions from both symmetric and asymmetric voltages , and the actual ishe voltage @xmath60 is estimated by a proper treatment of the measured data .
it was reported @xcite , that the symmetric voltage can also include contributions from phe and galvanomagnetic effects in the metallic fm layer due to the electric and magnetic field components of the microwave . however , the angular dependence of the emf w.r.t .
the external magnetic field showed that the contribution of the ishe to the symmetric voltages at least five times greater than the ones from the phe @xcite . in the above mentioned works , the thickness of the nm layer was of the order of the spin diffusion length @xmath70 , while the thickness of @xmath1-gan : si studied here is @xmath71 , @xmath24 much greater than @xmath7280 nm in @xmath1-gan .
the fact that the thickness of the nm is orders of magnitude greater than @xmath70 suppresses the backflow spin current into the metallic fm , further reducing the contributions to the emf from galvanomagnetic effects .
careful identification and estimation of the various components of the generated emf in the nm layer in a spin pumping experiment is therefore essential for the estimation of the spin - to - charge conversion efficiency of a material . here , for the in - plane magnetic field , after eliminating the heating effects , a @xmath73
is obtained and is exploited for the quantitative evaluation of the spin hall angle for @xmath1-gan : si .
the dependence of the measured emf on the applied microwave power under fmr conditions is shown in figs.[fig : fig3](a ) and [ fig : fig3](c ) for orientations of the magnetic field @xmath53 and @xmath54 , respectively . for a microwave power of 200 mw ,
the measured voltages at fmr conditions are + 2.849 @xmath74 for @xmath53 and -3.840 @xmath74 for @xmath54 .
it is to be noted here that the voltages reported in figs.[fig : fig3](a ) and [ fig : fig3](c ) contain both the symmetric and asymmetric contributions . after separating the symmetric and asymmetric voltage components , the symmetric part of the voltage , @xmath66 is plotted as a function of @xmath75 and shown in figs.[fig : fig3](b ) and [ fig : fig3](d ) for @xmath53 and @xmath54 , respectively
now , being @xmath60 proportional to the square of the microwave magnetic field ( @xmath76 ) , it is expected to be linearly proportional to the microwave power @xmath75 [ ] . this is validated in figs.[fig : fig3](b ) and [ fig : fig3](d ) for @xmath53 and @xmath54 , respectively .
the reversal of @xmath20 causes @xmath69 to change sign , which in turn induces the change in sign of the ishe electric field @xmath77 , as evidenced in figs.[fig : fig3](a ) and [ fig : fig3](c ) .
it can be concluded that the fundamental relation for the ishe , @xmath78 operates in the studied py/@xmath1-gan : si bilayer system . for an applied magnetic field perpendicular to the sample plane , @xmath24 @xmath79 , the amplitude of @xmath60
is quenched even though a fmr signal for the py is detected at the resonance field of 1245 mt .
this result provides the necessary and sufficient confirmation of spin pumping through she and ishe in @xmath1-gan : si .
the magnetization precession described by the llg eq . drives the spin pumping in the py/@xmath1-gan : si film . within the fundamental model of spin
pumping proposed by tserkovnyak @xmath4 @xcite , the dc component of the generated spin current density @xmath80 at the interface between py and @xmath1-gan : si is : @xmath81_{z}dt\label{eq.4}\ ] ] where @xmath82 and @xmath83 are the dirac constant and real part of the spin mixing conductance , while @xmath84_{z}$ ] is the @xmath85 component of @xmath84 $ ] .
now , @xmath83 is proportional to the reflection and transmission coefficients of the majority and minority spins of the nm electrons , which in turn depend on the transparency of the fm / nm interface @xcite .
thus , a transparent interface @xmath24 one with negligible spin scattering potential centres , would enhance @xmath83 , by augmenting the spin current density in the nm . the resonance condition obtained as a solution of the llg equation
is given by : @xmath86 where @xmath48 is the resonance field , @xmath87 the frequency of the microwave radiation and @xmath17 is the gyromagnetic ratio .
using eqs .
and and the dynamic components of the magnetization @xmath23 obtained as a solution of eq . with the components of @xmath18 as discussed above @xcite , the spin current density @xmath80 is : @xmath88}{8{\pi}\alpha^{2}\left[(4{\pi}m_{s})^2\gamma^2 + 4\omega^2\right]}\label{eq.6}\ ] ] the spin hall angle @xmath89 is related to the spin current @xmath80 and to the inverse spin hall voltage @xmath60 as : @xmath90 where @xmath91 , @xmath92 and @xmath70 are the thickness , conductivity and spin diffusion length of the @xmath1-gan : si layer , respectively , while @xmath93 and @xmath94 are the thickness and conductivity of the ferromagnet py .
the real part of the spin mixing conductance @xmath83 is given by @xcite : @xmath95 where @xmath96 and @xmath97 are the land @xmath96 factor and the bohr magneton , @xmath98 and @xmath99 the spectral width of the py / gan : si and py layers , as shown in fig.[fig : fig1](d ) . with @xmath100 , @xmath101 , @xmath102= ( @xmath103)m , @xmath104=(@xmath105)jt@xmath11 , @xmath87=(@xmath106)s@xmath11 and @xmath107=(@xmath108)h / m
, we calculate the spin mixing conductance @xmath83 to be ( @xmath109)m@xmath8 . using eqs . and and with the parameters @xmath17=@xmath110)t@xmath11s@xmath11 , @xmath111 , @xmath82=(@xmath112)js , @xmath113 , @xmath114 , @xmath115 , @xmath116 , @xmath92=(@xmath117)@xmath118m@xmath11 , @xmath119=(@xmath120)@xmath118m@xmath11 , @xmath121=(@xmath122)jm@xmath8 and @xmath123 we find the spin hall angle for @xmath1-gan : si to be @xmath124 , at least one order of magnitude higher than those reported for si [ ] , ge [ ] , zno [ ] and @xmath1-gaas [ ] .
the spin hall angle @xmath89 depends on both the side jump and the skew scattering mechanisms .
theoretically@xcite the side jump contribution to @xmath89 is given by @xmath125 , which corresponds to @xmath126 in the case under consideration .
the theoretical value is at least one order of magnitude higher than the one obtained experimentally .
defining the evolution of she in terms of mobility of the system in question according to vignale @xmath4 @xcite , with @xmath39 , @xmath1-gan : si falls in the limit of clean - ultraclean regime , which is dominated by the skew scattering .
considering the above mentioned overestimation of the magnitude of the side jump contribution and the consequent discrepancy between the theoretical and experimental values of @xmath2 , one can infer that she in @xmath1-gan : si is dominated by skew scattering .
@xmath1-gan : si control sample under a microwave excitation of power @xmath127 .
dots : experimental data ; solid lines : fitting according to eq .. ( b ) measured @xmath60 and @xmath62 as a function of the microwave power .
lines : linear fits.,width=6 ] the experiments discussed above for the @xmath71 thick @xmath1-gan : si layer have been carried out also on a @xmath128 thick @xmath1-gan : si film . like in the case of the @xmath71 @xmath1-gan : si sample , an @xmath43 passivated 10 nm py film is used as the source of spin . the spin hall voltage measured in the 150 nm thick @xmath1-gan
: si is reported in fig.[fig : fig4 ] . here , the intensity of the asymmetric component of the voltage is more than one order of magnitude higher than the one observed for the @xmath129 @xmath1-gan : si sample , as shown in figs.[fig : fig4](a ) and [ fig : fig4](b ) . for a nm layer thinner than its spin diffusion length @xmath70 , the spin backflow current @xmath130 from the nm to fm exceeds the forward spin current due to spin pumping @xmath131 [ ] . in such a scenario , @xmath60 due to the inverse spin hall effect can not be detected in the nm .
however , the asymmetric voltages can still be seen , being independent of spin pumping . in the present case of @xmath1-gan : si
, the spin diffusion length is @xmath70 @xmath7 80 nm .
it was reported that for nm metal systems , the asymmetry voltage contribution for a film thickness of the order of @xmath70 is close to 100% of the total emf measured .
this is due to the dominance of @xmath130 over @xmath131 [ ] . however , with a greater thickness of the nm film , the asymmetric contribution diminishes and approaches a minimum saturation value for a film thickness @xmath132 .
furthermore , it was reported by flovik @xmath4 @xcite that this strong @xmath62 could be assigned to eddy current effects and it was shown that in the case of a nm layer of pt , the asymmetry decreases considerably with increasing the thickness of the nm layer . in the systems considered here , for a thin layer of @xmath1-gan : si the spin backflow is likely to be the dominant mechanism that leads to the suppression of the @xmath133 signal , since the oersted fields induce a significant distortion of the fmr lineshape , not observed for the @xmath71 thick @xmath1-gan : si sample . for the case of zno , also a wide band gap wurtzite semiconductor like gan , dambrosio @xmath4 [ ] reported the measured emf from the zno layer in a py / zno bilayer under fmr conditions , as due to phe in the py layer . in our case of @xmath1-gan : si , the control experiments as previously mentioned did not reveal any measurable voltage due to phe in the py layer .
moreover , the thickness of the zno film used in case of dambrosio @xmath4 was 200 nm , making spin backflow a likely reason for the observed dominant contribution from phe .
on the other hand , in our case for calculating the spin hall angle @xmath89 we carry out measurements on a @xmath71 thick @xmath1-gan : si layer , which can be considered as a bulk system where the role of spin backflow is largely suppressed as discussed previously in detail upon comparison with the 150 nm @xmath1-gan : si layer . a detailed investigation of the dependence of spin pumping on the thickness and carrier concentration of @xmath1-gan
: si is needed for an in - depth understanding of this behaviour .
we have provided experimental demonstration of spin pumping induced spin current generation and its detection at room temperature using the ishe in the rashba semiconductor @xmath1-gan : si . from the fundamental model of spin
pumping , a spin mixing conductance of ( @xmath134 for the py/@xmath1-gan : si interface and a spin hall angle @xmath124 for wz @xmath1-gan : si are found .
the value obtained for @xmath89 is at least one order of magnitude higher than those of other semiconductors like si , ge , zno and @xmath1-gaas . the experimental demonstration of generation of pure spin current in @xmath1-gan : si and its enhanced spin - charge conversion efficiency over other functional semiconductors points at iii - nitrides as model systems for studies on spin - related phenomena in non - centrosymmetric semiconductors . moreover , this work paves the way to the realization of nitride - based low power optoelectronic , non - volatile and low dissipative spin devices like @xmath135 spin batteries .
this work was supported by the austrian science foundation fwf ( p22477 and p24471 and p26830 ) , by the nato science for peace programme ( project no .
984735 ) and by the eu @xmath136 framework programmes : capacities project regpot - ct-2013 - 316014 ( eagle ) , eit+ ( nanomat p2ig.01.01.02 - 02 - 002/08 ) and fundms advanced grant of the european research council ( erc grant no.227690 ) . 68 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ] @bib@innerbibempty link:\doibase 10.1103/physrevapplied.4.047001 [ * * , ( ) ] http://dx.doi.org/10.1038/nnano.2014.279 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4913887 [ * * , ( ) ] @noop * * , ( ) http://dx.doi.org/10.1038/nnano.2013.29 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.87.1213 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0038-1098(90)90963-c [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.035310 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.1539 [ * * , ( ) ] link:\doibase 10.1098/rspa.1929.0127 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.83.1834 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.85.393 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.126603 [ * * , ( ) ] link:\doibase 10.1126/science.1087128 [ * * , ( ) ] `` , '' ( , ) pp .
link:\doibase 10.1007/s10948 - 009 - 0547 - 9 [ * * , ( ) ] link:\doibase 10.1126/science.1105514 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.047204 [ * * , ( ) ] http://dx.doi.org/10.1038/nmat3279 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.88.117601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.224403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.060404 [ * * , ( ) ] http://stacks.iop.org/0268-1242/29/i=4/a=043002 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.77.1375 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2199473 [ * * , ( ) ] url = { http://link.aps.org/doi/10.1103/physrevlett.115.246602 } [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.014413 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3587173 [ * * , ( ) ]
link:\doibase 10.1103/physrevb.83.144402 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4922581 [ * * , ( ) ] link:\doibase doi = { 10.1103/physrevb.91.144412 } , [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.214419 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.140406 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.066604 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms3055 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms10296 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms1640 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.196602 [ * * , ( ) ] http://stacks.iop.org/1347-4065/54/i=9/a=093001 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4863750 [ * * , ( ) ] http://dx.doi.org/10.1021/acs.nanolett.5b03274 [ * * , ( ) ] http://scitation.aip.org/content/aip/journal/aplmater/4/3/10.1063/1.4941276 [ * * , ( ) ] http://dx.doi.org/10.1038/nature08876 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.216601 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.113.266602 [ * * , ( ) ] url = { http://link.aps.org/doi/10.1103/physrevlett.116.077203 } [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.057601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.035206 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.87.1311 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.89.205201 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.cap.2014.10.004 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms3944 [ * * , ( ) ] tp://dx.doi.org/10.1038/nmat3052 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1735851 [ * * , ( ) ] http://stacks.iop.org/1882-0786/7/i=1/a=013002 [ * * , ( ) ] url = { http://link.aps.org/doi/10.1103/physrevb.92.224436 } [ * * , ( ) ] http://dx.doi.org/10.1038/nphys3304 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.064426 [ * * , ( ) ] http://scitation.aip.org/content/aip/journal/apl/103/23/10.1063/1.4839475 [ * * , ( ) ] http://stacks.iop.org/1882-0786/6/i=2/a=023001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.205204 [ * * , ( ) ] link:\doibase 10.1109/tmag.2013.2262947 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.064414 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4917285 [ * * , ( ) ]
|
pure spin currents in semiconductors are essential for implementation in the next generation of spintronic elements .
heterostructures of iii- nitride semiconductors are currently employed as central building - blocks for lighting and high - power devices . moreover ,
the long relaxation times and the spin - orbit coupling ( soc ) in these materials indicate them as privileged hosts for spin currents and related phenomena .
spin pumping is an efficient mechanism for the inception of spin current and its conversion into charge current in non - magnetic metals and semiconductors with rashba soc @xmath0 spin hall effects .
we report on the generation in @xmath1-gan : si at room temperature and through spin pumping of pure spin current , fundamental for the understanding of the spin dynamics in these non - centrosymmetric rashba systems .
we find for @xmath1-gan : si a spin hall angle @xmath2=@xmath3 , exceeding by one order of magnitude those reported for other semiconductors , pointing at iii - nitrides as particularly efficient spin current generators . in the emerging field of spin - orbitronics @xcite spin - orbit coupling ( soc )
is employed in both magnetic and non - magnetic materials to generate , exploit and detect spin currents .
spin currents hold the key for the realization and implementation of the next generation of spin based nanoelectronic devices with properties like non - volatility , low power consumption and dissipation . while in magnetic materials
the soc is employed to create new classes of topological objects like magnetic skyrmions or dzyalonshinskii - moriya domain walls @xcite , spin - orbitronics in non - magnetic materials mostly addresses the spin - to - charge conversion through the spin hall effect ( she ) @xcite and the rashba - edelstein effect @xcite .
the concept of she is borrowed from the anomalous hall effect ( ahe ) where , due to the relativistic soc , asymmetric deflection of charge carriers takes place depending on their spin orientations @xcite .
while ahe is studied in magnetic systems , she is mostly observed in non - magnetic ones .
based on the concept of spin dependent mott scattering @xcite , the she was predicted nearly four decades ago by dyakonov and perel @xcite and was proposed to be the effective process for producing pure spin currents in solid state systems .
however , it was not until the theoretical work of hirsch @xmath4 @xcite and zhang @xmath4 @xcite , that the extrinsic she received renewed attention .
the possibility of an intrinsic mechanism was proposed by various theoretical groups @xcite .
the intrinsic she depends only on the electronic band structure of the material .
this effect arises from the nonequilibrium dynamics of the bloch electrons as they undergo spin precession due to an induced @xmath5dependent effective magnetic field , such as the rashba field .
the extrinsic she , on the other hand , is the mechanism in which the spins acquire transverse velocity due to soc during scattering of electrons @xcite .
the extrinsic she is classified according to two different underlying mechanisms @xmath6 the skew scattering and the side jump mechanism .
skew scattering is the asymmetric scattering of spin , within the scattering plane , due to an effective magnetic field gradient that arises as an effect of the soc .
the scattering plane defines the spin polarization direction of the resulting spin current .
the side jump mechanism is the velocity integrated over time of deflection of electrons in opposite directions by the electric fields experienced when approaching or withdrawing an impurity .
this phenomenon results in an effective transverse displacement of the electrons upon multiple scattering events . at low carrier mobilities @xmath7(10@xmath8 - 10@xmath9)@xmath10v@xmath11s@xmath11 ,
the intrinsic and extrinsic side jump mechanisms contribute mostly to the she , while for carrier mobilities greater than @xmath12 , extrinsic skew scattering is the dominant process @xcite .
moreover , for mobilites exceeding @xmath13 , spin coulomb drag becomes the dominant mechanism @xcite and compels the spin hall conductivity towards a saturation value .
the lack of direct electrical signals proved to be a major challenge in the observation of this effect , so that the initial experimental efforts were mostly accomplished using optical means @xcite . in a series of seminal publications , tserkovnyak , brataas and
bauer @xcite suggested a method for obtaining pure spin current in non - magnetic ( nm ) metals and semiconductors with non - negligible soc .
they proposed a spin battery @xcite based on adiabatic pumping of spins from a ferromagnetic metal or insulator ( fm ) grown in a fm / nm bilayer configuration when the system is driven to resonance under microwave irradiation .
such a battery leads to the dynamic generation of pure spin current in a nm with non trifling soc @xmath0 ferromagnetic resonance ( fmr ) of the fm @xcite .
the magnetization dynamics of ferromagnets is well described by the phenomenological landau - lifshitz - gilbert ( llg ) equation @xcite@xmath14 @xmath15 where @xmath16 is the unit vector of magnetization , @xmath17 the gyromagnetic ratio and @xmath18 the effective magnetic field expressed as @xmath19 , with @xmath20 the external magnetic field , @xmath21 the dynamic demagnetizing field and @xmath22 the ac field due to microwave radiation .
the magnetization @xmath23 , is expressed as a sum of the static and dynamic components , @xmath24 @xmath25 and the dimensionless coefficient @xmath26 is the gilbert damping parameter .
the first term on the right hand side of eq .
is the precession term , while the second term represents a damping component that impels the precession of magnetization @xmath23 to spiral down to a static magnetization axis due to the gilbert damping parameter @xmath26 . in a fm / nm hybrid bilayer , @xmath26 is enhanced due to the transfer of spin angular momentum from the fm to the nm through a dynamical process of adiabatic spin pumping @xcite .
the spins pumped in the nm are scattered by the effective spin - orbit field @xmath24 the rashba field and a spin accumulation is achieved in the nm , leading to a spin current in the nm through she .
an enhanced @xmath26 in a fm / nm bilyer is a signature of spin pumping and a fingerprint of the generation of pure spin current in the nm .
the spin current is converted into charge current , through a reciprocal process called the inverse spin hall effect ( ishe ) @xmath0 the relation : @xmath27 , where @xmath28 is the spin hall angle representing the efficiency of spin - to - charge conversion of a material .
the charge current in the nm with soc induces an electromotive force ( emf ) , whose direction is perpendicular to the spin current @xmath29 and to the spin polarization vector @xmath30 .
spin pumping is an efficient mechanism to convert spin current into charge current without any applied bias . over the last few years
several experimental works have been published demonstrating spin pumping in heavy metals like pt [ ] , ta [ ] , pd [ ] , au [ ] , in semiconductors like gaas [ ] , si [ ] , ge [ ] and zno [ ] and recently also in 3d topological insulators [ ] and ferroelectric rashba semiconductors [ ] .
apart from ferromagnetic spin pumping from a metallic fm like permalloy ( py ) , it was also shown that spin currents can be realised in bilayers like pt / yig [ ] based on ferromagnetic insulators .
recent reports also point to spin pumping from paramagnets@xcite and antiferromagnets @xcite . the successful generation and control of spin current in semiconductors using the mechanism of spin pumping would open wide perspectives for the integration of spin functionalities in state - of - the - art electronic and optoelectronic devices based on semiconductors like si , ge , iii - nitrides , iii - arsenides @xmath31 . amongst the conventional iii - v semiconductors , gan and its alloys algan and ingan have emerged as strategic materials for optoelectronic and electronic applications , primarily due to their tunable wide bandgap and structure induced polarization .
furthermore , transition metal doped iii - nitrides have been studied extensively in the past decade as workbench magnetic semiconductors @xcite .
another fundamental aspect , namely the presence of rashba spin orbit coupling ( rsoc ) @xcite , was recently demonstrated in degenerate wurtzite ( wz ) @xmath1-gan : si [ ] . using magnetotransport measurements it was shown that the rashba parameter , @xmath32 , linear in @xmath5 and accounting for the spin splitting of the conduction band in wz - gan , has a quantitative value measured for @xmath1-gan : si to be @xmath7 ( @xmath33)@xmath34 @xmath24 the same magnitude found for a 2-dimensional electron gas ( 2deg ) formed at the algan / gan interface @xcite . in this work it was shown that in polar wz - gan , the inversion asymmetry associated with the wurtzite crystal structure dominates over the interfacial electric field in the conduction band of gan .
the rsoc in degenerate @xmath1-gan : si makes this material a spin hall system for prospective spin - orbitronic applications based on iii - nitride semiconductors . here
we report on the generation of pure spin currents in @xmath1-gan : si using an adiabatic spin pumping technique and we estimate the spin hall angle for @xmath1-gan : si .
the evaluation of the spin - to - charge conversion efficiency in @xmath1-gan : si opens up the possibility to design and implement high performance spin based iii - nitride nanoelectronic and optoelectronic devices .
|
[ sec : intro ] over the past two decades the orbital architecture of giant planets has expanded from a single order of magnitude in the solar system ( 530 au ) to over five orders of magnitude among extrasolar planetary systems ( 0.015000 au ; figure [ fig : mass_sma ] ) .
high - contrast adaptive optics ( ao ) imaging has played a critical role in this advancement by probing separations beyond @xmath410 au and masses @xmath51 .
uncovering planetary - mass objects at hundreds and thousands of au has fueled novel theories of planet formation and migration , inspiring a more complex framework for the origin of giant planets in which multiple mechanisms ( core accretion , dynamical scattering , disk instability , and cloud fragmentation ) operate on different timescales and orbital separations .
in addition to probing unexplored orbital distances , imaging entails directly capturing photons that originated in planetary atmospheres , providing unparalleled information about the initial conditions , chemical composition , internal structure , atmospheric dynamics , photospheric condensates , and physical properties of extrasolar planets .
these three science goals the architecture , formation , and atmospheres of gas giants represent the main motivations to directly image and spectroscopically characterize extrasolar giant planets .
this pedagogical review summarizes the field of direct imaging in the era leading up to and transitioning towards extreme adaptive optics systems , the _ james webb space telescope _ , _ wfirst _ , and the thirty meter - class telescopes .
this `` classical '' period of high - contrast imaging spanning approximately 2000 to 2015 has set the stage and baseline expectations for the next generation of instruments and telescopes that will deliver ultra - high contrasts and reach unprecedented sensitivities .
in addition to the first images of _ bona fide _ extrasolar planets , this early phase experienced a number of surprising discoveries including planetary - mass companions orbiting brown dwarfs ; planets on ultra - wide orbits beyond 100 au ; enigmatic ( and still poorly understood ) objects like the optically - bright companion to fomalhaut ; and unexpectedly red , methane - free , and dust - rich atmospheres at low surface gravities . among the most important results
has been the gradual realization that massive planets are exceedingly rare on wide orbits ; only a handful of discoveries have been made despite thousands of hours spent on hundreds of targets spanning over a dozen surveys .
although dismaying , these null detections provide valuable information about the efficiency of planet formation and the resulting demographics at wide separations . making use of mostly general , non - optimized facility instruments and early adaptive optics systems
has also led to creative observing strategies and post - processing solutions for psf subtraction .
-1.2 in -.3 in -1.1 in distinguishing giant planets from low - mass brown dwarfs is a well - trodden intellectual exercise ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . except in the few rare cases where the architectures or abundance patterns of individual systems offer clues about a specific formation route , untangling the origin of imaged planetary - mass companions
must necessarily be addressed as a population and in a statistical manner .
this review is limited in scope to self - luminous companions detected in thermal emission at near- and mid - infrared wavelengths ( 15 @xmath6 m ) with masses between @xmath0113 with the understanding that multiple formation routes can probably produce objects in this `` planetary '' mass regime ( see section [ sec : dbl ] ) .
indeed , the separations regularly probed in high - contrast imaging surveys typically tens to hundreds of au lie beyond the regions in protoplanetary disks containing the highest surface densities of solids where core accretion operates most efficiently ( e.g. , @xcite ) .
direct imaging has therefore predominantly surveyed the wide orbital distances where alternative formation and migration channels like disk instability , cloud fragmentation , and planet - planet scattering are most likely to apply . in the future , the most efficient strategy to detect even smaller super - earths and terrestrial worlds close to their host stars will be in reflected light from a dedicated space - based optical telescope . by focusing on the optimal targets , early discoveries , largest surveys , and statistical results , this observationally - oriented overview aims to complement recent reviews on giant planet formation ( @xcite ; @xcite ; @xcite ; @xcite ) , atmospheric models ( @xcite ; @xcite , @xcite ; @xcite ) , evolutionary models , ( @xcite ; @xcite ) , observational results ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , and high contrast imaging instruments and speckle suppression techniques ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
-1.2 in -.2 in -1 in
planets radiatively cool over time by endlessly releasing the latent heat generated during their formation and gravitational contraction .
fundamental scaling relations for the evolution of brown dwarfs and giant planets can be derived analytically with basic assumptions of a polytropic equation of state and degenerate electron gas ( @xcite ; @xcite ) . neglecting the influence of lithium burning , deuterium burning , and atmospheres , which act as partly opaque wavelength - dependent boundary conditions , substellar objects with different masses cool in a similar monotonic fashion over time : @xmath7 here @xmath8 is the bolometric luminosity , @xmath9 is the object s age , and @xmath10 is its mass .
this steep mass - luminosity relationship means that luminosity tracks are compressed in the brown dwarf regime ( @xmath01375 ) and fan out in the planetary regime with significant consequences for high - contrast imaging .
a small gain in contrast in the brown dwarf regime results in a large gain in the limiting detectable mass , whereas the same contrast gain in the planetary regime has a much smaller influence on limiting mass ( figure [ fig : contrast_sensitivitylimits ] ) .
it is much more difficult , for example , to improve sensitivity from 10 to 1 than from 80 to 10 .
moreover , sensitivity to low masses and close separations is highly dependent on a star s youth and proximity . in terms of limiting detectable planet mass , observing younger and closer stars
is equivalent to improving speckle suppression or integrating for longer .
note that in a contrast - limited regime the absolute magnitude of the host star is also important .
the same contrast around low - mass stars and brown dwarfs corresponds to lower limiting masses compared to higher - mass stars .
young stars are therefore attractive targets for two principal reasons : planets are their most luminous at early ages , and the relative contrast between young giant planets and their host stars is lower than at older ages because stellar luminosities plateau on the main sequence while planets and brown dwarfs continue to cool , creating a luminosity bifurcation .
for example , evolutionary models predict the @xmath11-band contrast between a 5 planet orbiting a 1 star to be @xmath025 mag at 5 gyr but only @xmath010 mag at 10 myr ( @xcite ; @xcite ) . at old ages beyond @xmath41 gyr
, 110 planets are expected to have effective temperatures between 100500 k and cool to the late - t and y spectral classes with near - infrared absolute magnitudes @xmath518 mag ( @xcite ) . below are overviews of the most common classes of targets in direct imaging surveys highlighting the scientific context , strengths and drawbacks , and observational results for each category . in principle , younger stars make better targets for imaging planets . in practice ,
the youngest t tauri stars reside in star - forming regions beyond 100 pc . at these distances ,
the typical angular scales over which high - contrast imaging can probe planetary masses translate to wide physical separations beyond @xmath42050 au ( with some notable exceptions with non - redundant aperture masking and extreme ao systems ) . moreover , these extremely young ages of @xmath4110 myr correspond to timescales when giant planets may still be assembling through core accretion and therefore might have lower luminosities than at slightly later epochs ( e.g. , @xcite ; @xcite ; @xcite ) . on the other hand , the closest stars to the sun
probe the smallest physical scales but their old ages of @xmath4110 gyr mean that high contrast imaging only reaches brown dwarf masses .
-1.2 in -.6 in -1.2 in young moving groups coeval , kinematically comoving associations young stars and brown dwarfs in the solar neighborhood represent a compromise in age ( @xmath010150 myr ) and distance ( @xmath010100 pc ) between the nearest star forming regions and field stars ( figure [ fig : ymg_sensitivitylimits ] ; see @xcite , @xcite , and @xcite ) .
one distinct advantage they hold is that their members span a wide range of masses and can be used to age date each cluster from lithium depletion boundaries and isochrone fitting ( e.g. , @xcite ; @xcite ) . as a result ,
the ages of these groups are generally much better constrained than those of isolated young stars . for these reasons
young moving group members have emerged as the primary targets for high - contrast imaging planet searches over the past decade ( e.g. , @xcite ; @xcite ; @xcite ) identifying these nearby unbound associations of young stars is a difficult task .
each moving group s @xmath12 space velocities cluster closely together with small velocity dispersions of @xmath012 km s@xmath13 but individual stars in the same group can be separated by tens of parsecs in space and tens of degrees across the sky .
@xmath12 kinematics can be precisely determined if the proper motion , radial velocity , and parallax to a star are known .
incomplete knowledge of one or more of these parameters ( usually the radial velocity and/or distance ) means the @xmath12 kinematics are only partially constrained , making it challenging to unambiguously associate stars with known groups .
historically , most groups themselves and new members of these groups were found with the aid of the tycho catalog and _ hipparcos _ , which provided complete space velocities for bright stars together with ancillary information pointing to youth such as infrared excess from iras ; x - ray emission from the @xmath14 or @xmath15 space observatories ; strong h@xmath16 emission ; and/or @xmath176708 absorption . as a result ,
most of the faint low - mass stars and brown dwarfs have been neglected .
-.45 in -1.1 in -1.4 in in recent years the population of `` missing '' low - mass stars and brown dwarfs in young moving groups has been increasingly uncovered as a result of large all - sky dedicated searches ( figure [ fig : newmgmembers ] ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . parallaxes and radial velocities are generally not available for these otherwise anonymous objects , but by adopting the @xmath12 kinematics of known groups , it is possible to invert the problem and predict a distance , radial velocity , and membership probability .
radial velocities are observationally cheaper to acquire en masse compared to parallaxes , so membership confirmation has typically been accomplished with high - resolution spectroscopy .
the exceptions are for spectroscopic binaries , which require multiple epochs to measure a systemic velocity , and rapidly rotating stars with high projected rotational velocities ( @xmath18sin@xmath19 ) , which produce large uncertainties in radial velocity measurements .
the abundance of low - mass stars in the field means that some old interlopers will inevitably share similar space velocities with young moving groups .
these must be distilled from bona fide membership lists on a case - by - case basis ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . the current census of directly imaged planets and companions near the deuterium - burning limit
are listed in table [ tab : planets ] .
many of these host stars are members of young moving groups .
pic , 51 eri , and possibly tyc 9486 - 927 - 1 are members of the @xmath20 pic moving group ( @xcite ; @xcite ; @xcite ) .
hr 8799 and possibly @xmath21 and are thought to be members of columba ( @xcite ) .
2m12073932 is in the tw hydrae association ( @xcite ) .
gu psc and 2m01222439 are likely members of the ab dor moving group ( @xcite ; @xcite ; @xcite ) .
ab pic , 2m01035515 , and 2m02193925 are in tuc - hor ( @xcite ; @xcite ; @xcite ) , though the masses of their companions are somewhat uncertain and may not reside in the planetary regime .
in addition , the space motion of vhs 12561257 is well - aligned with the @xmath20 pic or possibly ab dor moving groups ( @xcite ; @xcite ) , but the lack of lithium in the host indicates the system is older and may be a kinematic interloper .
the number of moving groups in the solar neighborhood is still under debate , but at least five are generally considered to be well - established : the tw hydrae association , @xmath20 pic , tuc - hor , carina , and ab dor .
other associations have been proposed and may constitute real groups which formed together and are useful for age - dating purposes , but may require more scrutiny to better understand their size , structure , physical nature , and relationship to other groups . @xcite
provide a concise up - to - date census of their status and certitude .
soon , micro - arcsecond astrometry and parallaxes from @xmath22 will dramatically change the landscape of nearby young moving groups by readily identifying overlooked groups , missing members , and even massive planets on moderate orbits . despite their greater distances ( @xmath0120150 pc ) , the extreme youth ( @xmath0110 myr ) of t tauri stars and their massive counterparts ,
herbig ae / be stars , in nearby star - forming regions like taurus , the sco - cen complex , and @xmath23 oph have made them attractive targets to search for planets with direct imaging and probe the earliest stages of planet formation when gas giants are still assembling ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
one of the most surprising results from these efforts has been the unexpected discovery of planetary - mass companions on ultra - wide orbits at several hundred au from their host stars ( table [ tab : planets ] ) .
these wide companions pose challenges to canonical theories of planet formation via core accretion and disk instability and may instead represent the tail end of brown dwarf companion formation , perhaps as opacity - limited fragments of turbulent , collapsing molecular cloud cores ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
many ( and perhaps most ) of these young planetary - mass companions harbor accreting circum - planetary disks , which provide valuable information about mass accretion rates , circum - planetary disk structure , formation route , and the moon - forming capabilities of young planets .
accretion luminosity is partially radiated in line emission , making h@xmath16 a potentially valuable tracer to find and characterize protoplanets ( @xcite ) .
for example , @xcite find that up to 50% of the accretion luminosity in the @xmath015 companion gsc 6214 - 210 b is emitted at h@xmath16 . searching for these nascent protoplanets
has become a leading motivation to achieve ao correction in the optical and is actively being carried out with magao ( @xcite ) .
deep sub - mm observations with alma have opened the possibility of measuring the masses of these subdisks ( @xcite ) and possibly even indirect identification via gas kinematics ( @xcite ) .
larger disks may be able to be spatially resolved and a dynamical mass for the planet may be measured from keplerian motion . the relationship between protoplanetary disks and young planets is also being explored in detail at these extremely young ages .
in particular , transition disks young stars whose spectral energy distributions indicate they host disks with large optically thin cavities generally depleted of dust ( e.g. , see reviews by @xcite , @xcite , @xcite , and @xcite) have been used as signposts to search for embedded protoplanets .
this approach has been quite fruitful , resulting in the discovery of companions within these gaps spanning the stellar ( coku tau 4 : @xcite ; hd 142527 : @xcite , @xcite , @xcite , @xcite ) , brown dwarf ( hd 169142 : @xcite , @xcite ) , and planetary ( lkca 15 : @xcite , @xcite , @xcite ; hd 100546 : @xcite , @xcite , @xcite ; @xcite , @xcite ) mass regimes using a variety of techniques . however , environmental factors can severely complicate the interpretation of these detections . extinction and reddening , accretion onto and from circum - planetary disks , extended emission from accretion streams , and circumstellar disk sub - structures seen in scattered light can result in false alarms , degenerate interpretations , and large uncertainties in the mass estimates of actual companions .
t cha offers a cautionary example ; @xcite discovered a candidate substellar companion a mere 62 mas from the transition - disk host star with aperture masking interferometry , but additional observations did not show orbital motion as expected for a real companion .
additional modeling indicates that the signal may instead be a result of scattering by grains in the outer disk or possibly even noise in the data ( @xcite ; @xcite ; @xcite ) .
this highlights an additional complication with aperture masking : because model fits to closure phases usually consist of binary models with two or more point sources , it can difficult to discern actual planets from other false positives . in these situations
the astrometric detection of orbital motion is essential to confirm young protoplanets embedded in disks .
other notable examples of ambiguous candidate protoplanets at wider separations include fw tau b , an accreting low - mass companion to the taurus binary fw tau ab orbiting at a projected separation of 330 au ( @xcite ; @xcite ) , and tmr-1c , a faint , heavily extincted protoplanet candidate located @xmath01400 au from the taurus protostellar binary host tmr-1ab showing large - amplitude photometric variability and circumstantial evidence of a dynamical ejection ( @xcite ; @xcite ; @xcite ; @xcite ) .
follow - up observations of both companions suggest they may instead be low - mass stars or brown dwarfs with edge - on disks ( @xcite ; @xcite ; @xcite ; @xcite ) , underscoring a few of the difficulties that arise when interpreting candidate protoplanets at extremely young ages .
altogether the statistics of planets orbiting the youngest t tauri stars from direct imaging are still fairly poorly constrained .
quantifying this occurrence rate is important because it can be compared with the same values at older ages to determine the _ evolution _ of this population .
planet - planet scattering , for example , implies an increase in the frequency of planets on ultra - wide orbits over time .
@xcite found the frequency of 620 companions from @xmath0200500 au to be @xmath44@xmath24% in upper scorpius . combining these results with those from @xcite and their own shallow imaging survey ,
@xcite find that the frequency of 540 companions between 50250 au is @xmath31.8% and between 2501000 au is 4.0@xmath25% assuming hot - start evolutionary models .
in future surveys it will be just as important to report nondetections together with new discoveries so this frequency can be measured with greater precision .
young brown dwarfs ( @xmath01375 ) have low circum - substellar disk masses ( @xcite ; @xcite ) and are not expected to host giant planets as frequently as stars . nevertheless
, their low luminosities make them especially advantageous for high - contrast imaging because lower masses can be probed with contrast - limited observations .
several deep imaging surveys with ground - based ao or @xmath26 have included brown dwarfs in their samples ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . a handful of companions in the 515 range
have been discovered with direct imaging : 2m12073932 b ( @xcite ; @xcite ) , 2m0441 + 2301 b ( @xcite ) , and possibly both fu tau b ( @xcite ) and vhs 12561257 ( @xcite ) depending on their ages .
a few other low - mass companions ( and in some cases the primaries themselves ) to late - t and early - y field brown dwarfs may also reside in the planetary regime depending on the system ages : cfbdsir j1458 + 1013 b ( @xcite ) , wise j1217 + 1626 b ( @xcite ) , and wise j0146 + 4234 b ( @xcite ) .
the low mass ratios of these systems ( @[email protected] ) bear a closer resemblance to binary stars than canonical planetary systems ( @[email protected] ) , and the formation route of these very low - mass binaries is probably quite different than around stars ( @xcite ) .
high - order multiple systems with low total masses like 2m0441 + 2301 aabbab and vhs 12561257 abb suggest that cloud fragmentation can form objects in the planetary - mass domain ( @xcite ; @xcite ; @xcite ; @xcite )
. continued astrometric monitoring of ultracool binaries will eventually yield orbital elements and dynamical masses for these intriguing systems to test formation mechanisms ( @xcite ) and giant planet evolutionary models .
close stellar binaries ( @xmath00.15@xmath29 ) are generally avoided in direct imaging surveys .
multiple similarly - bright stars can confuse wavefront sensors , which are optimized for single point sources , and deep coronagraphic imaging generally saturates nearby stellar companions .
physically , binaries carve out large dynamically - unstable regions that are inhospitable to planets and there is strong evidence that they inhibit planet formation by rapidly clearing or truncating protoplanetary disks ( e.g. , @xcite ; @xcite ; @xcite ) .
nevertheless , many planets have been found in binary systems in both s - type orbital configurations ( a planet orbiting a single star ; e.g. , @xcite ) and p - type orbits ( circumbinary planets ; see @xcite for a recent summary ) .
binaries are common products of star formation , so understanding how stellar multiplicity influences the initial conditions ( protoplanetary disk mass and structure ) , secular evolution ( kozai - lidov interactions ) and end products ( dynamically relaxed planetary systems ) of planet formation has important consequences for the galactic census of exoplanets ( e.g. , @xcite ; @xcite ) .
several planetary - mass companions have been imaged around binary stars on wide circumbinary orbits : roxs
42b b ( @xcite ; @xcite ) , ross 458 c ( @xcite ; @xcite ) , sr 12 c ( @xcite ) , hd 106906 b ( @xcite ; @xcite ) , and vhs 12561257 ( @xcite ; @xcite ) .
2m01035515 b ( @xcite ) and fw tau b ( @xcite ) orbit close , near - equal mass stellar binaries , but the masses of the wide tertiaries are highly uncertain ( @xcite ) .
on the other hand , few imaged planets orbiting single stars also have wide stellar companions .
51 eri ab is orbited by the pair of m dwarf binaries gj 3305 ab ( @xcite ; @xcite ; @xcite ) at @xmath42000 au .
fomalhaut has two extremely distant stellar companions at @xmath057 kau and @xmath0158 kau ( @xcite ) .
2m0441 + 2301 bb orbits a low - mass brown dwarf and is part of a hierarchical quadruple system with a distant star - brown dwarf pair at a projected separation of 1800 au ( @xcite ) .
there has been little work comparing the occurrence rate of imaged planets in binaries and single stars .
however , several surveys and post - processing techniques are now expressly focusing on binary systems and should clarify the statistical properties of planets in these dynamically complicated arrangements ( @xcite ; @xcite ; @xcite ) .
debris disks are extrasolar analogs of the asteroid and kuiper belts in the solar system .
they are the continually - replenished outcomes of cascading planetesimal collisions that result in large quantities of transient dust heated by the host star .
observationally , debris disks are identified from unresolved infrared or sub - mm excesses over stellar photospheric emission .
deep observations spanning the optical , ir , and sub - mm can spatially resolved the largest and most luminous disks in scattered light and thermal emission to investigate disk morphology and grain properties .
topical reviews by @xcite , @xcite , @xcite , @xcite , and @xcite highlight recent theoretical and observational progress on the formation , modeling , and evolution of debris disks .
debris disks are intimately linked to planets , which can stir planetesimals , sculpt disk features , produce offsets between disks and their host stars , and carve gaps to form belts with spectral energy distributions showing multiple temperature components .
the presence of debris disks , and especially those with features indicative of a massive perturber , may therefore act as signposts for planets .
the four directly imaged planetary systems @xmath20 pic , hr 8799 , 51 eri , and hd 95086 all possess debris disks , the latter three having multiple belts interior and exterior to the imaged planet(s ) .
this remarkably consistent configuration is analogous to the solar system s architecture in which gas giants are flanked by ( very low - level ) zodiacal emission .
anecdotal signs point to a possible correlation between disks and imaged planets but this relationship has not yet been statistically validated .
there have been hints of a correlation between debris disks and low - mass planets detected via radial velocities ( @xcite ; @xcite ) , but these were not confirmed in a recent analysis by @xcite . indeed , many stars hosting multi - component debris disks have now been targeted with high - contrast imaging and do not appear to harbor massive planets ( @xcite ; @xcite ; @xcite ; @xcite ) . given the high incidence of debris disks around main - sequence stars ( @xmath51620% : @xcite ; @xcite ) , with even higher rates at younger ages ( @xcite ; @xcite ) , any correlation of imaged giant planets and debris disks will be difficult to discern because the overall occurrence rate of massive planets on wide orbits is extremely low ( @xmath281% ; see section [ sec : occurrencerate ] ) .
perhaps more intriguing would be a subset of this sample with additional contextual clues , for example the probability of an imaged planet given a two - component debris disk compared to a diskless control sample .
the fabulous four vega , @xmath20 pic , fomalhaut , and @xmath30 eridani host the brightest debris disks discovered by iras ( @xcite ; @xcite ) and have probably been targeted more than any other stars with high - contrast imaging over the past 15 years , except perhaps for hr 8799 . despite having similarly large and luminous disks , their planetary systems are quite different and demonstrate a wide diversity of evolutionary outcomes .
fomalhaut s disk possesses a sharply truncated , offset , and eccentric ring about 140 au in radius suggesting sculpting from a planet ( @xcite ; @xcite ; @xcite ) . a comoving optical source ( `` fomalhaut b '' )
was discovered interior to the ring by @xcite and appears to be orbiting on a highly inclined and eccentric orbit not coincident with the ring structure ( @xcite ) .
the nature of this intriguing companion remains puzzling ; it may be a low - mass planet with a large circum - planetary disk , a swarm of colliding irregular satellites , or perhaps a recent collision of protoplanets ( e.g. , @xcite ; @xcite ; @xcite ) .
massive planets have been ruled out from deep imaging down to about 20 au ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
vega s nearly face - on debris disk is similar to fomalhaut s in terms of its two - component structure comprised of warm and cold dust belts and wide gaps with orbital ratios @xmath510 , possibly indicating the presence of multiple low - mass planets ( e.g. , @xcite ; @xcite ; @xcite ) .
deep imaging of vega over the past 15 years has thus far failed to identify planets with detection limits down to a few jupiter masses ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
@xmath20 pic hosts an extraordinarily large , nearly edge - on disk spanning almost 2000 au in radius ( @xcite ; @xcite ) .
its proximity , brightness , and spatial extent make it one of the best - studied debris disks , showing signs of multiple belts ( e.g. , @xcite ) , asymmetries ( e.g. , @xcite ) , molecular gas clumps ( @xcite ) , and an inner warp ( @xcite ) predicted to be caused by a close - in inclined planet ( @xcite ) .
@xcite uncovered a possible massive planet at @xmath49 au and , despite immediate follow - up ( @xcite ; @xcite ) , it was not until it reemerged on the other side of the star that @xmath20 pic b was unambiguously confirmed ( @xcite ) .
@xmath30 eridani is another particularly fascinating example of a nearby , relatively young k2 star hosting a bright debris disk with spatially resolved ring structure and a warm inner component ( @xcite ; @xcite ; @xcite ) . at 3.2 pc
this star harbors the closest debris disk to the sun , has a jovian - mass planet detected by radial velocity and astrometric variations ( @xcite ; @xcite ) , and possesses a long - term rv trend pointing to an additional long - period giant planet .
because of its favorable age and proximity , @xmath30 eridani has been exhaustively imaged with adaptive optics on the largest ground - based telescopes in an effort to recover the known planets and search for others ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
together with long - baseline radial velocity monitoring , these deep observations have ruled out planets @xmath313 anywhere in this system . at the old ages of field stars ( @xmath4110 gyr ) , giant planets
have cooled to late spectral types and low luminosities where high - contrast imaging does not regularly reach the planetary - mass regime .
nevertheless , several surveys have focused on this population because their proximity means very close separations can be probed and their old ages provide information about potential dynamical evolution of substellar companions and giant planets over time ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . of particular interest
are stars showing low - amplitude , long - baseline radial velocity changes ( doppler `` trends '' ) .
these accelerations are regularly revealed in planet searches and point to the existence of unseen stars , brown dwarfs , or giant planets on wide orbits .
high - contrast imaging is a useful tool to diagnose the nature of these companions and , in the case of non - detections , rule out massive objects at wide projected separations ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
when a companion is detected , its minimum mass can be inferred from the host star s acceleration ( @xmath32 ) , the distance to the system ( @xmath33 ) , and the angular separation of the companion ( @xmath23 ) following @xcite and @xcite : @xmath34 the coefficient is 0.0145 when expressed in .
this equation assumes an instantaneous radial velocity slope , but longer baseline coverage or a change in the acceleration ( `` jerk '' ) can provide better constraints on a companion s mass and period ( @xcite ) .
if a significant fraction of the orbit is measured with both astrometry and radial velocities , simultaneous modeling of both data sets can yield a robust dynamical mass measurement .
perhaps the best example of this is from @xcite , who measured the mass and three - dimensional orbit of the brown dwarf companion hr 7672 b , initially discovered by @xcite based on an acceleration from the host star .
many stellar and white dwarf companions have been discovered in this fashion but only a few substellar companions have been found ( table [ tab : trends ] ) . figure [ fig : trends ] shows the population of known companions inducing shallow rv trends on their host stars and which have also been recovered with high resolution ( and often high - contrast ) imaging .
most of these are m dwarfs with masses between @xmath40.10.5 at separations of @xmath410100 au .
this is primarily due to the two competing methods at play : at these old ages , direct imaging is insensitive to low masses and close separations , while small accelerations induced from wide - separation and low - mass companions are difficult to measure even for long - baseline , precision radial velocity planet searches .
the trends program ( e.g. , @xcite ; @xcite ) is the largest survey to combine these two methods and demonstrates the importance of both detections and non - detections to infer the population of planets on moderate orbits out to @xmath420 au ( @xcite ) .
-.5 in -0.7 in -.7 in dynamical masses of planets may eventually be measured by combing radial velocity monitoring of the host star and direct imaging , effectively treating the system like a spatially resolved single - lined spectroscopic binary .
stellar jitter is a limiting factor at very young ages and at older ages the low luminosities of planets generally preclude imaging
. the intermediate ages of moving group members may provide an adequate solution , and at least one ambitious survey by @xcite is currently underway to search for planets and long - term radial velocity trends for this population .
another solution is to image planets in reflected light at optical wavelengths , which requires a space - based telescope and coronagraph like _ wfirst _
( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
similarly , astrometric accelerations can be used to identify and measure the masses of substellar companions when combined with high - contrast imaging .
this will be particularly relevant in the near - future with @xmath22 as thousands of planets are expected to be found from the orbital reflex motion of their host stars ( @xcite ; @xcite ) .
young stars in the field not necessarily associated with coherent moving groups have also been popular targets for high - contrast imaging planet searches .
the advantage of this population is that they are numerous and often reside at closer distances than actual members of young moving groups , but their ages and metallicities are generally highly uncertain so substellar companions uncovered with deep imaging can have a wide range of possible masses ( e.g. , @xcite ) .
gj 504 and @xmath21 and are recent examples of young field stars with faint companions that were initially thought to have planetary masses ( @xcite ; @xcite ) but which follow - up studies showed are probably older ( @xcite ; @xcite ; @xcite ; @xcite ) , implying the companions are likely more massive and reside in the brown dwarf regime .
sirius is another example of a young massive field star extensively targeted with high - contrast imaging ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
this system is particularly noteworthy for possible periodic astrometric perturbations to the orbit of its white dwarf companion sirius b that may be caused by an still - hidden giant planet or brown dwarf ( @xcite ) .
[ sec : masses ] the masses of directly imaged planets are generally highly uncertain , heavily model - dependent , and difficult to independently measure . yet mass is fundamentally important to test models of giant planet formation and empirically calibrate substellar evolutionary models .
this section describes how observables like bolometric luminosity , color , and absolute magnitude coupled with evolutionary models and semi - empirical quantities like age are used to infer the masses of planets .
although no imaged planet has yet had its mass directly measured , there are several promising routes to achieve this which will eventually enable rigorous tests of giant planet cooling models .
-1.2 in -.5 in -0.7 in like white dwarfs and brown dwarfs , giant planets cool over time so evolutionary models along with two physical parameters luminosity , age , effective temperature , or radius are needed to infer a planet s mass . among these ,
luminosity and age are usually better constrained and less reliant on atmospheric models than effective temperature and radius , which can substantially vary with assumptions about cloud properties , chemical composition , and sources of opacity .
below are summaries of the major assumptions ( in roughly descending order ) involved in the inference of planet masses using atmospheric and evolutionary models along with notable advantages , drawbacks , and limitations of various techniques . * * initial conditions and formation pathway . *
the most important assumption is the amount of initial energy and entropy a planet begins with following its formation .
this defines its evolutionary pathway , which is embodied in three broad classes informed by formation mechanisms .
+ hot - start models begin with arbitrarily large radii and oversimplified , idealized initial conditions that generally ignore the effects of accretion and mass assembly . as such
, they represent the most luminous outcome and correspond to the most optimistically ( albeit unrealistically ) low mass estimates .
ironically , hot - start grids are nearly unanimously adopted for estimating the masses of young brown dwarfs and giant planets even though the early evolution in these models is the least reliable .
the most widely used hot - start models for imaged planets are the cond and dusty grids from @xcite and @xcite , @xcite , and @xcite . + cold - start models were made prominent by @xcite and @xcite in the context of direct imaging as an attempt to emulate a more realistic formation scenario for giant planets through core accretion . in this model
, accretion shocks radiate the gravitational potential energy of infalling gas as a giant planet grows . after formation , these planets begin cooling with much lower luminosities and initial entropies compared to the hot - start scenario , taking between @xmath410@xmath35 and @xmath410@xmath36 years to converge with hot - start cooling models depending on the planet mass .
the observational implications of this are severe : planets formed from core accretion may be orders of magnitude less luminous than those produced from cloud fragmentation or disk instability .
while this may offer a diagnostic for the formation route if the mass of a planet is independently measured , it also introduces considerable uncertainty in the more typical case when only an age and luminosity are known .
for example , 51 eri b may be as low as 2 or as high as 12 depending on which cooling model ( hot or cold ) is assumed ( @xcite ) . +
this picture is made even more complicated by large uncertainties in the details of cold - start models .
the treatment of accretion shocks , circumplanetary disks , core mass , and even deuterium burning for the most massive planets can dramatically influence the initial entropy and luminosity evolution of planets ( @xcite ; @xcite ; @xcite ; @xcite ) .
this motivated a class of warm - start models with intermediate initial entropies that probably better reflect dissipative accretion shocks that occur in nature ( @xcite ; @xcite ) .
unfortunately , the relevant details of giant planet assembly are poorly constrained by observations .
there is also likely to be intrinsic scatter in the initial conditions for a given planet which may result in large degeneracies in the planet mass , core mass , and accretion history for young gas giants with the same age and luminosity .
it is quite possible , for instance , that the hr 8799 c , d , and e planets which all share the same age and nearly the same luminosity could have very different masses .
+ -.4 in -1.5 in -0.05 in + -.1 in * * stellar age . * after bolometric luminosity , which is generally uncomplicated to estimate for imaged planets , the age of the host star is the most sensitive parameter on which the mass of an imaged companion depends .
it is also one of the most difficult quantities to accurately determine and usually relies on stellar evolutionary models or empirical calibrations .
recent reviews on this topic include @xcite , @xcite , and @xcite .
figure [ fig : lum_age ] shows the current census of imaged companions near and below the deuterium - burning limit with both age and luminosity measurements ( from table [ tab : planets ] ) .
apart from uncertainties in the formation history of these objects , age uncertainties dominate the error budget for inferring masses .
+ clusters of coeval stars spanning a wide range of masses provide some of the best age constraints but are still dominated by systematic errors .
several star - forming regions and young moving groups in particular have been systematically adjusted to older ages over the past few years , which has propagated to the ages and masses of planets in those associations ( @xcite ; @xcite ; @xcite ; @xcite ) .
the implied hot - start mass for @xmath20 pic b , for example , increases by several jupiter masses ( corresponding to several tens of percent ) assuming the planet s age is @xmath023 myr instead of @xmath012 myr ( @xcite ; although see the next bullet point ) .
+ for young field stars , distant stellar companions can help age - date the entire system .
for example , the age of fomalhaut was recently revised to @xmath4400 myr from @xmath4200 myr in part due to constraints from its wide m dwarf companions ( @xcite ; @xcite ) . ultimately , if the age of a host star is unknown , the significance and interpretation of a faint companion is limited if basic physical properties like its mass are poorly constrained . *
* epoch of planet formation .
* planets take time to form so they are not exactly coeval with their host stars .
their ages may span the stellar age to the stellar age minus @xmath410 myr depending on the timescale for giant planets to assemble .
planets formed via cloud fragmentation or disk instability might be nearly coeval with their host star , but those formed by core accretion are expected to build mass over several myr . while this difference is negligible at intermediate and old ages beyond a few tens of myrs , it can have a large impact on the inferred masses of the youngest planets ( @xmath2820 myr ) .
for example , if the age of the young planetary - mass companion 2m12073932 b is assumed to be coeval with the tw hyrdae association ( @xmath37 = 10 @xmath38 3 myr ) then its hot - start mass is @xmath05 . on the other hand ,
if its formation was delayed by 8 myr ( @xmath37 = 2 myr ) then its mass is only @xmath02.5 .
+ -.6 in -0 in -1.15 in * * atmospheric models . *
atmospheric models can influence the inferred masses of imaged exoplanets in several ways .
they act as surface boundary conditions for evolutionary models and regulate radiative cooling through molecular and continuum opacity sources .
this in turn impacts the luminosity evolution of giant planets , albeit minimally because of the weak dependence on mean opacity ( @xmath39 @xmath40 @xmath41 ; @xcite ; @xcite ) .
even the unrealistic cases of permanently dusty and perpetually condensate - free photospheres do not dramatically affect the luminosity evolution of cooling models or mass determinations using age and bolometric luminosity ( @xcite ; @xcite ) , although more realistic ( `` hybrid '' ) models accounting for the evolution and dissipation of clouds at the l / t transition can influence the shape of cooling curves in slight but significant ways ( @xcite ; @xcite ) .
+ on the other hand , mass determinations in color - magnitude space are highly sensitive to atmospheric models and can result in changes of several tens of percent depending on the specific treatment of atmospheric condensates .
dust reddens spectra and can modify the near - infrared colors and absolute magnitudes of ultracool objects by several magnitudes .
this introduces another source of uncertainty if the spectral shape is poorly constrained , though the difference between dusty and cloud - free models is smaller at longer wavelengths and higher temperatures .
+ one of the most important and unexpected empirical results to emerge from direct imaging has been the realization that young brown dwarfs and massive planets retain photospheric clouds even at low effective temperatures where older , high - gravity brown dwarfs have already transitioned to t dwarfs ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
this is demonstrated in figure [ fig : cmd ] , which shows the location of imaged companions near and below the deuterium - burning limit on the near - infrared color - color diagram . at young ages , warm giant planets
are significantly redder than the field population of brown dwarfs , and several of the most extreme examples have anomalously low absolute magnitudes . for old brown dwarfs , this evolution from dusty ,
co - bearing l dwarfs to cloud - free , methane - dominated t dwarfs takes place over a narrow temperature range ( @xmath412001400 k ) but occurs at a lower ( albeit still poorly constrained ) temperature for young gas giants .
the lack of methane is likely caused by disequilibrium carbon chemistry at low surface gravities as a result of vigorous vertical mixing ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) , while the preservation of photospheric condensates can be explained by a dependency of cloud base pressure and particle size on surface gravity ( @xcite ) .
unfortunately , the dearth of known planets between @xmath4l5t5 is the main limitation to understanding this transition in detail ( figure [ fig : spt_age ] ) .
+ in principle , the mass of a planet can also be inferred by fitting synthetic spectra to the planet s observed spectrum or multi - band photometry .
the mass can then be obtained from best - fitting model as follows : + @xmath42 + here @xmath43 is the surface gravity ( in cm s@xmath44 ) and @xmath45 is the planet s radius . the radius can either be taken from evolutionary models or alternatively from the multiplicative factor that scales the emergent model spectrum to the observed flux - calibrated spectrum ( or photometry ) of the planet .
this scale factor corresponds to the planet s radius over its distance , squared ( @xmath46/@xmath47 ; see @xcite for details ) .
+ clearly the inferred mass is very sensitive to both the surface gravity and the radius . in practice ,
gravity is usually poorly constrained for model fits to brown dwarf and giant planet spectra because its influence on the emergent spectrum is more subtle ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . in addition
, the scale factor strongly depends on the model effective temperature ( @xmath40 @xmath48 ) , which is typically not known to better than @xmath4100 k. altogether , the current level of systematic imperfections present in atmospheric models and observed spectra of exoplanets ( e.g. , @xcite ) mean that masses can not yet be reliably measured from fitting grids of synthetic spectra . *
* deuterium burning history . * as brown dwarfs with masses between about 13 and 75 contract , their core temperatures become hot enough to burn deuterium , though not at sufficient rates to balance surface radiative losses ( e.g. , @xcite ; @xcite).63 can also burn lithium .
this limit changes slightly for non - solar values ( @xcite ) . ]
the onset and timescale of deuterium burning varies primarily with mass but also with metallicity , helium fraction , and initial entropy ( e.g. , @xcite ) ; lower - mass brown dwarfs take longer to initiate deuterium burning than objects near the hydrogen - burning minimum mass .
this additional transient energy source delays the otherwise invariable cooling and causes luminosity tracks to overlap .
thus , objects with the same luminosity and age can differ in mass depending on their deuterium - burning history .
many substellar companions fall in this ambiguous region , complicating mass determinations by up to a factor of @xmath42 ( figure [ fig : lum_age ] and table [ tab : planets ] ) . with a large sample of objects in this region , spectroscopy may ultimately be able to distinguish higher- and lower - mass scenarios through relative surface gravity measurements ( @xcite ) . *
* planet composition .
* the gas and ice giants in the solar system are enriched in heavy elements compared to solar values .
the specific mechanism for this enhancement is still under debate but exoplanets formed via core accretion are expected to show similar compositional and abundance ratio differences compared to their host stars , whereas planets formed through cloud fragmentation or disk instability are probably quite similar to the stars they orbit .
the bulk composition of planets modifies their atmospheric opacities and influences both their emergent spectra and luminosity evolution ( @xcite ) .
a common practice when deriving masses for imaged planets is to assume solar abundances , which is largely dictated by the availability of published atmospheric and evolutionary models .
many of these assumptions can be removed with atmospheric retrieval methods by directly fitting for atomic and molecular abundances ( @xcite ; @xcite ; @xcite ) . *
* additional sources of uncertainty . *
a number of other factors and implicit assumptions can also introduce random and systematic uncertainties in mass derivations .
different methods of psf subtraction can bias photometry if planet self - subtraction or speckle over - subtraction is not properly corrected ( e.g. , @xcite ; @xcite ; @xcite ) .
photometric variability from rapidly changing or rotationally - modulated surface features can introduce uncertainties in relative photometry ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
the host star can also be variable if it has unusually large starspot coverage or if it is very young and happens to have an edge - on disk ( e.g. , twa 30a ; @xcite ) .
+ multiplicity can also bias luminosity measurements .
roughly 20% of brown dwarfs are close binaries with separation distributions peaking near 4.5 au and mass ratios approaching unity ( e.g. , @xcite ; @xcite ; @xcite ) . if some planetary - mass companions form in the same manner as brown dwarfs , and
if the same trends in multiplicity continue into the planetary regime , then a small fraction of planetary - mass companions are probably close , unresolved , equal - mass binaries .
these systems will appear twice as luminous .
+ if atmospheric chemistry or cloud structure varies latitudinally then orientation and viewing angle could be important . for the youngest protoplanets embedded in their host stars circumstellar disks
, accretion streams might dominate over thermal photospheric emission , complicating luminosity measurements and mass estimates ( e.g. , lkca15 b and hd 100546 b ; @xcite ; @xcite ) .
approaches to applying bolometric corrections , measuring partly opaque coronagraphs and neutral density filters , finely interpolating atmospheric and evolutionary model grids , or converting models between filter systems ( for example , @xmath49 to @xmath50 ) may vary . finally , additional energy sources like radioactivity or stellar insolation are assumed to be negligible but could impact the luminosity evolution of some exoplanets .
no imaged planet has yet had its mass measured .
the most robust , model - independent way to do so is through dynamical interactions with other objects . because planets follow mass - luminosity - age relationships , knowledge of all three parameters
are needed to test cooling models .
once a mass is measured , its age ( from the host star ) and bolometric luminosity ( from its distance and spectral energy distribution ) enable precision model tests , although an assumption about energy losses from accretion via hot- , warm- , or cold - start must be made .
nevertheless , if all hot - start models overpredict the luminosities of giant planets , that would suggest that accretion history is indeed an important factor in both planet formation and realistic cooling models . below is a summary of methods to measure substellar masses . *
* dynamical masses . * most close - in ( @xmath3100 au ) planets have shown significant orbital motion since their discoveries ( table [ tab : planets ] ) .
this _ relative _ motion provides a measure of the total mass of the system ( @xmath51+@xmath52 ) .
if stationary background stars can simultaneously be observed with the planet - star pair then _ absolute _ astrometry is possible .
this then gives individual masses for each component ( @xmath51 and @xmath52 separately ) .
unfortunately the long orbital periods and lack of nearby background stars for the present census of imaged planets means this method is currently impractical to measure masses .
+ relative astrometry can also be combined with radial velocities to measure a planet s mass . assuming the visual orbit and total mass are well constrained from imaging
, the mass of the companion can be measured by monitoring the line of sight reflex motion of the host star ( e.g. , @xcite ) .
this treats the system as a single - lined spectroscopic binary , giving the mass function @xmath53/@xmath54 , where @xmath55 is the measured total mass , @xmath19 is the measured inclination , and @xmath56 is the mass of the planet .
if precise radial velocities are not possible for the host star because it has an early spectral type ( with few absorption lines ) or high levels of stellar activity ( rv jitter ) then rv monitoring of the planet can also yield its mass .
this can be achieved by combining adaptive optics imaging and high - resolution near - infrared spectroscopy to spatially separate the star and planet , as has been demonstrated with @xmath20 pic b ( @xcite ) . soon
@xmath22 will produce precise astrometric measurements of the host stars of imaged planets .
together with orbit monitoring though high - contrast imaging , this may offer another way to directly constrain the masses of imaged planets .
+ close substellar binaries offer another approach .
their orbital periods are typically faster and , in rare cases when such binaries themselves orbit a star , the age of the tertiary components can be adopted from the host star .
several brown dwarf - brown dwarf masses have been measured in this fashion : hd 130948 bc ( @xcite ) , gl 417 bc ( @xcite ) , and preliminary masses for @xmath30 indi bab ( @xcite ) .
isolated substellar pairs are also useful for dynamical mass measurements but their ages are generally poorly constrained unless they are members of young clusters or moving groups .
no binaries with both components unambiguously residing in the planetary - mass regime are known , but there is at least one candidate ( wise j014656.66 + 423410.0 ; @xcite ) . *
* keplerian disk rotation .
* dynamical masses for young protoplanets may eventually be possible using alma through keplerian rotation of circumplanetary disks .
this requires resolving faint gas emission lines ( e.g. , co @xmath57=32 or co @xmath57=21 ) from the planet both spatially and spectrally , something that has yet to be achieved for known young planets harboring subdisks ( e.g. , @xcite ; @xcite ) .
although challenging , this type of measurement can act as a detailed probe of the initial conditions of giant planet formation and evolution .
* * stability analysis . *
numerical modeling of planets and their interactions with debris disks , protoplanetary disks , or additional planets offers another way to constrain the mass of an imaged planet .
if their masses are too low , planets will not be able to gravitationally shape dust and planetesimals in a manner consistent with observations .
modeling of the disks and companions orbiting fomalhaut and @xmath20 pic illustrate this approach ; independent constraints on the orbit and masses of these companions can be made by combining spatially - resolved disk structures ( a truncated , offset dust ring encircling fomalhaut and a warped inner disk surrounding @xmath20 pic b ) and observed orbital motion ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
likewise , if a planet s mass is too high then it carves a larger disk gap or may destabilize other planets in the system through mutual interactions .
for example , detailed @xmath58-body simulations of hr 8799 s planets have shown that they must have masses @xmath281020 consistent with giant planet evolutionary models or they would have become dynamically unstable by the age of the host star ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . *
* disk morphology . * large - scale structures in disks
clumps , asymmetries , warps , gaps , rings , truncated edges , spiral arms , and geometric offsets can also be used to indirectly infer the presence of unseen planets and predict their masses and locations ( e.g. , @xcite ; @xcite ; @xcite ) .
this approach relies on assumptions about disk surface density profiles and grain properties , so it is not a completely model - free measurement , but it is potentially sensitive to planet masses as low as a few tens of earth masses ( @xcite ) .
it also enables an immediate mass evaluation without the need for long - term orbit monitoring .
recently , @xcite and @xcite presented a novel approach along these lines to predict the locations and masses widely - separated companions inducing spiral arms on a circumstellar disk ( see also @xcite , @xcite , and @xcite ) .
this may prove to be a valuable way to constrain masses of planetary companions at extremely wide orbital distances .
-.4 in -1 . in -0.3 in
[ sec : surveys ] myriad large high - contrast imaging surveys have been carried out over the past decade .
the most impactful programs are highly focused , carefully designed with well - understood biases , and have meticulously - selected target lists to address specific science questions .
the advantages of large surveys include homogeneous observations , instrument setups , data reduction pipelines , and statistical treatments of the results . below are summaries of the most substantial high - contrast imaging surveys carried out to date with a focus on deep adaptive optics imaging programs that routinely reach planetary masses and employ modern observing and post - processing techniques to suppress speckle noise .
these surveys produced the first wave of discoveries ( figure [ fig : planetgallery ] ) , opening the door to directly characterizing the atmospheres of exoplanets as well as their orbits through astrometric monitoring ( figure [ fig : astrometry ] and appendix [ tab : astrometry ] ) this section follows an historical approach by outlining early ground- and space - based experiments , the first generation of planet - finding instruments and associated surveys , and the next generation of instruments characterized by extreme adaptive optics systems with exceptionally high strehl ratios .
-1 in -1.2 in -2.1 in early high - contrast imaging surveys in search of closely - separated brown dwarf companions and giant planets were conducted with speckle interferometry ( @xcite ) , image stabilizers ( @xcite ) , @xmath26 ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , speckle cameras ( @xcite ) , or newly - commissioned adaptive optics systems from the ground with facility instruments ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
when psf subtraction was performed , it usually entailed roll - subtraction ( for @xmath26 ; e.g. , @xcite ) , self - subtraction with a rotated psf , or reference star subtraction .
some of these pioneering programs are especially noteworthy for their depth and emphasis on statistical results .
@xcite targeted 45 single young a m stars with nicmos in the @xmath59 filter ( 1.6 @xmath6 m ) on board @xmath26 .
two brown dwarfs were uncovered , twa 5 b ( @xcite ) and hr 7372 b ( @xcite ) , as well as gl 577 bc , a tight binary companion near the hydrogen - burning limit .
@xcite used naco at the vlt to obtain deep adaptive optics imaging of 28 young nearby stars .
no substellar companions were found , but the importance of thoroughly reporting survey results is highlighted , even for non - detections , a theme that continues today .
focusing exclusively on young moving group members in @xmath60-band enabled @xcite to reach exceptionally low limiting masses for a sample of 22 stars with naco .
the palomar and keck adaptive optics survey by @xcite is another especially valuable contribution ; they imaged 266 fgk stars and discovered two brown dwarf companions , hd 49197 b ( @xcite ) and hd 203030 b ( @xcite ) , implying a substellar occurrence rate of 3.2@xmath61% .
hd 203030 b was the first young brown dwarf for which signs of a discrepancy between the field spectral type - effective temperature sequence was recognized , now understood as a retention of clouds to lower effective temperatures at low surface gravities .
these groundbreaking surveys helped define the scientific motivation , framework , and early expectations for the first generation planet - finding instruments and larger observing programs .
[ sec : firstgen ] high - contrast imaging is largely driven by advances in instrumentation and speckle suppression . the first wave of instruments specifically designed to image giant planets gave rise to large surveys ( @xmath58@xmath050500 ) targeting mostly young nearby stars .
deep observations in pupil - tracking mode ( angular differential imaging ) have become standardized as a way to distinguish quasi - static speckles from planets ( @xcite ) .
this era is also characterized by the advent of advanced psf subtraction techniques to optimally remove speckles during post - processing .
two especially important algorithms are the locally optimized combination of images ( loci ; @xcite ) , which is based on least - squares minimization of residual speckle noise , and karhunen - love image projection ( klip ; @xcite ) , a computationally - fast method based on principal component analysis .
the introduction of these new methods gave rise to an array of sophisticated data reduction pipelines with additional features aimed at minimizing biases and avoiding both self- and over - subtraction of planet flux in adi and sdi datasets ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
the suite of instrumentation for high - contrast imaging has ballooned over the past 15 years and includes dual - channel imagers , infrared wavefront sensors , non - redundant aperture masking interferometry , adaptive secondary mirrors , integral field units , high - order adaptive optics systems , and specialized coronagraphs ( e.g. , apodized lyot coronagraph , annular groove phase mask coronagraph , vector vortex coronagraph , apodizing phase plate , and four quadrant phase mask ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
many of these have been implemented in the first generation of instruments in part as testbeds for regular use in second - generation systems .
these instruments are reviewed in detail in @xcite , @xcite , @xcite , @xcite , and @xcite .
[ sec : gdps ] this survey ( pi : b biller ) targeted 45 young stars between 20032006 with ages
@xmath28250 myr and distances within 50 pc using simultaneous differential imagers mounted on the vlt and mmt ( @xcite ) .
it was among the first to utilize simultaneous differential imaging to search for cold planets around a large sample of young stars .
the sdi method takes advantage of expected spectral differences between the star , which has a nearly flat continuum , and cool , methanated planets by simultaneously imaging in multiple narrow - band filters across this deep absorption feature at 1.6 @xmath6 m . because speckles radially scale with wavelength while real objects remain stationary , their observations also had some sensitivity to warmer planets without methane ( though it is now clear that the onset of methane occurs at lower temperatures for giant planets than for brown dwarfs ) .
no substellar companions were found , which ruled out a linearly - flat extension of close - in giant planets out to 45 au with high confidence .
[ sec : gdps ] gdps ( pi : d. lafrenire ) was a large high - contrast imaging program at the gemini - north 8.1-m telescope with the niri camera and altair ao system focusing on 85 stars , 16 of which were identified as close multiples ( @xcite ) .
the sample contained a mix of nearby gkm stars within 35 pc comprising then - known or suspected nearby young moving group members , stars with statistically young ages , and several others harboring circumstellar disks .
altogether the ages span 10 myr to @xmath45 gyr .
the observations were taken in adi mode with the @xmath49 filter , and psf subtraction was carried out with the loci algorithm ( @xcite ) .
no substellar companions were discovered , implying an occurrence rate of @xmath323% for @xmath312 planets between 25420 au and @xmath312% for @xmath312 planets between 50295 au at the 95% confidence level .
[ sec : youngaustral ] @xcite carried out a deep @xmath60- and @xmath10-band survey of 54 nearby fgk stars at the mmt with clio .
the mmt adaptive optics system uses a deformable secondary mirror which reduces the thermal background by minimizing the number of optical elements along the light path .
observations were carried out between 20062007 with angular differential imaging .
the image processing pipeline is described in @xcite and @xcite .
the target ages are generally older ( @xmath40.12 gyr ) but the long wavelengths of the observations and proximity of the sample ( @xmath2825 pc ) enabled sensitivity to planetary masses for most of the targets .
one new low - mass stellar companion was discovered and the binary brown dwarf hd 130948 bc was recovered in the survey .
the statistical results are detailed in @xcite ; they find that no more than 50% of sun - like stars host @xmath625 planets between 3094 au and
no more than 15% host @xmath6210 planets between 22100 au at the 90% confidence level .
[ sec : youngaustral ] this program utilized naco at the vlt between 20022007 to target 88 young gkm stars within 100 pc ( @xcite ) .
17 new close multiple systems were uncovered and deep imaging was obtained for 65 single young stars .
observations were taken with a lyot coronagraph in @xmath11 and @xmath63 bands and psf subtraction was performed with azimuthally - averaged subtraction and high - pass filtering .
the most important discovery from this survey was 2m12073932 b , a remarkable 5 @xmath38 2 companion to a 25 brown dwarf in the 10 myr twa moving group ( @xcite ; @xcite ) enabled with infrared wavefront sensing .
the unusually red colors and spectral shape of 2m12073932 b ( @xcite ) have made it the prototype of young dusty l dwarfs , now understood as a cloudy extension of the l dwarf sequence to low temperatures ( @xcite ; @xcite ) .
this system is also unusual from the perspective of brown dwarf demographics ; the mass ratio of @xmath40.2 and separation of @xmath441 au make it an outlier compared to brown dwarf mass ratio and separation distributions in the field ( @xcite ) .
two other substellar companions were discovered in this survey : gsc 08047 - 00232 b ( @xcite ) , also independently found by @xcite , and ab pic b ( @xcite ) , which resides near the deuterium - burning limit .
[ sec : youngdusty ] this vlt / naco survey targeted 59 young nearby afgk stars with ages @xmath28200 myr and distances within 65 pc ( @xcite ) .
most of the sample are members of young moving groups and the majority ( 76% ) were chosen to have mid - infrared excesses , preferentially selected for having debris disks .
observations were carried out in @xmath60-band between 20092012 using angular differential imaging .
four targets in the sample had known substellar companions ( hr 7329 , ab pic , hr 8799 , and @xmath20 pic ) .
no new substellar companions were discovered but eight new visual binaries were resolved . a statistical analysis of af stars between 5320 au and 314 implies a giant planet occurrence rate of 7.4@xmath64% ( 68% confidence level ) .
[ sec : seeds ] the seeds survey ( pi : m. tamura ) was a 125-night program on the 8.2-m subaru telescope targeting about 500 stars to search for giant planets and spatially resolve circumstellar disks ( @xcite ) .
@xcite provide an overview of the observing strategy , target samples , and main results .
observations were carried out with the hiciao camera behind subaru s ao188 adaptive optics system over five years beginning in 2009 .
the sample contained a mixture of young stars in star - forming regions , moving groups , and open clusters ; nearby stars and white dwarfs ; and stars with protoplanetary disks and debris disks .
most of the observations were taken in @xmath11-band in angular differential imaging mode as well as polarimetric differential imaging for young disk - bearing stars .
the adi reduction pipeline is described in @xcite .
three new substellar companions were found in seeds : gj 758 b ( @xcite ) , @xmath21 and b ( @xcite ) , and gj 504 b ( @xcite ) .
the masses of gj 504 b and @xmath21 and b may fall in the planetary regime depending on the ages and metallicities of the system , which are still under debate .
two brown dwarf companions found in the pleiades ( hd 23514 b and hii 1348 b ; @xcite ) had also independently been discovered by other groups .
seeds resolved a remarkable number of protoplanetary and transition disks in polarized light over two dozen in total revealing previously unknown gaps , rings , and spiral structures down to 0@xmath651 with exceptional clarity ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
the statistical results for debris disks are presented in @xcite .
at the 95% confidence level , they find that @xmath31530% of stars host @xmath3110 planets at the gap edge .
@xcite inferred a frequency of 1.03.1% at the 68% confidence level for 570 companions between 10100 au by combining results from the seeds moving group sample ( @xcite ) , the seeds disk sample ( @xcite ) , the seeds pleiades sample ( @xcite ) , gdps ( @xcite ) , and the nici campaign moving group sample ( @xcite ) .
[ sec : nici ] the gemini nici planet - finding campaign ( pi : m. liu ) was a 500-hour survey targeting about 230 young stars of all spectral classes with deep imaging using the near - infrared coronagraphic imager on the gemini - south 8.1-m telescope ( @xcite ) .
nici is an imaging instrument encompassing an adaptive optics system , tapered and partly - translucent lyot coronagraph , and dual - channel camera ( @xcite ) .
campaign observations spanned 20082012 and were carried out in two modes : single - channel @xmath11-band with angular differential imaging , and simultaneous dual - channel ( @xmath49 at 1.578 @xmath6 m and @xmath66 at 1.652 @xmath6 m ) angular and spectral differential imaging to maximize sensitivity to methane - dominated planets . the observing strategy and reduction pipeline
are detailed in @xcite and @xcite , and nici astrometric calibration is discussed in @xcite .
one previously - known brown dwarf companion was resolved into a close binary , hip 79797 bab ( @xcite ) , and three new substellar companions were found : pz tel b , a highly eccentric brown dwarf companion in the @xmath20 pic moving group ( @xcite ) ; cd-35 2722 b , a young mid - l dwarf in the ab dor moving group ( @xcite ) ; and hd 1160 b , a substellar companion orbiting a young massive star ( @xcite ) .
no new planets were discovered but @xmath20 pic b was recovered during the survey and its orbit was shown to be misaligned with the inner and outer disks ( @xcite ; @xcite ) .
two debris disks surrounding hr 4796 a and hd 141569 were also resolved with unprecedented detail ( @xcite ; @xcite ) .
the statistical results are organized in several studies . from a sample of 80 members of young moving groups ,
@xcite measured the frequency of 120 planets between 10150 au to be @xmath3618% at the 95.4% confidence level , depending on which hot - start evolutionary models are adopted .
the high - mass sample of 70 b and a - type stars was described in @xcite ; they found that the frequency of @xmath314 planets between 59460 au is @xmath320% at 95% confidence .
@xcite found that @xmath313% of debris disk stars have @xmath625 planets beyond 80 au at 95% confidence from observations of 57 targets .
[ sec : idps ] idps is an expansive imaging survey carried out at the vlt with naco , keck with nirc2 , gemini - south with nici , and gemini - north with niri targeting @xmath0300 young a m stars ( pi : c. marois ) .
this 14-year survey was mostly carried out in @xmath50 band , though much of the survey comprised a mix of broad- and narrow - band near - infrared filters .
target ages were mostly @xmath28300 myr and encompassed distances from @xmath41080 pc ( galicher et al .
2016 , submitted ) .
the main result from this survey was the discovery of the hr 8799 planets ( @xcite ; @xcite ) .
altogether over 1000 unique point sources were found , most of which were meticulously shown to be background stars from multi - epoch astrometry ( galicher et al .
2016 , submitted ) .
the preliminary analysis of a subset of high - mass a and f stars spanning @xmath01.53.0 was presented in @xcite .
39 new observations in @xmath11 , @xmath50 , and @xmath67 filters were carried out in angular differential imaging mode between 20072012 and were combined with three high - mass targets from the literature .
stellar ages span 8400 myr with distances out to 90 pc and comprise a mix of young moving group members , young field stars , and debris disk hosts .
the subsample of 42 massive stars includes three hosts of substellar companions : hr 8799 , @xmath20 pic , and hr 7329 , a @xmath20 pic moving group member with a wide brown dwarf companion ( @xcite ) . including the detections of planets around hr 8799 and @xmath20 pic , @xcite measure the occurrence rate of 314 planets between 5320 au to be 8.7@xmath68% at 68% confidence .
the complete statistical analysis for the entire sample is presented in galicher et al .
( 2016 , submitted ) .
they merge their own results for 292 stars with the gdps and naco - lp surveys , totaling a combined sample of 356 targets . from this
they infer an occurrence rate of 1.05@xmath69% ( 95% confidence interval ) for 0.514 planets between 20300 au .
they do not find evidence that this frequency depends on stellar host mass .
in addition , 16 of the 59 binaries resolved in idps are new . [
sec : palms ] the palms survey ( pi : b. bowler ) is a deep imaging search for planets and brown dwarfs orbiting low - mass stars ( 0.10.6 ) carried out at keck observatory with nirc2 and subaru telescope with hiciao .
deep coronagraphic observations were acquired for 78 single young m dwarfs in @xmath11- and @xmath70-bands between 20102013 using angular differential imaging .
an additional 27 stars were found to be close binaries .
targets largely originate from @xcite , @xcite , and an additional @xmath71-selected sample ( e. shkolnik et al . , in preparation ) .
most of these lie within 40 pc and have ages between 20620 myr ; about one third of the sample are members of young moving groups .
the observations and psf subtraction pipeline are described in @xcite .
four substellar companions were found in this program : 1rxs j235133.3 + 312720 b ( @xcite ) , gj 3629 b ( @xcite ) , 1rxs j034231.8 + 121622 b ( @xcite ) , and 2mass j15594729 + 4403595 b ( @xcite ) .
1rxs j235133.3 + 312720 b is a particularly useful benchmark brown dwarf because it orbits a member of a young moving group ( ab dor ) and therefore has a well - constrained age ( @xmath0120 myr ) .
the statistical results from the survey are presented in @xcite .
no planets were found , implying an occurrence rate of @xmath310.3% for 113 planets between 10100 au at the 95% confidence level assuming hot - start models and @xmath316.0% assuming cold - start models .
for the most massive planets between 513 , the upper limits are @xmath36.0% and @xmath39.9% for hot- and cold - start cooling models .
the second , parallel phase of the palms survey is an ongoing program targeting a larger sample of @xmath4400 young m dwarfs primarily at keck with shallower contrasts ( bowler et al .
, in prep . ) .
initial discoveries include two substellar companions : 2mass j012250932439505 b , an l - type member of ab dor at the planet / brown dwarf boundary with an unusually red spectrum ( @xcite ; @xcite ) , and 2mass j021558920929121 c ( @xcite ) , a brown dwarf in a close quadruple system which probably belongs to the tuc - hor moving group . [
sec : vltlp ] the naco - lp survey was a large program at the vlt focused on 86 young , bright , primarily fgk stars ( pi : j .- l .
@xmath11-band observations were carried out with naco in adi mode between 20092013 ( @xcite ) .
the target sample is described in detail in @xcite ; stars were chosen to be single , have ages @xmath28200 myr , and lie within 100 pc .
many of these stars were identified as new members of young moving groups .
although no new substellar companions were discovered , an intriguing white dwarf was found orbiting hd 8049 , an ostensibly young k2 star that may instead be much older due to mass exchange with its now evolved companion ( @xcite ) .
new observations of the spatially resolved debris disk around hd 61005 ( `` the moth '' ) were presented by @xcite , and 11 new close binaries were resolved during this program ( @xcite ) .
the statistical analysis of the sample of single stars was performed in @xcite .
based on a subsample of 51 young fgk stars , they found that @xmath315% of sun - like stars host planets with masses @xmath315 between 100500 au and @xmath310% host @xmath3110 planets between 50500 au at the 95% confidence level .
@xcite use these naco - lp null results together with additional deep archival observations to study the companion mass function as it relates to binary star formation and planet formation . from their full sample of 199 sun - like stars
, they find that the results from direct imaging are consistent with the superposition of the planet mass function determined from radial velocity surveys and the stellar companion mass ratio distribution down to 5 , suggesting that many planetary - mass companions uncovered with direct imaging may originate from the tail of the brown dwarf mass distribution instead of being the most massive representatives of the giant planet population .
[ sec : otherfirst ] several smaller , more focused surveys have also been carried out with angular differential imaging : @xcite targeted 8 debris disk hosts with naco using simultaneous differential imaging ; @xcite imaged 38 high - mass stars primarily with naco ; @xcite observed 15 b and a stars with niri ; @xcite presented observations of 16 young m dwarfs in @xmath60-band with naco ; @xcite targeted 16 young afgk stars with naco s four - quadrant phase mask , simultaneous differential imaging , and angular differential imaging ; @xcite used naco s apodizing phase plate coronagraph in @xmath60 band to image six young debris disk hosts with gaps as part of the holey debris disk survey ; and @xcite also used the app at naco to image a sample of 13 a- and f - type main sequence stars in search of planets , uncovering a probable brown dwarf around hd 984 ( @xcite ) .
the lyot project is another survey with important contributions for its early use of coronagraphy behind an extreme adaptive optics system ( @xcite ) .
this survey was carried out at the 3.6-m aeos telescope equipped with a 941-actuator deformable mirror and targeted 86 nearby bright stars using angular differential imaging ( @xcite ) .
the transition to second - generation planet - finding instruments began over the past few years .
this new era is characterized by regular implementation of high - order ( `` extreme '' ) adaptive optics systems with thousands of actuators and exceptionally low residual wavefront errors ; pyramid wavefront sensors providing better sensitivity and higher precision wavefront correction ; strehl ratios approaching ( and often exceeding ) 90% at near - infrared wavelengths ; high - contrast integral field units designed for on - axis observations enabling speckle subtraction and low - resolution spectroscopy ; sensitivity to smaller inner working angles than first - generation instruments ; and advanced coronagraphy . [ sec : leech ] project 1640 is a large ongoing survey ( pi : r. oppenheimer ) and high - contrast imaging instrument with the same name located behind the palm-3000 second - generation adaptive optics system at the palomar observatory 200-inch ( 5.1-meter ) hale telescope .
the instrument itself contains an apodized - pupil lyot coronagraph and integral field unit that samples 32 spectral channels across @xmath72 , @xmath57 , and @xmath11 bands ( @xcite ) , producing a low - resolution spectrum to broadly characterize the physical properties of faint companions ( @xcite ; @xcite ) .
the survey consists of two phases , the first ( now concluded ) with the original palomar adaptive optics system ( palm-241 ; @xcite ) and a second ongoing three - year program focusing on nearby massive stars with the upgraded palm-3000 adaptive optics system ( @xcite ) .
the data reduction pipeline is described in @xcite and a detailed treatment of speckle subtraction , precision astrometry , and robust spectrophotometry can be found in @xcite , @xcite , @xcite , @xcite , and @xcite .
results from the project 1640 survey include several discoveries of faint stellar companions to massive a stars ( @xcite ; @xcite ) and follow - up astrometric and spectral characterization of known substellar companions ( @xcite ; @xcite ) .
in addition , @xcite and @xcite presented detailed spectroscopic and astrometric analysis of the hr 8799 planets and found intriguing evidence for mutually dissimilar spectral properties and signs of non - coplanar orbits .
[ sec : leech ] leech is an ongoing @xmath470-night high - contrast imaging program ( pi : a. skemer ) at the twin 8.4-m large binocular telescope .
survey observations began in 2013 and are carried out in angular differential imaging mode in @xmath60-band with lmircam utilizing deformable secondary mirrors to maximize sensitivity at mid - ir wavelengths by limiting thermal emissivity from warm optics ( @xcite ) .
the target sample focuses on intermediate - age stars @xmath31 gyr including members of the @xmath4500 myr ursa majoris moving group , massive ba stars , and nearby young fgk stars .
in addition to searching for new companions , this survey is also characterizing known planets using the unique mid - ir instrumentation , sensitivity , and filter suite at the lbt .
@xcite refined the orbits of the hr 8799 planets and found them to be consistent with 8:4:2:1 mean motion resonances .
@xcite observed gj 504 b in three narrow - band filters spanning 3.74.0 @xmath6 m .
model fits indicate an exceptionally low effective temperature of @xmath0540 k and enhanced metallicity , possibly pointing to an origin through core accretion .
additionally , @xcite presented leech observations and dynamical mass measurements the ursa majoris binary no uma .
recently the integral field unit arizona lenslets for exoplanet spectroscopy ( ales ; @xcite ) was installed inside lmircam and will enable integral - field spectroscopy of planets between 35 @xmath6 m for the first time .
[ sec : gpies ] gpies is an ongoing 890-hour , 600-star survey to image extrasolar giant planets and debris disks with the gemini planet imager at gemini - south ( pi : b. macintosh ) .
gpi is expressly built to image planets at small inner working angles ; its high - order adaptive optics system incorporates an apodized pupil lyot coronagraph , integral field spectrograph , imaging polarimeter , and ( imminent ) non - redundant masking capabilities ( @xcite ) .
survey observations targeting young nearby stars began in 2014 and will span three years .
@xcite presented the discovery of 51 eri b , the first exoplanet found in gpies and the lowest - mass planet imaged in thermal emission to date . this remarkable young , methanated t dwarf has a contrast of 14.5 mag in @xmath11-band at a separation of 0@xmath6545 , which translates into a mass of only 2 at 13 au assuming hot - start cooling models .
it is also the only imaged planet consistent with the most pessimistic cold - start evolutionary models , in which case its mass may be as high as 12 .
@xcite obtained follow - up observations with gpi and showed that 51 eri b shares a common proper motion with its host and exhibits slight ( but significant ) orbital motion .
other initial results from this survey include astrometry and a refined orbit for @xmath20 pic b ( @xcite ) , as well as resolved imaging of the debris disks around hd 106906 ( @xcite ) and hd 131835 ( @xcite ) .
[ sec : sphere ] sphere ( spectro - polarimetric high - contrast exoplanet research ) is an extreme adaptive optics system ( saxo ) and versatile instrument for high - contrast imaging and spectroscopy at the vlt with a broad range of capabilities ( @xcite ) .
irdis ( @xcite ) offers classical and dual - band imaging ( @xcite ) , dual - polarization imaging ( @xcite ) , and long - slit spectroscopy ( @xcite ) ; ifs provides low - resolution ( @xmath45@xmath43050 ) integral field spectroscopy spanning 0.951.65 @xmath6 m ( @xcite ) ; and zimpol enables diffraction - limited imaging and polarimetry in the optical ( @xcite ) . as part of a guaranteed time observing program , the sphere gto team is carrying out a large ongoing survey ( pi : j .- l .
beuzit ) with a range of science goals .
about 200 nights of this time are devoted to a deep near - infrared imaging survey ( shine : sphere infrared survey ) to search for and characterize exoplanets around 400600 stars , while the remaining @xmath460 nights will be used for a broad range of science including spatially - resolved observations of circumstellar disks and optical imaging of planets .
initial results from the shine survey include observations of the brown dwarf gj 758 b ( @xcite ) , pz tel b and hd 1160 b ( @xcite ) , hd 106906 ( @xcite ) , and the hr 8799 planets ( @xcite ; @xcite ) .
other results have focused on resolved imaging of the debris disk surrounding hd 61005 , which may be a product of a recent planetesimal collision ( @xcite ) , and hd 135344 b , host of a transition disk with striking spiral arm structure ( @xcite ) .
@xcite uncovered intriguing and temporally evolving features in au mic s debris disk .
in addition , @xcite presented deep irdis near - infrared images and visible zimpol polarimetric observations of hd 100546 revealing a complex disk environment with considerable structure and resolved @xmath50-band emission at the location of the candidate protoplanet hd 100546 b. [ sec : sphere ] a number of other novel instruments and forthcoming surveys bear highlighting .
magao ( pi : l. close ) at the magellan 6.5-m clay telescope is a versatile adaptive optics system consisting of a 585-actuator adaptive secondary mirror , pyramid wavefront sensor , and two science cameras offering simultaneous diffraction - limited imaging spanning the visible ( 0.61.05 @xmath6 m ) with visao and near - infrared ( 15.3 @xmath6 m ) with clio2 ( @xcite ; @xcite ; @xcite ) .
strehl ratios of @xmath42030% in the optical are opening up new science fronts including deep red - optical observations of exoplanets ( @xcite ; @xcite ) , characterization of accreting protoplanets in h@xmath16 ( @xcite ) , and high spatial resolution imaging down to @xmath420 mas ( @xcite ) .
vector apodizing phase plate coronagraphs were recently installed in magao and other upgrades such as an optical integral field unit are possible in the future .
subaru coronagraphic extreme adaptive optics ( scexao ; pi : o. guyon ) is being built for the subaru telescope and is the newest extreme adaptive optics system on a large telescope .
a detailed description of all facets of this instrument is described in @xcite . in short ,
a pyramid wavefront sensor is coupled with a 2000-element deformable mirror to produce strehl ratios in excess of 90% .
the instrument is particularly flexible , allowing for a variety of setups and instrument subcomponents including speckle nulling to suppress static and slowly changing speckles ( @xcite ) , a near - infrared science camera ( currently hiciao ) , sub - diffraction - limited interferometric science in the visible with vampires ( @xcite ) and first ( @xcite ) , high - contrast integral field spectroscopy ( @xcite ) , and coronagraphy with phase - induced amplitude apodization ( @xcite ) and vector vortex coronagraphs ( @xcite ) .
-.8 in -.1 in -0.7 in -.2 in -.6 in -0.2 in [ sec : occurrencerate ] the frequency and mass - period distribution of planets spanning various orbital distances , stellar host masses , and system ages provides valuable clues about the dominant processes shaping the formation and evolution of planetary systems .
these measurements are best addressed with large samples and uniform statistical analyses .
@xcite carried out the first such large - scale study based on adaptive optics imaging surveys from @xcite and @xcite . from their sample of 60 unique stars they found an upper limit of 20% for @xmath314 planets between 20100 au at the 95% confidence level .
this was expanded to 118 targets in @xcite by including the gdps survey of @xcite , resulting in the same upper limit and planet mass regime but for a broader range of separations of 8911 au at 68% confidence . @xcite and
@xcite combined their own observations of high - mass stars with previous surveys and measured occurrence rates of 8.7@xmath68% ( for 314 planets between 5320 au ) and 16.1@xmath73% ( for 113 planets between 11000 au ) , respectively .
@xcite incorporated the seeds , gdps , and the nici moving group surveys and found a frequency of 1.03.1% for 570 companions between 10100 au .
recently , galicher et al . ( 2016 , submitted ) combined results from idps , gdps , and the naco survey of young nearby austral stars and found an occurrence rate of 1.05@xmath69% for 0.514 companions between 20300 au based on a sample of 356 unique stars . breaking this into stellar mass bins
did not reveal any signs of a trend with stellar host mass .
-1.2 in -.2 in -.6 in here i reexamine the occurrence rate of giant planets with a meta - analysis of the largest and deepest high - contrast imaging surveys .
696 contrast curves are assembled from the literature from the programs outlined in section [ sec : firstgen ] . for stars with more than one observation , the deeper contrast curve at 1@xmath29
is chosen .
targets with stellar companions within 100 au are removed from the sample because binaries can both inhibit planet formation and dynamically disturb planetary orbits .
most candidate planets uncovered during these surveys are rejected as background stars from second epoch observations , but some candidates are either not recovered or are newly revealed in follow - up data . because of finite telescope allocation , some of these candidates remain untested for common proper motion .
these ambiguous candidates can not be ignored in a statistical analysis because one ( or more ) could be indeed be bound . in these cases ,
contrast curves are individually truncated one standard deviation above the brightest candidate .
ages are taken from the literature except for members of young moving groups , for which the most recent ( and systematically older ) ages of young moving groups from @xcite are adopted .
most ages in the sample are less than 300 myr and within 100 pc .
altogether this leaves 384 unique stars spanning b2m6 spectral types : 76 from @xcite , 72 from @xcite , 61 from @xcite , 54 from @xcite , 45 from @xcite , 30 from @xcite , 25 from @xcite , 14 from @xcite , and 7 from @xcite .
sensitivity maps and planet occurrence rates are derived following @xcite . for a given planet mass and semimajor axis ,
a population of artificial planets on random circular orbits are generated in a monte carlo fashion and converted into apparent magnitudes and separations using cond hot - start evolutionary models from @xcite , the age of the host star , and the distance to the system , including uncertainties in age and distance .
these are compared with the measured contrast curve to infer the fractional sensitivity at each grid point spanning 30 logarithmically - uniform bins in mass and separation between 11000 au and 0.5100 .
when available , fractional field of view coverage is taken into account . contrasts measured in @xmath49 filters are converted to @xmath11-band using an empirical color - spectral type relationship based on synthetic colors of ultracool dwarfs from the spex prism library ( @xcite ) as well as the spectral type - effective temperature sequence from @xcite and @xmath74 filter profiles yields the following relation : @xmath49@xmath74 = @xmath75@xmath76spt@xmath77 , where @xmath78=0.03913178 , @xmath79=0.008678245 , @xmath80=0.001542768 , @xmath81=0.0001033761 , @xmath82=2.902588@xmath8310@xmath84 , and spt is the numerical near - infrared spectral type ( m0=1.0 , l0=10.0 , t0=20.0 ) .
this relation is valid from m3t8 and the rms of the fit is 0.025 mag . @xcite
provide an empirical @xmath85(spt ) relationship , but the inverse spt(@xmath85 ) is necessary for this filter conversion at a given mass and age . refitting the same data from golimowski et al .
yields the following : spt = @xmath75@xmath76@xmath85@xmath77 , where @xmath78=36.56779 , @xmath79=0.004666549 , @xmath80=9.872890@xmath8310@xmath84 , @xmath81=4.108142@xmath8310@xmath86 , @xmath82= 4.854263@xmath8310@xmath87 .
this is valid for 700 k @xmath3 @xmath85 @xmath3 3900 k , the rms is 1.89 mag , and spt is the same numerical near - infrared spectral type as above . ] .
the mean sensitivity maps for all 384 targets and separate bins containing ba stars ( 110 targets ) , fgk stars ( 155 targets ) , and m dwarfs ( 118 targets ) are shown in figure [ fig : meansensitivity ] .
in general , surveys of high - mass stars probe higher planet masses than deep imaging around m dwarfs owing to differences in the host stars intrinsic luminosities .
the most sensitive region for all stars is between @xmath430300 au , with less coverage at extremely wide separations because of limited fields of view and at small separations in contrast - limited regimes .
the occurrence rate of giant planets for all targets and for each stellar mass bin are listed in table [ tab : gpfreq ] , which assumes logarithmically - uniform distributions in mass and separation ( see section 6.5 of @xcite for details ) . the mode and 68.3% minimum credible interval ( also known as the highest posterior density interval ) of the planet frequency probability distribution
are reported .
two massive stars in the sample host planets that were either discovered or successfully recovered in these surveys : @xmath20 pic , with a planet at 9 au , and hr 8799 , with planets spanning 1570 au .
hr 8799 is treated as a single detection .
the most precise occurrence rate measurement is between 513 and 30300 au . over these ranges ,
the frequency of planets orbiting ba , fgk , and m stars is 2.8@xmath2% , @xmath34.1% , and @xmath33.9% , respectively ( figure [ fig : plfrequency ] ) . here
upper limits are 95% confidence intervals .
although there are hints of a higher giant planet occurrence rate around massive stars analogous to the well - established correlation at small separations ( @xcite ; @xcite ; @xcite ; @xcite ) , this trend is not yet statistically significant at wide orbital distances and requires larger sample sizes in each stellar mass bin to unambiguously test this correlation .
marginalizing over stellar host mass , the overall giant planet occurrence rate for the full sample of 384 stars is 0.6@xmath1% , which happens to be comparable to the frequency of hot jupiters around fgk stars in the field ( 1.2 @xmath38 0.4% ; @xcite ) and in the @xmath88 sample ( 0.5 @xmath38 0.1% ; @xcite ) .
however , compared to the high occurrence rate of giant planets ( 0.310 ) with orbital periods out to 2000 days ( @xmath410% ; @xcite ) , massive gas giants are clearly quite rare at wide orbital distances .
[ sec : dbl ] direct imaging has shown that planetary - mass companions exist at unexpectedly wide separations but the provenance of these objects remains elusive .
there is substantial evidence that the tail - end of the star formation process can produce objects extending from low - mass stars at the hydrogen burning limit ( @xmath075 ) to brown dwarfs at the opacity limit for fragmentation ( @xmath0510 ) , which corresponds to the minimum mass of a pressure - supported fragment during the collapse of a molecular cloud core ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . indeed ,
isolated objects with inferred masses below 10 have been found in a range of contexts over the past decade : in star - forming regions ( @xcite ; @xcite ; @xcite ; @xcite ) , among closer young stellar associations ( @xcite ; @xcite ; @xcite ; @xcite ) , and at much older ages as y dwarfs in the field ( @xcite ; @xcite ; @xcite ) . similarly ,
several systems with _ companions _ below @xmath010 are difficult to explain with any formation scenario other than cloud fragmentation : 2m12073932 ab is a @xmath425 brown dwarf with a @xmath45 companion at an orbital distance of 40 au ( @xcite ) and 2m0441 + 2301 aabbab is a quadruple system comprising a low - mass star , two brown dwarfs , and a 10 object in a hierarchical and distinctly non - planetary configuration ( @xcite ) . from the radial velocity perspective , the distribution of gas giant minimum masses is generally well - fit with a decaying power law ( @xcite ; @xcite ; @xcite ) or exponential function ( @xcite ) that tapers off beyond @xmath410 .
this is evident in figure [ fig : mass_sma_exoplanets ] , although inhomogeneous radial velocity detection biases which exclude lower - mass planets at wide separations are not taken into account .
the dominant formation channel for this population of close - in giant planets is thought to be core accretion plus gas capture , in which growing cores reach a critical mass and undergo runaway gas accretion ( e.g. , @xcite ) .
the totality of evidence indicates that the decreasing brown dwarf companion mass function almost certainly overlaps with the the rising giant planet mass function in the 520 mass range .
no strict mass cutoff can therefore unambiguously divide giant planets from brown dwarfs , and many of the imaged companions below 13 listed in table [ tab : planets ] probably originate from the dwindling brown dwarf companion mass function .
another approach to separate these populations is to consider formation channel : planets originate in disks while brown dwarfs form like stars from the gravitational collapse of molecular cloud cores .
however , not only are the relic signatures of formation difficult to discern for individual discoveries , but objects spanning the planetary up to the stellar mass regimes may also form in large toomre - unstable circumstellar disks at separations of tens to hundreds of au ( e.g. , @xcite ; @xcite ) .
any binary narrative based on origin in a disk versus a cloud core is therefore also problematic .
furthermore , both giant planets and brown dwarf companions may migrate , dynamically scatter , or undergo periodic kozai - lidov orbital oscillations if a third body is present , further mixing these populations and complicating the interpretation of very low - mass companions uncovered with direct imaging .
the deuterium - burning limit at @xmath013 is generally acknowledged as a nebulous , imperfect , and ultimately artificial division between brown dwarfs and giant planets .
moreover , this boundary is not fixed and may depend on planet composition , core mass , and accretion history ( @xcite ; @xcite ; @xcite ) .
uncertainties in planet luminosities , evolutionary histories , metallicities , and ages can also produce large systematic errors in inferred planet masses ( see section [ sec : masses ] ) , rendering inconsequential any sharp boundary set by mass .
however , despite these shortcomings , this border lies in the planet / brown dwarf mass valley " and may still serve as a pragmatic ( if flawed ) qualitative division between two populations formed _
predominantly _ with their host stars and _ predominantly _ in protoplanetary disks .
observational tests of formation routes will eventually provide the necessary tools to understand the relationship between these populations .
this can be carried out at an individual level with environmental clues such as coplanarity of multi - planet systems or orbital alignment within a debris disk ; enhanced metallicities or abundance ratios relative to host stars ( @xcite ) ; or overall system orbital architecture .
similarly , the statistical properties of brown dwarfs and giant planets can be used to identify dominant formation channels : the separation distribution of objects formed through cloud fragmentation should resemble that of binary stars ; disk instability and core accretion may result in a bimodal period distribution for giant planets ( @xcite ) ; planet scattering to wide orbits should produce a rising mass function at low planet masses as opposed to a truncated mass distribution at the fragmentation limit for cloud fragmentation and disk instability ; and the companion mass function and mass ratio distribution are expected to smoothly extend from low - mass stars down to the fragmentation limit if a common formation channel in at play ( @xcite ; @xcite ) . testing these scenarios will require much larger sample sizes given the low occurrence rates uncovered in direct imaging surveys .
[ sec : intro ] high - contrast imaging is still in its nascence .
radial velocity , transit , and microlensing surveys have unambiguously demonstrated that giant planets are much rarer than super - earths and rocky planets at separations @xmath2810 au . in that light ,
the discovery of truly massive planets at tens , hundreds , and even thousands of au with direct imaging is fortuitous , even if the overall occurrence rate of this population is quite low .
each detection technique has produced many micro paradigm shifts over the past twenty years that disrupt and rearrange perceptions about the demographics and architectures of planetary systems .
hot jupiters , correlations with stellar mass and metallicity , the ubiquity of super - earths , compact systems of small planets , resonant configurations , orbital misalignments , the prevalence of habitable - zone earth - sized planets , circumbinary planets , and featureless clouds and hazes are an incomplete inventory within just a few au ( e.g. , @xcite ) .
the most important themes to emerge from direct imaging are that massive planets exist but are uncommon at wide separations ( @xmath3110 au ) , and at young ages the low - gravity atmospheres of giant planets do not resemble those of older , similar - temperature brown dwarfs .
there are many clear directions forward in this field .
deeper contrasts and smaller inner working angles will probe richer portions of planetary mass- and separation distributions .
thirty meter - class telescopes with extreme adaptive optics systems will regularly probe sub - jovian masses at separations down to 5 au .
this next generation will uncover more planets and enable a complete mapping of the evolution of giant planet atmospheres over time .
other fertile avenues for high - contrast imaging include precise measurements of atmospheric composition ( @xcite ; @xcite ) , doppler imaging ( @xcite ; @xcite ) , photometric monitoring to map variability of rotationally - modulated features ( e.g. , @xcite ) , synergy with other detection methods ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) , advances in stellar age - dating at the individual and population levels , merging high - contrast imaging with high - resolution spectroscopy ( @xcite ; @xcite ) , surveying the companion mass function to sub - jovian masses , polarimetric observations of photospheric clouds ( e.g. , @xcite ; @xcite ) , statistical correlations with stellar host properties , probing the earliest stages of protoplanet assembly ( @xcite ; @xcite ) , astrometric orbit monitoring and constraints on dynamical histories , and robust dynamical mass measurements to test evolutionary models and probe initial conditions ( e.g. , @xcite ; @xcite ) .
high - contrast imaging has a promising future and will play an ever - growing role in investigating the architecture , atmospheres , and origin of exoplanets .
it is a pleasure to thank the referee , rebecca oppenheimer , as well as lynne hillenbrand , dimitri mawet , sasha hinkley , and trent dupuy for their thoughtful comments and constructive feedback on this review .
michael liu , arthur vigan , christian marois , motohide tamura , gal chauvin , andy skemer , adam kraus , and bruce macintosh contributed helpful suggestions on past and ongoing imaging surveys .
bruce macintosh , eric nielsen , andy skemer , and raphal galicher kindly provided images for figure [ fig : planetgallery ] .
trent dupuy generously shared his compilation of late - t and y dwarfs used in figure [ fig : cmd ] .
this research has made use of the exoplanet orbit database , the exoplanet data explorer at exoplanets.org , and the spex prism spectral libraries maintained by adam burgasser .
nasa s astrophysics data system bibliographic services together with the vizier catalogue access tool and simbad database operated at cds , strasbourg , france , were invaluable resources for this work .
natexlab#1#1 absil , o. , & mawet , d. 2009 , astron astrophys rev , 18 , 317 absil , o. , le bouquin , j .- b . , berger , j .-
, et al . 2011 , a&a , 535 , a68 absil , o. , milli , j. , mawet , d. , et al . 2013 , a&a , 559 , l12 ahmic , m. , jayawardhana , r. , brandeker , a. , et al .
2007 , the astrophysical journal , 671 , 2074 allard , f. , homeier , d. , & freytag , b. 2012 , philosophical transactions of the royal society a : mathematical , physical and engineering sciences , 370 , 2765 allers , k. n. , & liu , m. c. 2013 , the astrophysical journal , 772 , 79 allers , k. n. , liu , m. c. , dupuy , t. j. , & cushing , m. c. 2010 , the astrophysical journal , 715 , 561 amara , a. , & quanz , s. p. 2012
, monthly notices ras , 427 , 948 anderson , e. , & francis , c. 2012 , astronomy letters , 38 , 331 andrews , s. m. 2015 , publications of the astronomical society of the andrews , s. m. , rosenfeld , k. a. , kraus , a. l. , & wilner , d. j. 2013 , apj , 771 , 129 apai , d. , janson , m. , moro - martn , a. , et al .
2008 , the astrophysical journal , 672 , 1196 apai , d. , kasper , m. , skemer , a. , et al .
2016 , apj , 820 , 1 artigau , e. , gagn , j. , faherty , j. , et al .
2015 , apj , 806 , 1 aumann , h. h. 1985 , publ .
, 97 , 885 aumann , h. h. , beichman , c. a. , gillet , f. c. , et al .
1984 , astrophys .
j. , 278 , l23 backman , d. , marengo , m. , & stapelfeldt , k. 2009 , the astrophysical bailey , j. 2014 , publications of the astronomical society of australia , 31 , e043 bailey , v. , meshkat , t. , reiter , m. , et al .
2014 , the astrophysical journal , 780 , l4 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 2002 , a&a , 382 , 563 baraffe , i. , chabrier , g. , barman , t. s. , allard , f. , & hauschildt , p. h. 2003 , a&a , 402 , 701 baraffe , i. , homeier , d. , allard , f. , & chabrier , g. 2015 , arxiv , 1503.04107v1 barenfeld , s. a. , bubar , e. j. , mamajek , e. e. , & young , p. a. 2013 , the astrophysical journal , 766 , 6 barman , t. s. , konopacky , q. m. , macintosh , b. , & marois , c. 2015 , apj , 804 , 1 barman , t. s. , macintosh , b. , konopacky , q. m. , & marois , c. 2011 , the astrophysical journal , 733 , 65 . 2011 , the astrophysical journal , 735 , l39 basri , g. , & brown , m. e. 2006 , annual review of earth and planetary sciences , 34 , 193 bate , m. r. 2009 , monthly notices ras , 392 , 590 bate , m. r. , bonnell , i. a. , & bromm , v. 2002 , monthly notices ras , 332 , l65 beichman , c. , gelino , c. r. , kirkpatrick , j. d. , et al .
2013 , apj , 764 , 101 . 2014 , apj , 783 , 68 bell , c. p. m. , mamajek , e. e. , & naylor , t. 2015 , monthly notices ras , 454 , 593 benedict , g. f. , mcarthur , b. e. , gatewood , g. , et al .
2006 , the astronomical journal , 132 , 2206 benest , d. , & duvent , j. l. 1995 , a&a , 299 , 621 bergfors , c. , brandner , w. , janson , m. , khler , r. , & henning , t. 2011 , a&a , 528 , a134 beust , h. , augereau , j .- c .
, bonsor , a. , et al .
2014 , a&a , 561 , a43 beuzit , j .- l . , mouillet , d. , oppenheimer , b. r. , & monnier , j. d. 2007 , protostars and planets v , 717 beuzit , j .- l . ,
sgransan , d. , forveille , t. , et al .
2004 , a&a , 425 , 997 beuzit , j .- l . , feldt , m. , dohlen , k. , et al .
2008 , 7014 , 701418 biller , b. , allers , k. , liu , m. , close , l. m. , & dupuy , t. 2011 , the astrophysical journal , 730 , 39 biller , b. , artigau , . , wahhaj , z. , et al .
2008 , proc .
spie , 7015 , 70156q biller , b. , lacour , s. , juhsz , a. , et al .
2012 , the astrophysical journal letters , 753 , l38 biller , b. a. , close , l. m. , masciadri , e. , et al . 2007 , the astrophysical journal supplement series , 173 , 143 biller , b. a. , liu , m. c. , wahhaj , z. , et al . 2010 , the astrophysical journal , 720 , l82 . 2013 , apj , 777 , 160 biller , b. a. , males , j. , rodigas , t. , et al . 2014 , the astrophysical journal letters , 792 , l22 biller , b. a. , vos , j. , bonavita , m. , et al .
2015 , the astrophysical journal letters , 813 , 1 biller , b. a. , liu , m. c. , rice , k. , et al .
2015 , monthly notices ras , 450 , 4446 binks , a. s. , & jeffries , r. d. 2014 , mnras letters , 438 , l11 binks , a. s. , jeffries , r. d. , & maxted , p. f. l. 2015 , monthly notices ras , 452 , 173 boccaletti , a. , lagrange , a .- m . , bonnefoy , m. , galicher , r. , & chauvin , g. 2013 , a&a , 551 , l14 boccaletti , a. , thalmann , c. , lagrange , a .- m . , et al
2015 , nature , 526 , 230 bodenheimer , p. , dangelo , g. , lissauer , j. j. , fortney , j. j. , & saumon , d. 2013 , the astrophysical journal , 770 , 120 boley , a. c. 2009 , apjl , 695 , l53 boley , a. c. , payne , m. j. , corder , s. , et al .
2012 , the astrophysical journal , 750 , l21 bond , h. e. , gilliland , r. l. , schaefer , g. h. , et al .
2015 , apj , 813 , 1 bonnefoy , m. , chauvin , g. , rojo , p. , et al .
2010 , a&a , 512 , a52 bonnefoy , m. , lagrange , a .- m . , boccaletti , a. , et al .
2011 , a&a , 1 bonnefoy , m. , boccaletti , a. , lagrange , a .- m .
, et al . 2013 , a&a , 555 , a107 bonnefoy , m. , currie , t. , marleau , g. d. , et al . 2014 , a&a , 562 , a111 bonnefoy , m. , marleau , g. d. , galicher , r. , et al .
2014 , a&a , 567 , l9 bonnefoy , m. , zurlo , a. , baudino , j. l. , et al .
2016 , a&a , 587 , a58 bonnet - bidaud , j. m. , & pantin , e. 2008 , a&a , 489 , 651 boss , a. p. 2001 , the astrophysical journal letters , 551 , l167 bowler , b. p. , andrews , s. m. , kraus , a. l. , et al .
2015 , the astrophysical journal letters , 805 , 1 bowler , b. p. , & hillenbrand , l. a. 2015 , the astrophysical journal letters , 811 , l30 bowler , b. p. , liu , m. c. , dupuy , t. j. , & cushing , m. c. 2010 , the astrophysical journal , 723 , 850 bowler , b. p. , liu , m. c. , kraus , a. l. , & mann , a. w. 2014 , apj , 784 , 65 bowler , b. p. , liu , m. c. , kraus , a. l. , mann , a. w. , & ireland , m. j. 2011 , the astrophysical journal , 743 , 148 bowler , b. p. , liu , m. c. , shkolnik , e. l. , & dupuy , t. j. 2013 , the astrophysical journal , 774 , 55 bowler , b. p. , liu , m. c. , shkolnik , e. l. , et al .
2012 , the astrophysical journal , 753 , 142 bowler , b. p. , liu , m. c. , shkolnik , e. l. , & tamura , m. 2012 , the astrophysical journal , 756 , 69 . 2015 , the astrophysical journal supplement series , 216 , 7 bowler , b. p. , johnson , j. a. , marcy , g. w. , et al .
2010 , the astrophysical journal , 709 , 396 bowler , b. p. , shkolnik , e. l. , liu , m. c. , et al .
2015 , the astrophysical journal , 806 , 62 brandner , w. , zinnecker , h. , alcal , j. m. , et al .
2000 , aj , 120 , 950 brandt , t. d. , mcelwain , m. w. , turner , e. l. , et al .
2013 , the astrophysical journal , 764 , 183 . 2014 , apj , 794 , 159 brandt , t. d. , mcelwain , m. w. , janson , m. , et al .
2014 , in proc .
spie , 9148 , 1 , 91849 brandt , t. d. , kuzuhara , m. , mcelwain , m. w. , et al .
2014 , apj , 786 , 1 brown , r. a. 2015 , apj , 805 , 1 bryan , m. l. , knutson , h. a. , howard , a. w. , et al .
2016 , apj , 821 , 1 buenzli , e. , thalmann , c. , vigan , a. , et al .
2010 , a&a , 524 , l1 burgasser , a. j. 2014 , arxiv , 1406.4887v1 burgasser , a. j. , reid , i. n. , siegler , n. , et al .
2007 , protostars and planets v , 427 burgasser , a. j. , simcoe , r. a. , bochanski , j. j. , et al .
2010 , the astrophysical journal , 725 , 1405 burningham , b. , leggett , s. k. , homeier , d. , et al .
2011 , monthly notices ras , 414 , 3590 burrows , a. , hubbard , w. b. , lunine , j. i. , & liebert , j. 2001 , reviews of modern physics , 73 , 719 burrows , a. , & liebert , j. 1993 , reviews of modern physics , 65 , 301 burrows , a. , marley , m. , hubbard , w. b. , et al . 1997 , astrophysical journal , 491 , 856 butler , r. p. , wright , j. t. , marcy , g. w. , et al . 2006 , the astrophysical journal , 646 , 505 caceres , c. , hardy , a. , schreiber , m. r. , et al
. 2015 , the astrophysical journal letters , 806 , 1 cantalloube , f. , mouillet , d. , mugnier , l. m. , et al .
2015 , a&a , 582 , a89 cardoso , c. v. , mccaughrean , m. j. , king , r. r. , et al .
2009 , in aip conf .
proc 1094 , 15th cambridge workshop on coolstars , stellar systems and the sun , ed .
e. stempels ( melville , ny : aip ) , , 509 carson , j. , thalmann , c. , janson , m. , et al .
2013 , the astrophysical journal , 763 , l32 carson , j. c. , eikenberry , s. s. , brandl , b. r. , wilson , j. c. , & hayward , t. l. 2005 , the astronomical journal , 130 , 1212 carson , j. c. , eikenberry , s. s. , smith , j. j. , & cordes , j. m. 2006 , the astronomical journal , 132 , 1146 chabrier , g. 2001 , the astrophysical journal , 554 , 1274 chabrier , g. , baraffe , i. , selsis , f. , et al .
2007 , protostars and planets v , 623 chabrier , g. , johansen , a. , janson , m. , & rafikov , r. 2014 , in protostars and planets vi , henrik beuther , ralf s. klessen , cornelis p. dullemond , and thomas henning ( eds . ) , university of arizona press , tucson , 914 pp .
, p.619 - 642 chauvin , g. , beust , h. , lagrange , a .- m . , & eggenberger , a. 2011 , a&a , 528 , a8 chauvin , g. , lagrange , a .- m . , dumas , c. , et al .
2004 , a&a , 425 , l29 . 2005 , a&a , 438 , l25 chauvin , g. , lagrange , a .- m . , udry , s. , et al .
2006 , a&a , 456 , 1165 chauvin , g. , lagrange , a .- m . , udry , s. , & mayor , m. 2007 , a&a , 475 , 723 chauvin , g. , thomson , m. , dumas , c. , et al .
2003 , a&a , 404 , 157 chauvin , g. , lagrange , a .- m . , zuckerman , b. , et al .
2005 , a&a , 438 , l29 chauvin , g. , lagrange , a .- m . ,
lacombe , f. , et al .
2005 , a&a , 430 , 1027 chauvin , g. , lagrange , a .- m . , bonavita , m. , et al .
2010 , a&a , 509 , a52 chauvin , g. , lagrange , a .- m . , beust , h. , et al .
2012 , a&a , 542 , a41 chauvin , g. , vigan , a. , bonnefoy , m. , et al .
2015 , a&a , 573 , a127 cheetham , a. , hulamo , n. , lacour , s. , de gregorio - monsalvo , i. , & tuthill , p. 2015
, monthly notices ras letters , 450 , l1 chiang , e. , kite , e. , kalas , p. , graham , j. r. , & clampin , m. 2009 , the astrophysical journal , 693 , 734 chun , m. , toomey , d. , wahhaj , z. , et al .
2008 , spie , 7015 , 1 cieza , l. a. , padgett , d. l. , allen , l. e. , et al .
2009 , apjl , 696 , l84 clanton , c. , & gaudi , b. s. 2016 , apj , 819 , 1 claudi , r. u. , turatto , m. , gratton , r. g. , et al .
2008 , in spie astronomical telescopes + instrumentation , ed .
i. s. mclean & m. m. casali ( spie ) , 70143e close , l. m. , males , j. r. , kopon , d. a. , et al .
2012 , adaptive optics systems iii .
proceedings of the spie , 8447 close , l. m. , males , j. r. , morzinski , k. , et al .
2013 , apj , 774 , 94 close , l. m. , follette , k. b. , males , j. r. , et al .
2014 , the astrophysical journal , 781 , l30 crepp , j. r. , gonzales , e. j. , bechter , e. b. , et al .
2016 , arxiv , arxiv:1604.00398 crepp , j. r. , johnson , j. a. , howard , a. w. , et al .
2014 , apj , 781 , 29 . 2013 , the astrophysical journal , 774 , 1 crepp , j. r. , pueyo , l. , brenner , d. , et al .
2011 , the astrophysical journal , 729 , 132 crepp , j. r. , johnson , j. a. , fischer , d. a. , et al .
2012 , the astrophysical journal , 751 , 97 crepp , j. r. , johnson , j. a. , howard , a. w. , et al .
2012 , the astrophysical journal , 761 , 39 . 2013 , apj , 771 , 46 crepp , j. r. , rice , e. l. , veicht , a. , et al .
2015 , the astrophysical journal letters , 798 , l43 crossfield , i. j. m. 2014 , a&a , 566 , a130 . 2015 , publications of the astronomical society of the pacific , 127 , 941 crossfield , i. j. m. , biller , b. , schlieder , j. e. , et al .
2014 , nature , 505 , 654 cumming , a. , butler , r. p. , marcy , g. w. , et al .
2008 , pasp , 120 , 531 currie , t. , bailey , v. , fabrycky , d. , et al .
2010 , the astrophysical journal , 721 , l177 currie , t. , burrows , a. , & daemgen , s. 2014 , apj , 787 , 104 currie , t. , cloutier , r. , brittain , s. , et al .
2015 , the astrophysical journal letters , 814 , 1 currie , t. , cloutier , r. , debes , j. h. , kenyon , s. j. , & kaisler , d. 2013 , the astrophysical journal letters , 777 , l6 currie , t. , daemgen , s. , debes , j. , et al .
2014 , the astrophysical journal , 780 , l30 currie , t. , fukagawa , m. , thalmann , c. , matsumura , s. , & plavchan , p. 2012
, the astrophysical journal , 755 , l34 currie , t. , thalmann , c. , matsumura , s. , et al .
2011 , the astrophysical journal , 736 , l33 currie , t. , burrows , a. , itoh , y. , et al .
2011 , the astrophysical journal , 729 , 128 currie , t. , debes , j. , rodigas , t. j. , et al .
2012 , the astrophysical journal letters , 760 , l32 currie , t. , burrows , a. , girard , j. h. , et al .
2014 , apj , 795 , 133 currie , t. , muto , t. , kudo , t. , et al .
2014 , the astrophysical journal letters , 796 , l30 cushing , m. c. , marley , m. s. , saumon , d. , et al .
2008 , the astrophysical journal , 678 , 1372 cushing , m. c. , kirkpatrick , j. d. , gelino , c. r. , et al .
2011 , the astrophysical journal , 743 , 50 daemgen , s. , bonavita , m. , jayawardhana , r. , lafreniere , d. , & janson , m. 2015 , apj , 799 , 155 davies , r. , & kasper , m. 2012 , annu .
astrophys . , 50 , 305 dawson , r. i. , murray - clay , r. a. , & fabrycky , d. c. 2011 , the astrophysical journal , 743 , l17 de rosa , r. j. , nielsen , e. l. , blunt , s. c. , et al .
2015 , the astrophysical journal letters , 814 , 1 de rosa , r. j. , rameau , j. , patience , j. , et al .
2016 , arxiv , 1604.01411v1 deacon , n. r. , schlieder , j. e. , & murphy , s. j. 2016 , monthly notices ras , 457 , 3191 deacon , n. r. , liu , m. c. , magnier , e. a. , et al .
2014 , apj , 792 , 119 dekany , r. , roberts , j. , burruss , r. , et al .
2013 , apj , 776 , 130 delorme , p. , lagrange , a. m. , chauvin , g. , et al .
2012 , a&a , 539 , a72 delorme , p. , gagn , j. , girard , j. h. , et al .
2013 , a&a , 553 , l5 dent , w. r. f. , walker , h. j. , holland , w. s. , & greaves , j. s. 2000 , monthly notices ras , 314 , 702 dent , w. r. f. , wyatt , m. c. , roberge , a. , et al .
2014 , science , 1490 desidera , s. , covino , e. , messina , s. , et al .
2015 , a&a , 573 , a126 dohlen , k. , langlois , m. , saisse , m. , et al .
2008 , in spie astronomical telescopes + instrumentation , ed .
i. s. mclean & m. m. casali , lab .
dastrophysique de marseille , cnrs , univ . de provence ( france ) ( spie ) , 70143l dong , r. , fung , j. , & chiang , e. 2016 , arxiv , 1602.04814v1 dong , r. , hall , c. , rice , k. , & chiang , e. 2015 , the astrophysical journal letters , 812 , l32 dong , r. , zhu , z. , fung , j. , et al .
2016 , the astrophysical journal letters , 816 , l12 dou , j. , ren , d. , zhao , g. , et al .
2015 , apj , 802 , 1 duchne , g. 2010 , the astrophysical journal , 709 , l114 duchne , g. , & kraus , a. 2013 , annu
astrophys . , 51 , 269 dupuy , t. j. , kratter , k. m. , kraus , a. l. , et al .
2016 , apj , 817 , 1 dupuy , t. j. , & kraus , a. l. 2013 , science , 341 , 1492 dupuy , t. j. , & liu , m. c. 2011 , the astrophysical journal , 733 , 122 . 2012 , the astrophysical journal supplement , 201 , 19 dupuy , t. j. , liu , m. c. , & ireland , m. j. 2009 , the astrophysical journal , 692 , 729 . 2014 , apj , 790 , 133 dupuy , t. j. , liu , m. c. , & leggett , s. k. 2015 , apj , 803 , 1 dupuy , t. j. , liu , m. c. , leggett , s. k. , et al .
2015 , apj , 805 , 1 durisen , r. h. , boss , a. p. , mayer , l. , et al .
2007 , in protostars and planets v , e. b. reipurth , d. jewitt , & k. keil ( tucson , az : univ .
arizona press ) , 607 ehrenreich , d. , lagrange , a .- m . , montagnier , g. , et al .
2010 , a&a , 523 , a73 eiroa , c. , marshall , j. p. , mora , a. , et al .
2013 , a&a , 555 , a11 els , s. g. , sterzik , m. f. , marchis , f. , et al .
2001 , a&a , 370 , l1 espaillat , c. , muzerolle , j. , najita , j. , et al .
2014 , protostars and planets vi , 497 esposito , s. , mesa , d. , skemer , a. , et al . 2012 , a&a , 549 , a52 fabrycky , d. c. , & murray - clay , r. a. 2010 , the astrophysical journal , 710 , 1408 faherty , j. k. , rice , e. l. , cruz , k. l. , mamajek , e. e. , & nez , a. 2012 , the astronomical journal , 145 , 2 farihi , j. , bond , h. e. , dufour , p. , et al .
2013 , monthly notices ras , 430 , 652 feigelson , e. d. , lawson , w. a. , & stark , m. 2006 , the astronomical fergus , r. , hogg , d. w. , oppenheimer , r. , brenner , d. , & pueyo , l. 2014 , apj , 794 , 161 filippazzo , j. c. , rice , e. l. , faherty , j. , et al .
2015 , apj , 810 , 1 fitzgerald , m. p. , kalas , p. g. , & graham , j. r. 2009 , the astrophysical journal , 706 , l41 fortney , j. j. , marley , m. s. , saumon , d. , & lodders , k. 2008 , the astrophysical journal , 683 , 1104 fortney , j. j. , & nettelmann , n. 2009 , space sci rev , 1 forveille , t. , s gransan , d. , delorme , p. , et al .
2004 , a&a , 427 , l1 fuhrmann , k. , & chini , r. 2015 , apj , 806 , 1 fukagawa , m. , itoh , y. , tamura , m. , et al .
2009 , the astrophysical journal , 696 , l1 gagn , j. , burgasser , a. j. , faherty , j. k. , et al .
2015 , the astrophysical journal letters , 808 , 1 gagn , j. , lafreniere , d. , doyon , r. , malo , l. , & artigau , e. 2014 , apj , 783 , 121 . 2015 , apj , 798 , 73 gagn , j. , faherty , j. k. , cruz , k. l. , et al .
2015 , the astrophysical journal supplement series , 219 , 1 gaidos , e. , mann , a. w. , lpine , s. , et al . 2014 , monthly notices ras , 443 , 2561 galicher , r. , marois , c. , macintosh , b. , barman , t. , & konopacky , q. 2011 , the astrophysical journal , 739 , l41 galicher , r. , rameau , j. , bonnefoy , m. , et al .
2014 , a&a , 565 , l4 garcia , e. v. , dupuy , t. j. , allers , k. n. , liu , m. c. , & deacon , n. r. 2015 , apj , 804 , 1 garufi , a. , quanz , s. p. , schmid , h. m. , et al .
2016 , arxiv , 1601.04983v1 gauza , b. , bejar , v. j. s. , prez - garrido , a. , et al .
2015 , the astrophysical journal , 804 , 96 geiler , k. , kellner , s. , brandner , w. , et al .
2007 , a&a , 461 , 665 ginski , c. , mugrauer , m. , seeliger , m. , et al .
2016 , monthly notices ras , 457 , 2173 gizis , j. e. 2002 , the astrophysical journal , 575 , 484 goldman , b. , marsat , s. , henning , t. , clemens , c. , & greiner , j. 2010 , monthly notices ras , 405 , 1140 golimowski , d. a. , leggett , s. k. , marley , m. s. , et al .
2004 , the astronomical journal , 127 , 3516 golimowski , d. a. , henry , t. j. , krist , j. e. , et al .
2004 , the astronomical journal , 128 , 1733 gonzalez , c. a. g. , absil , o. , absil , p. a. , et al .
2016 , arxiv , 1602.08381v1 godziewski , k. , & migaszewski , c. 2009 , monthly notices ras letters , 397 , l16 . 2014 , monthly notices ras , 440 , 3140 greaves , j. s. , holland , w. s. , moriarty - schieven , g. , et al .
1998 , the astrophysical journal , 506 , l133 greaves , j. s. , holland , w. s. , wyatt , m. c. , et al .
2005 , apjl , 619 , l187 greco , j. p. , & brandt , t. d. 2016 , arxiv , 1602.00691v1 greco , j. p. , & burrows , a. 2015 , apj , 808 , 1 guyon , o. 2003 , a&a , 404 , 379 guyon , o. , pluzhnik , e. a. , galicher , r. , et al .
2005 , the astrophysical journal , 622 , 744 guyon , o. , pluzhnik , e. a. , kuchner , m. j. , collins , b. , & ridgway , s. t. 2006 , the astrophysical journal supplement series , 167 , 81 hagelberg , j. , sgransan , d. , udry , s. , & wildi , f. 2016 , monthly notices ras , 455 , 2178 han , e. , wang , s. x. , wright , j. t. , et al .
2014 , publications of the astronomical society of the pacific , 126 , 827 hashimoto , j. , tamura , m. , muto , t. , et al .
2011 , the astrophysical journal , 729 , l17 hatzes , a. p. , cochran , w. d. , mcarthur , b. , et al .
2000 , the astrophysical journal , 544 , l145 hayward , t. l. , biller , b. a. , liu , m. c. , et al .
2014 , publications of the astronomical society of the pacific , 126 , 1112 heap , s. r. , lindler , d. j. , lanz , t. m. , et al .
2000 , the astrophysical journal , 539 , 435 heinze , a. n. , hinz , p. m. , kenworthy , m. , et al .
2010 , the astrophysical journal , 714 , 1570 heinze , a. n. , hinz , p. m. , kenworthy , m. , miller , d. , & sivanandam , s. 2008 , the astrophysical journal , 688 , 583 heinze , a. n. , hinz , p. m. , sivanandam , s. , et al .
2010 , the astrophysical journal , 714 , 1551 helled , r. , bodenheimer , p. , podolak , m. , et al .
2013 , in protostars and planets vi , ed
. h. beuther et al .
( tucson , az : univ .
arizona press ) , 643 helling , c. , & casewell , s. 2014 , astron astrophys rev , 22 , 80 helling , c. , woitke , p. , rimmer , p. , et al .
2014 , life , 4 , 142 helling , c. , ackerman , a. , allard , f. , et al .
2008 , monthly notices ras , 391 , 1854 henry , t. j. , & mccarthy , d. w. j. 1990 , astrophysical journal , 350 , 334 herczeg , g. j. , & hillenbrand , l. a. 2015 , the astrophysical journal , 808 , 23 hinkley , s. , oppenheimer , b. r. , soummer , r. , et al .
2007 , the astrophysical journal , 654 , 633 hinkley , s. , oppenheimer , b. r. , brenner , d. , et al .
2010 , the astrophysical journal , 712 , 421 hinkley , s. , oppenheimer , b. r. , zimmerman , n. , et al .
2011 , publications of the astronomical society of the pacific , 123 , 74 hinkley , s. , pueyo , l. , faherty , j. k. , et al .
2013 , apj , 779 , 153 hinkley , s. , bowler , b. p. , vigan , a. , et al .
2015 , the astrophysical journal letters , 805 , 1 hinkley , s. , kraus , a. l. , ireland , m. j. , et al .
2015 , the astrophysical journal letters , 806 , 1 hinz , p. m. , heinze , a. n. , sivanandam , s. , et al .
2006 , apj , 653 , 1486 hinz , p. m. , rodigas , t. j. , kenworthy , m. a. , et al .
2010 , the astrophysical journal , 716 , 417 howard , a. w. , johnson , j. a. , marcy , g. w. , et al .
2010 , the astrophysical journal , 721 , 1467 howard , a. w. , marcy , g. w. , bryson , s. t. , et al .
2012 , the astrophysical journal supplement series , 201 , 15 howard , a. w. , marcy , g. w. , fischer , d. a. , et al .
2014 , apj , 794 , 51 huby , e. , perrin , g. , marchis , f. , et al .
2012 , a&a , 541 , a55 hulamo , n. , lacour , s. , tuthill , p. , et al .
2011 , a&a , 528 , l7 hung , l .- w . , duchene , g. , arriaga , p. , et al .
2015 , the astrophysical journal letters , 815 , 1 ingraham , p. , marley , m. s. , saumon , d. , et al .
2014 , the astrophysical journal letters , 794 , l15 ireland , m. j. , kraus , a. , martinache , f. , law , n. , & hillenbrand , l. a. 2011 , the astrophysical journal , 726 , 113 ireland , m. j. , & kraus , a. l. 2008 , the astrophysical journal , 678 , l59 . 2014 , iau , 8 , 199 irwin , a. w. , fletcher , j. m. , yang , s. l. s. , walker , g. a. h. , & goodenough , c. 1992 , astronomical society of the pacific , 104 , 489 isella , a. , chandler , c. j. , carpenter , j. m. , prez , l. m. , & ricci , l. 2014 , apj , 788 , 129 itoh , y. , oasa , y. , & fukagawa , m. 2006 , the astrophysical journal itoh , y. , hayashi , m. , tamura , m. , et al .
2005 , the astrophysical journal , 620 , 984 itoh , y. , tamura , m. , hayashi , m. , et al .
2008 , publications of the astronomical society of japan , 60 , 209 janson , m. , bonavita , m. , klahr , h. , et al .
2011 , the astrophysical journal , 736 , 89 janson , m. , carson , j. c. , lafrenire , d. , et al .
2012 , the astrophysical journal , 747 , 116 janson , m. , lafreniere , d. , jayawardhana , r. , et al .
2013 , apj , 773 , 170 janson , m. , quanz , s. p. , carson , j. c. , et al .
2015 , a&a , 574 , a120 janson , m. , reffert , s. , brandner , w. , et al .
2008 , a&a , 488 , 771 janson , m. , brandner , w. , henning , t. , et al .
2007 , the astronomical journal , 133 , 2442 janson , m. , apai , d. , zechmeister , m. , et al .
2009 , monthly notices ras , 399 , 377 janson , m. , carson , j. , thalmann , c. , et al .
2011 , apj , 728 , 85 janson , m. , brandt , t. d. , kuzuhara , m. , et al .
2013 , the astrophysical journal , 778 , l4 janson , m. , brandt , t. d. , moro - martn , a. , et al .
2013 , apj , 773 , 73 janson , m. , bergfors , c. , brandner , w. , et al .
2014 , the astrophysical journal supplement series , 214 , 17 jeffries , r. d. 2014 , eas publications series , 65 , 289 jenkins , j. s. , jones , h. r. a. , biller , b. , et al .
2010 , a&a , 515 , a17 jenkins , j. s. , jones , h. r. a. , tuomi , m. , et al .
2016 , arxiv , 1603.09391v1 jensen - clem , r. , millar - blanchaer , m. , mawet , d. , et al .
2016 , apj , 820 , 1 jilkova , l. , & zwart , s. p. 2015 , monthly notices ras , 451 , 804 johnson , j. a. 2009 , publications of the astronomical society of the pacific , 121 , 309 johnson , j. a. , aller , k. m. , howard , a. w. , & crepp , j. r. 2010 , publications of the astronomical society of the pacific , 122 , 905 johnson , j. a. , butler , r. p. , marcy , g. w. , et al .
2007 , the astrophysical journal , 670 , 833 jones , j. , white , r. j. , quinn , s. , et al .
2016 , the astrophysical journal letters , 822 , 1 jovanovic , n. , martinache , f. , guyon , o. , et al .
2015 , publications of the astronomical society of pacific , 127 , 890 kalas , p. , graham , j. , chiang , e. , & fitzgerald , m. 2008 , science express kalas , p. , graham , j. r. , & clampin , m. 2005 , nature , 435 , 1067 kalas , p. , graham , j. r. , fitzgerald , m. p. , & clampin , m. 2013 , apj , 775 , 56 kalas , p. , & jewitt , d. 1995 , astronomical journal v.110 , 110 , 794 kalas , p. , graham , j. r. , chiang , e. , et al .
2008 , science , 322 , 1345 kalas , p. g. , rajan , a. , wang , j. j. , et al .
2015 , apj , 814 , 1 kasper , m. , apai , d. , janson , m. , & brandner , w. 2007 , a&a , 472 , 321 kellogg , k. , metchev , s. , gagn , j. , & faherty , j. 2016 , the astrophysical journal letters , 821 , 1 kennedy , g. m. , & wyatt , m. c. 2011 , monthly notices ras , 412 , 2137 kenworthy , m. a. , codona , j. l. , hinz , p. m. , et al .
2007 , apj , 660 , 762 kenworthy , m. a. , mamajek , e. e. , hinz , p. m. , et al .
2009 , the astrophysical journal , 697 , 1928 kenworthy , m. a. , meshkat , t. , quanz , s. p. , et al .
2013 , apj , 764 , 7 kenyon , s. j. , & bromley , b. c. 2004 , the astronomical journal kenyon , s. j. , currie , t. , & bromley , b. c. 2014 , apj , 786 , 70 kirkpatrick , j. d. , gelino , c. r. , cushing , m. c. , et al .
2012 , the astrophysical journal , 753 , 156 kiss , l. l. , mor , a. , szalai , t. , et al .
2010 , monthly notices ras , 411 , 117 knutson , h. a. , fulton , b. j. , montet , b. t. , et al .
2014 , apj , 785 , 126 konopacky , q. m. , barman , t. s. , macintosh , b. a. , & marois , c. 2013 , science , 339 , 1398 konopacky , q. m. , marois , c. , macintosh , b. a. , et al .
2016 , arxiv , 1604.08157v1 kratter , k. m. , & lodato , g. 2016 , arxiv , 1603.01280v1 kraus , a. l. , andrews , s. m. , bowler , b. p. , et al .
2015 , the astrophysical journal , 798 , l23 kraus , a. l. , & hillenbrand , l. a. 2012 , the astrophysical journal , 757 , 141 kraus , a. l. , & ireland , m. j. 2012 , the astrophysical journal , 745 , 5 kraus , a. l. , ireland , m. j. , cieza , l. a. , et al .
2014 , apj , 781 , 20 kraus , a. l. , ireland , m. j. , hillenbrand , l. a. , & martinache , f. 2012 , the astrophysical journal , 745 , 19 kraus , a. l. , ireland , m. j. , huber , d. , mann , a. w. , & dupuy , t. j. 2016 , arxiv , 1604.05744v1 kraus , a. l. , ireland , m. j. , martinache , f. , & lloyd , j. p. 2008
, the astrophysical journal , 679 , 762 kraus , a. l. , shkolnik , e. l. , allers , k. n. , & liu , m. c. 2014 , aj , 147 , 146 kraus , a. l. , white , r. j. , & hillenbrand , l. a. 2005 , apj , 633 , 452 krivov , a. v. 2010 , research in astron .
astrophys . , 383 kuchner , m. j. , & brown , m. e. 2000 , publications of the astronomical society of the pacific , 112 , 827 kumar , s. s. 1963 , astrophysical journal , 137 , 1121 kuzuhara , m. , tamura , m. , ishii , m. , et al .
2011 , the astronomical journal , 141 , 119 kuzuhara , m. , tamura , m. , kudo , t. , et al .
2013 , the astrophysical journal , 774 , 11 lachapelle , f .-
r . , lafreniere , d. , gagn , j. , et al .
2015 , apj , 802 , 1 lacour , s. , biller , b. , cheetham , a. , et al .
2015 , arxiv , 1511.09390v1 lafrenire , d. , jayawardhana , r. , & van kerkwijk , m. h. 2008 , the astrophysical journal , 689 , l153 . 2010 , the astrophysical journal , 719 , 497 lafrenire , d. , jayawardhana , r. , van kerkwijk , m. h. , brandeker , a. , & janson , m. 2014 , apj , 785 , 47 lafrenire , d. , marois , c. , doyon , r. , & barman , t. 2009 , the astrophysical journal , 694 , l148 lafrenire , d. , marois , c. , doyon , r. , nadeau , d. , & artigau , . 2007 , the astrophysical journal , 660 , 770 lafrenire , d. , doyon , r. , marois , c. , et al .
2007 , the astrophysical journal , 670 , 1367 lagrange , a .- m . 2014 , philosophical transactions of the royal society a : mathematical , physical and engineering sciences , 372 , 20130090 lagrange , a .- m . , beust , h. , udry , s. , chauvin , g. , & mayor , m. 2006 , a&a , 459 , 955 lagrange , a .- m . , meunier , n. , chauvin , g. , et al .
2013 , a&a , 559 , a83 lagrange , a .- m . ,
gratadour , d. , chauvin , g. , et al .
2009 , a&a , 493 , l21 lagrange , a .- m . , kasper , m. , boccaletti , a. , et al .
2009 , a&a , 506 , 927 lagrange , a .- m . , bonnefoy , m. , chauvin , g. , et al .
2010 , science , 329 , 57 lagrange , a .- m . , langlois , m. , gratton , r. , et al .
2016 , a&a , 586 , l8 langlois , m. , dohlen , k. , vigan , a. , et al .
2014 , in spie astronomical telescopes + instrumentation , ed .
s. k. ramsay , i. s. mclean , & h. takami ( spie ) , 91471r larwood , j. d. , & kalas , p. g. 2001 , mnras , 323 , 402 leconte , j. , soummer , r. , hinkley , s. , et al .
2010 , the astrophysical journal , 716 , 1551 lee , j .- m . , heng , k. , & irwin , p. g. j. 2013 , apj , 778 , 97 lpine , s. , & simon , m. 2009 , the astronomical journal , 137 , 3632 line , m. r. , fortney , j. j. , marley , m. s. , & sorahana , s. 2014 , apj , 793 , 33 liu , m. c. 2004 , science , 305 , 1442 liu , m. c. , dupuy , t. j. , bowler , b. p. , leggett , s. k. , & best , w. m. j. 2012 , the astrophysical journal , 758 , 57 liu , m. c. , fischer , d. a. , graham , j. r. , et al .
2002 , the astrophysical journal , 571 , 519 liu , m. c. , wahhaj , z. , biller , b. a. , et al .
2010 , spie , 7736 , 77361k liu , m. c. , delorme , p. , dupuy , t. j. , et al .
2011 , the astrophysical journal , 740 , 108 liu , m. c. , magnier , e. a. , deacon , n. r. , et al .
2013 , the astrophysical journal , 777 , l20 lodato , g. , delgado - donate , e. , & clarke , c. j. 2005 , monthly notices ras letters , 364 , l91 looper , d. l. , mohanty , s. , bochanski , j. j. , et al .
2010 , the astrophysical journal , 714 , 45 lopez , s. , & jenkins , j. s. 2012 , apj , 756 , 177 lovis , c. , & mayor , m. 2007 , a&a , 472 , 657 low , c. , & lynden - bell , d. 1976 , royal astronomical society , 176 , 367 lowrance , p. j. , mccarthy , c. , becklin , e. e. , et al .
1999 , the astrophysical journal , 512 , l69 lowrance , p. j. , schneider , g. , kirkpatrick , j. d. , et al .
2000 , the astrophysical journal , 541 , 390 lowrance , p. j. , becklin , e. e. , schneider , g. , et al .
2005 , the astronomical journal , 130 , 1845 lucas , p. w. , roche , p. f. , allard , f. , & hauschildt , p. h. 2001 , monthly notices ras , 326 , 695 luhman , k. l. , burgasser , a. j. , & bochanski , j. j. 2011 , the astrophysical journal , 730 , l9 luhman , k. l. , burgasser , a. j. , labb , i. , et al .
2012 , the astrophysical journal , 744 , 135 luhman , k. l. , & jayawardhana , r. 2002 , the astrophysical journal , 566 , 1132 luhman , k. l. , mamajek , e. e. , allen , p. r. , & cruz , k. l. 2009 , the astrophysical journal , 703 , 399 luhman , k. l. , mamajek , e. e. , allen , p. r. , muench , a. a. , & finkbeiner , d. p. 2009
, the astrophysical journal , 691 , 1265 luhman , k. l. , mcleod , k. k. , & goldenson , n. 2005 , the astrophysical journal , 623 , 1141 luhman , k. l. , wilson , j. c. , brandner , w. , et al . 2006 , the astrophysical journal , 649 , 894 luhman , k. l. , patten , b. m. , marengo , m. , et al .
2007 , the astrophysical journal , 654 , 570 macintosh , b. , graham , j. r. , ingraham , p. , et al .
2014 , proceedings of the national academy of sciences , 111 , 12661 macintosh , b. , graham , j. r. , barman , t. , et al .
2015 , science macintosh , b. a. , becklin , e. e. , kaisler , d. , konopacky , q. , & zuckerman , b. 2003 , the astrophysical journal , 594 , 538 macintosh , b. a. , max , c. , zuckerman , b. , et al .
2001 , young stars near earth : progress and prospects , 244 , 309 madhusudhan , n. , knutson , h. , fortney , j. , & barman , t. 2014 , arxiv , 1402.1169v1 maire , a. l. , boccaletti , a. , rameau , j. , et al .
2014 , a&a , 566 , a126 maire , a. l. , skemer , a. j. , hinz , p. m. , et al .
2015 , a&a , 576 , a133 . 2015 , a&a , 579 , c2 maire , a. l. , bonnefoy , m. , ginski , c. , et al .
2016 , a&a , 587 , a56 males , j. r. , close , l. m. , morzinski , k. m. , et al .
2014 , apj , 786 , 32 malo , l. , artigau , e. , doyon , r. , et al .
2014 , apj , 788 , 81 malo , l. , doyon , r. , lafrenire , d. , et al .
2013 , apj , 762 , 88 mamajek , e. e. 2012 , the astrophysical journal , 754 , l20 . 2016 , in young stars & planets near the sun , proceedings of the international astronomical union , iau symposium , 314 , 21 mamajek , e. e. , & bell , c. p. m. 2014 , monthly notices ras , 445 , 2169 mamajek , e. e. , bartlett , j. l. , seifahrt , a. , et al . 2013 , the astronomical journal , 146 , 154 marengo , m. , megeath , s. t. , fazio , g. g. , et al . 2006 , the astrophysical journal , 647 , 1437 marengo , m. , stapelfeldt , k. , werner , m. w. , et al
. 2009 , the astrophysical journal , 700 , 1647 marleau , g. d. , & cumming , a. 2013 , monthly notices ras , 437 , 1378 marley , m. s. , ackerman , a. s. , cuzzi , j. n. , & kitzmann , d. 2013 , arxiv , astro-ph.ep marley , m. s. , fortney , j. , seager , s. , & barman , t. 2007 , protostars and planets v , 733 marley , m. s. , fortney , j. j. , hubickyj , o. , bodenheimer , p. , & lissauer , j. j. 2007 , the astrophysical journal , 655 , 541 marley , m. s. , & robinson , t. d. 2015 , annu .
astrophys . , 53 , 279 marley , m. s. , saumon , d. , cushing , m. , et al .
2012 , the astrophysical journal , 754 , 135 marois , c. , correia , c. , galicher , r. , et al .
2014 , in spie astronomical telescopes + instrumentation , ed .
e. marchetti , l. m. close , & j .-
veran ( spie ) , 91480u marois , c. , lafrenire , d. , doyon , r. , macintosh , b. , & nadeau , d. 2006 , the astrophysical journal , 641 , 556 marois , c. , macintosh , b. , barman , t. , et al .
2008 , science , 322 , 1348 marois , c. , macintosh , b. , & vran , j .- p .
2010 , proc .
spie , 7736 , 77361j marois , c. , zuckerman , b. , konopacky , q. m. , macintosh , b. , & barman , t. 2010 , nature , 468 , 1080 marshall , j. p. , moro - martin , a. , eiroa , c. , et al .
2014 , a&a , 565 , a15 martinache , f. , guyon , o. , jovanovic , n. , et al .
2014 , publications of the astronomical society of pacific , 126 , 565 masciadri , e. , mundt , r. , henning , t. , alvarez , c. , & barrado y navascus , d. 2005 , the astrophysical journal , 625 , 1004 matthews , b. c. , krivov , a. v. , wyatt , m. c. , bryden , g. , & eiroa , c. 2014 , in protostars and planets vi , henrik beuther , ralf s. klessen , cornelis p. dullemond , and thomas henning ( eds . ) , university of arizona press , tucson , 521 matthews , c. t. , crepp , j. r. , skemer , a. , et al .
2014 , the astrophysical journal letters , 783 , l25 mawet , d. , riaud , p. , absil , o. , & surdej , j. 2005 , the astrophysical journal mawet , d. , serabyn , e. , liewer , k. , et al .
2010 , the astrophysical journal , 709 , 53 mawet , d. , absil , o. , montagnier , g. , et al .
2012 , a&a , 544 , a131 mawet , d. , pueyo , l. , lawson , p. , et al .
2012 , in spie astronomical telescopes + instrumentation , ed .
m. c. clampin , g. g. fazio , h. a. macewen , & j. m. oschmann ( spie ) , 844204 mawet , d. , milli , j. , wahhaj , z. , et al .
2014 , apj , 792 , 97 mawet , d. , david , t. , bottom , m. , et al .
2015 , apj , 811 , 1 mayama , s. , hashimoto , j. , muto , t. , et al .
2012 , the astrophysical journal letters , 760 , l26 mccarthy , c. , & zuckerman , b. 2004 , the astronomical journal , 127 , 2871 mccarthy , k. , & wilhelm , r. j. 2014 , the astronomical journal , 148 , 70 mennesson , b. , serabyn , e. , hanot , c. , et al .
2011 , apj , 736 , 14 meshkat , t. , bailey , v. p. , su , k. y. l. , et al .
2015 , apj , 800 , 5 meshkat , t. , kenworthy , m. a. , quanz , s. p. , & amara , a. 2013 , apj , 780 , 17 meshkat , t. , kenworthy , m. a. , reggiani , m. , et al .
2015 , monthly notices ras , 453 , 2534 meshkat , t. , bailey , v. , rameau , j. , et al .
2013 , the astrophysical journal , 775 , l40 meshkat , t. , bonnefoy , m. , mamajek , e. e. , et al .
2015 , monthly notices ras , 453 , 2379 metchev , s. , marois , c. , & zuckerman , b. 2009 , the astrophysical journal , 705 , l204 metchev , s. a. , & hillenbrand , l. a. 2004 , the astrophysical journal , 617 , 1330 . 2006 , the astrophysical journal , 651 , 1166 . 2009 , the astrophysical journal supplement series , 181 , 62 metchev , s. a. , hillenbrand , l. a. , & white , r. j. 2003 , the astrophysical journal , 582 , 1102 metchev , s. a. , heinze , a. , apai , d. , et al .
2015 , apj , 799 , 154 meyer , m. r. , carpenter , j. m. , mamajek , e. e. , et al .
2008 , apjl , 673 , l181 millar - blanchaer , m. a. , graham , j. r. , pueyo , l. , et al .
2015 , apj , 811 , 1 milli , j. , lagrange , a .- m . , mawet , d. , et al .
2014 , a&a , 566 , a91 mohanty , s. , greaves , j. , mortlock , d. , et al .
2013 , apj , 773 , 168 mollire , p. , & mordasini , c. 2012 , a&a , 547 , a105 montet , b. t. , crepp , j. r. , johnson , j. a. , howard , a. w. , & marcy , g. w. 2014 , apj , 781 , 28 montet , b. t. , bowler , b. p. , shkolnik , e. l. , et al .
2015 , the astrophysical journal letters , 813 , 1 moor , a. , szabo , g. m. , kiss , l. l. , et al .
2013 , monthly notices ras , 1 mordasini , c. 2013 , a&a , 558 , a113 mordasini , c. , alibert , y. , klahr , h. , & henning , t. 2012 , a&a , 547 , a111 moro - martin , a. , wyatt , m. c. , malhotra , r. , & trilling , d. e. 2008 , the solar system beyond neptune , 465 moro - martin , a. , marshall , j. p. , kennedy , g. , et al .
2015 , apj , 801 , 1 morzinski , k. m. , close , l. m. , males , j. r. , et al .
2014 , in spie astronomical telescopes + instrumentation , ed .
e. marchetti , l. m. close , & j .-
veran ( spie ) , 914804 morzinski , k. m. , males , j. r. , skemer , a. j. , et al .
2015 , apj , 815 , 1 mouillet , d. , larwood , j. d. , papaloizou , j. c. b. , & lagrange , a .- m .
1997 , monthly notices ras , 292 , 896 mugrauer , m. , & neuhuser , r. 2005 , monthly notices ras letters , 361 , l15 mugrauer , m. , neuh user , r. , guenther , e. w. , et al .
2004 , a&a , 417 , 1031 muto , t. , grady , c. a. , hashimoto , j. , et al .
2012 , the astrophysical journal , 748 , l22 muzic , k. , scholz , a. , geers , v. c. , & jayawardhana , r. 2015 , apj , 810 , 1 nakajima , t. , durrance , s. t. , golimowski , d. a. , & kulkarni , s. r. 1994 , the astrophysical journal , 428 , 797 nakajima , t. , morino , j .-
i . , tsuji , t. , et al .
2005 , astronomische nachrichten , 326 , 952 naud , m .- e . , artigau , e. , malo , l. , et al .
2014 , apj , 787 , 5 neuhuser , r. , guenther , e. w. , alves , j. , et al .
2003 , astronomische nachrichten , 324 , 535 neuhuser , r. , mugrauer , m. , fukagawa , m. , torres , g. , & schmidt , t. 2007 , a&a , 462 , 777 ngo , h. , knutson , h. a. , hinkley , s. , et al .
2015 , apj , 800 , 138 nielsen , e. l. , & close , l. m. 2010 , the astrophysical journal , 717 , 878 nielsen , e. l. , close , l. m. , biller , b. a. , masciadri , e. , & lenzen , r. 2008 , the astrophysical journal , 674 , 466 nielsen , e. l. , liu , m. c. , wahhaj , z. , et al .
2012 , the astrophysical journal , 750 , 53 . 2013 , apj , 776 , 4 . 2014 , apj , 794 , 158 norris , b. , schworer , g. , tuthill , p. , et al .
2015 , monthly notices ras , 447 , 2894 oberg , k. i. , murray - clay , r. , & bergin , e. a. 2011 , the astrophysical journal , 743 , l16 olofsson , j. , benisty , m. , le bouquin , j .- b . , et al .
2013 , a&a , 552 , a4 olofsson , j. , samland , m. , avenhaus , h. , et al .
2016 , arxiv , 1601.07861v1 oppenheimer , b. r. , golimowski , d. a. , kulkarni , s. r. , et al .
2001 , the astronomical journal , 121 , 2189 oppenheimer , b. r. , & hinkley , s. 2009 , annu .
astrophys . , 47 , 253 oppenheimer , b. r. , kulkarni , s. r. , & stauffer , j. r. 2000 , protostars and planets iv ( book - tucson : university of arizona press ; eds mannings , 1313 oppenheimer , b. r. , digby , a. p. , newburgh , l. , et al .
2004 , in astronomical telescopes and instrumentation , ed .
d. bonaccini calia , b. l. ellerbroek , & r. ragazzoni ( spie ) , 433442 oppenheimer , b. r. , baranec , c. , beichman , c. , et al .
2013 , the astrophysical journal , 768 , 24 owen , j. e. 2016 , publications of the astronomical society of australia , 33 , e005 owen , j. e. , & menou , k. 2016 , the astrophysical journal letters , 819 , 1 ozernoy , l. m. , gorkavyi , n. n. , mather , j. c. , & taidakova , t. a. 2000 , the astrophysical journal , 537 , l147 patience , j. , king , r. r. , de rosa , r. j. , & marois , c. 2010 , a&a , 517 , a76 pecaut , m. j. , mamajek , e. e. , & bubar , e. j. 2012 , the astrophysical journal , 746 , 154 perez , s. , dunhill , a. , casassus , s. , et al .
2015 , the astrophysical journal letters , 811 , 1 perryman , m. 2011 , the exoplanet handbook ( cambridge university press ) perryman , m. , hartman , j. , bakos , g. .
, & lindegren , l. 2014 , apj , 797 , 14 petr - gotzens , m. g. , cuby , j. g. , smith , m. d. , & sterzik , m. f. 2010 , a&a , 522 , a78 pueyo , l. , crepp , j. r. , vasisht , g. , et al .
2012 , the astrophysical journal supplement series , 199 , 6 pueyo , l. , soummer , r. , hoffmann , j. , et al .
2015 , apj , 803 , 31 quanz , s. p. 2015 , astrophys .
space sci .
, 1 quanz , s. p. , amara , a. , meyer , m. r. , et al .
2015 , apj , 807 , 1 . 2013 , the astrophysical journal , 766 , l1 quanz , s. p. , meyer , m. r. , kenworthy , m. a. , et al .
2010 , the astrophysical journal letters , 722 , l49 queloz , d. , mayor , m. , weber , l. , et al . 2000 , a&a , 354 , 99 radigan , j. , lafreniere , d. , jayawardhana , r. , & artigau , e. 2014 , apj , 793 , 75 rajan , a. , barman , t. , soummer , r. , et al .
2015 , the astrophysical journal letters , 809 , 1 rameau , j. , chauvin , g. , lagrange , a .- m .
, et al .
2015 , a&a . 2013 , a&a , 553 , a60 . 2013 , the astrophysical journal , 779 , l26 . 2013 , the astrophysical , 772 , l15 rameau , j. , nielsen , e. l. , de rosa , r. j. , et al .
2016 , arxiv , 1604.05139v1 rebolo , r. , zapatero osorio , m. r. , madruga , s. , et al .
1998 , science , 282 , 1309 reggiani , m. , quanz , s. p. , meyer , m. r. , et al .
2014 , the astrophysical journal letters , 792 , l23 reggiani , m. , meyer , m. r. , chauvin , g. , et al .
2016 , a&a , 586 , a147 reid , i. n. , cruz , k. l. , kirkpatrick , j. d. , et al .
2008 , the astronomical journal , 136 , 1290 reid , i. n. , & walkowicz , l. m. 2006 , the publications of the astronomical society of the pacific , 118 , 671 riaz , b. , & martn , e. l. 2011 , a&a , 525 , a10 riaz , b. , martn , e. l. , petr - gotzens , m. g. , & monin , j .- l .
2013 , a&a , 559 , a109 rice , e. l. , oppenheimer , r. , zimmerman , n. , roberts , jr , l. c. , & hinkley , s. 2015 , publications of the astronomical society of the pacific , 127 , 479 riedel , a. r. , finch , c. t. , henry , t. j. , et al .
2014 , the astronomical journal , 147 , 85 rieke , g. h. , su , k. y. l. , stansberry , j. a. , et al .
2005 , the astrophysical journal , 620 , 1010 roberts , l. c. , rice , e. l. , beichman , c. a. , et al .
2012 , the astronomical journal , 144 , 14 robinson , t. d. , stapelfeldt , k. r. , & marley , m. s. 2016 , publications of the astronomical society of the pacific , 128 , 1 rodigas , t. j. , follette , k. b. , weinberger , a. , close , l. , & hines , d. c. 2014 , the astrophysical journal letters , 791 , l37 rodigas , t. j. , males , j. r. , hinz , p. m. , mamajek , e. e. , & knox , r. p. 2011
, the astrophysical journal , 732 , 10 rodigas , t. j. , weinberger , a. , mamajek , e. e. , et al .
2015 , apj , 811 , 1 rodigas , t. j. , arriagada , p. , faherty , j. , et al .
2016 , apj , 818 , 1 rodriguez , d. r. , bessell , m. s. , zuckerman , b. , & kastner , j. h. 2011 , the astrophysical journal , 727 , 62 rodriguez , d. r. , zuckerman , b. , kastner , j. h. , et al .
2013 , apj , 774 , 101 rodriguez , d. r. , zuckerman , b. , melis , c. , & song , i. 2011 , the astrophysical journal , 732 , l29 rosotti , g. p. , juhasz , a. , booth , r. a. , & clarke , c. j. 2016 , mnras , 1 rouan , d. , riaud , p. , boccaletti , a. , clnet , y. , & labeyrie , a. 2000 , the publications of the astronomical society of the pacific , 112 , 1479 ryu , t. , sato , b. , kuzuhara , m. , et al .
2016 , arxiv , 1603.02017v1 sallum , s. , follette , k. b. , eisner , j. a. , et al .
2015 , nature , 527 , 342 sallum , s. , eisner , j. a. , close , l. m. , et al .
2015 , apj , 801 , 85 santos , n. c. , mayor , m. , naef , d. , et al .
2002 , a&a , 392 , 215 sartoretti , p. , brown , r. a. , & latham , d. w. 1998 , astronomy and saumon , d. , & marley , m. s. 2008 , the astrophysical journal , 689 , 1327 savransky , d. 2015 , apj , 800 , 100 schlieder , j. e. , lpine , s. , & simon , m. 2010 , the astronomical journal , 140 , 119 . 2012 , the astronomical journal , 143 , 80 . 2012 , the astronomical journal , 144 , 109 schlieder , j. e. , skemer , a. j. , maire , a .-
, et al . 2016 , apj , 818 , 1 schneider , a. c. , windsor , j. , cushing , m. c. , kirkpatrick , j. d. , & wright , e. l. 2016 , the astrophysical journal letters , 822 , 1 schnupp , c. , bergfors , c. , brandner , w. , et al .
2010 , arxiv , astro-ph.sr scholz , a. , jayawardhana , r. , muzic , k. , et al .
2012 , apj , 756 , 24 scholz , r .- d .
2010 , a&a , 515 , a92 schroeder , d. j. , golimowski , d. a. , brukardt , r. a. , et al .
2000 , the astronomical
serabyn , e. , mawet , d. , & burruss , r. 2010 , nature , 464 , 1018 shkolnik , e. , liu , m. c. , & reid , i. n. 2009 , the astrophysical journal , 699 , 649 shkolnik , e. l. , anglada - escud , g. , liu , m. c. , et al .
2012 , the astrophysical journal , 758 , 56 silk , j. 1977 , astrophysical journal , 214 , 152 skemer , a. j. , & close , l. m. 2011 , apj , 730 , 53 skemer , a. j. , marley , m. s. , hinz , p. m. , et al .
2014 , apj , 792 , 17 skemer , a. j. , hinz , p. , esposito , s. , et al .
2014 , in spie astronomical telescopes + instrumentation , ed .
e. marchetti , l. m. close , & j .-
veran ( spie ) , 91480l skemer , a. j. , hinz , p. , montoya , m. , et al .
2015 , in spie optical engineering + applications , ed .
s. shaklan ( spie ) , 96051d skemer , a. j. , morley , c. v. , zimmerman , n. t. , et al .
2016 , apj , 817 , 1 smith , b. a. , & terrile , r. j. 1984 , science snellen , i. , de kok , r. , birkby , j. l. , et al .
2015 , a&a , 576 , a59 snellen , i. a. g. , brandl , b. r. , de kok , r. j. , et al .
2014 , nature , 508 , 63 soderblom , d. r. 2010 , annu .
astrophys . , 48 , 581 soderblom , d. r. , hillenbrand , l. a. , jeffries , r. d. , mamajek , e. e. , & naylor , t. 2014 , and planets vi song , i. , zuckerman , b. , & bessell , m. s. 2003 , the astrophysical journal , 599 , 342 soummer , r. 2005 , the astrophysical journal , 618 , l161 soummer , r. , hagan , j. b. , pueyo , l. , et al .
2011 , the astrophysical journal , 741 , 55 soummer , r. , pueyo , l. , & larkin , j. 2012 , the astrophysical journal , 755 , l28 sozzetti , a. , giacobbe , p. , lattanzi , m. g. , et al .
2013 , monthly notices ras , 437 , 497 spergel , d. , gehrels , n. , baltay , c. , et al .
2015 , arxiv , 1 spiegel , d. s. , & burrows , a. 2012 , the astrophysical journal , 745 , 174 spiegel , d. s. , burrows , a. , & milsom , j. a. 2011 , the astrophysical journal , 727 , 57 stevenson , d. j. 1991 , in : annual review of astronomy and astrophysics .
29 ( a92 - 18081 05 - 90 ) .
palo alto , 29 , 163 stolker , t. , dominik , c. , avenhaus , h. , et al .
2016 , arxiv , 1603.00481v1 stone , j. m. , skemer , a. j. , kratter , k. m. , et al .
2016 , the astrophysical journal letters , 818 , 1 stumpf , m. b. , brandner , w. , joergens , v. , et al .
2010 , the astrophysical journal , 724 , 1 su , k. , rieke , g. h. , & misselt , k. a. 2005 , the astrophysical su , k. y. l. , rieke , g. h. , malhotra , r. , et al .
2013 , apj , 763 , 118 sudol , j. j. , & haghighipour , n. 2012 , apj , 755 , 38 tamura , m. 2016 , proceedings of the japan academy .
b : physical and biological sciences , 92 , 45 tamura , m. , seeds team , usuda , t. , tamura , m. , & ishii , m. 2009 , in exoplanets and disks : their formation and diversity : proceedings of the international conference , aa(national astronomical observatory of japan , mitaka , tokyo 181 - 8588 motohide.tamuraatnao.ac.jp ) ( aip ) , 1116 tanner , a. , beichman , c. , akeson , r. , et al .
2007 , the publications of the astronomical society of the pacific , 119 , 747 tanner , a. m. , gelino , c. r. , & law , n. m. 2010 , publications of the astronomical society of the pacific , 122 , 1195 terebey , s. , van buren , d. , matthews , k. , & padgett , d. l. 2000 , the astronomical journal , 119 , 2341 terebey , s. , van buren , d. , padgett , d. l. , hancock , t. , & brundage , m. 1998 , the astrophysical journal , 507 , l71 thalmann , c. , schmid , h. m. , boccaletti , a. , et al .
2008 , in spie astronomical telescopes + instrumentation , ed .
i. s. mclean & m. m. casali ( spie ) , 70143f thalmann , c. , carson , j. , janson , m. , et al .
2009 , the astrophysical journal , 707 , l123 thalmann , c. , grady , c. a. , goto , m. , et al .
2010 , the astrophysical journal letters , 718 , l87 thalmann , c. , usuda , t. , kenworthy , m. , et al .
2011 , the astrophysical journal , 732 , l34 thalmann , c. , desidera , s. , bonavita , m. , et al .
2014 , a&a , 572 , a91 thomas , s. , belikov , r. , & bendek , e. 2015 , apj , 810 , 1 tinney , c. g. , faherty , j. k. , kirkpatrick , j. d. , et al .
2014 , apj , 796 , 39 todorov , k. , luhman , k. l. , & mcleod , k. k. 2010 , the astrophysical journal , 714 , l84 todorov , k. o. , line , m. r. , pineda , j. e. , et al .
2015 , arxiv , 1504.00217v3 todorov , k. o. , luhman , k. l. , konopacky , q. m. , et al .
2014 , apj , 788 , 40 tokovinin , a. 2014 , the astronomical journal , 147 , 87 torres , c. a. o. , quast , g. r. , melo , c. h. f. , & sterzik , m. f. 2008 , in young nearby loose associations , ed . b. reipurth ( san francisco , ca : asp ) , 757 torres , g. 1999 , the publications of the astronomical society of the pacific , 111 , 169 traub , w. a. , & oppenheimer , b. r. 2010 , exoplanets , 111 traub , w. a. , belikov , r. , guyon , o. , et al .
2014 , in spie astronomical telescopes + instrumentation , ed .
j. m. oschmann , m. clampin , g. g. fazio , & h. a. macewen ( spie ) , 91430n trilling , d. e. , bryden , g. , beichman , c. a. , et al .
2008 , the astrophysical journal , 674 , 1086 troy , m. , dekany , r. g. , brack , g. , et al .
2000 , proc .
4007 , 4007 , 31 vigan , a. , gry , c. , salter , g. , et al .
2015 , monthly notices ras , 454 , 129 vigan , a. , langlois , m. , moutou , c. , & dohlen , k. 2008 , a&a , 489 , 1345 vigan , a. , moutou , c. , langlois , m. , et al . 2010 ,
monthly notices ras , 407 , 71 vigan , a. , patience , j. , marois , c. , et al .
2012 , a&a , 544 , 9 vigan , a. , bonnefoy , m. , ginski , c. , et al .
2016 , a&a , 587 , a55 wahhaj , z. , koerner , d. w. , ressler , m. e. , et al .
2003 , the astrophysical journal , 584 , l27 wahhaj , z. , liu , m. c. , biller , b. a. , et al .
2011 , the astrophysical journal , 729 , 139 . 2013 , apj , 779 , 80 wahhaj , z. , liu , m. c. , nielsen , e. l. , et al .
2013 , the astrophysical journal , 773 , 179 wahhaj , z. , liu , m. c. , biller , b. a. , et al .
2014 , arxiv , 1404.6525v1 wahhaj , z. , cieza , l. a. , mawet , d. , et al .
2015 , a&a , 581 , a24 wang , j. , xie , j .- w . ,
barclay , t. , & fischer , d. a. 2014 , apj , 783 , 4 white , r. j. , & ghez , a. m. 2001 , the astrophysical journal , 556 , 265 williams , j. p. , & cieza , l. a. 2011 , annu .
astrophys . , 49 , 67 wilner , d. j. , holman , m. j. , & kuchner , m. j. 2002 , the astrophysical winn , j. n. , & fabrycky , d. c. 2015 , annu .
astrophys . , 53 , 409 winters , j. g. , henry , t. j. , lurie , j. c. , et al .
2015 , the astronomical journal , 149 , 1 wllert , m. , brandner , w. , reffert , s. , et al .
2014 , a&a , 564 , a10 wright , j. t. , marcy , g. w. , howard , a. w. , et al .
2012 , apj , 753 , 160 wright , j. t. , fakhouri , o. , marcy , g. w. , et al .
2011 , publications of the astronomical society of the pacific , 123 , 412 wu , y .- l . , close , l. m. , males , j. r. , et al .
2015 , the astrophysical journal letters , 807 , 1 wyatt , m. c. 2008 , annu . rev .
astrophys . , 46 , 339 wyatt , m. c. , holmes , e. k. , pia , r. k. , et al . 1999 ,
apj , 527 , 918 wyatt , m. c. , kennedy , g. , sibthorpe , b. , et al .
2012 , monthly notices ras , 424 , 1206 yamamoto , k. , matsuo , t. , shibai , h. , et al .
2013 , pasj , 1306.3100 zahnle , k. j. , & marley , m. s. 2014 , apj , 797 , 41 zapatero osorio , m. r. , rebolo , r. , bihain , g. , et al .
2010 , the astrophysical journal , 715 , 1408 zhou , y. , apai , d. , schneider , g. h. , marley , m. s. , & showman , a. p. 2016 , apj , 818 , 1 zhou , y. , herczeg , g. j. , kraus , a. l. , metchev , s. , & cruz , k. l. 2014 , the astrophysical journal , 783 , l17 zhu , z. , dong , r. , stone , j. m. , & rafikov , r. r. 2015 , apj , 813 , 1 zimmerman , n. , brenner , d. , oppenheimer , b. r. , et al .
2011 , publications of the astronomical society of the pacific , 123 , 746 zimmerman , n. , oppenheimer , b. r. , hinkley , s. , et al .
2010 , the astrophysical journal , 709 , 733 zucker , s. , mazeh , t. , santos , n. c. , udry , s. , & mayor , m. 2003 , a&a , 404 , 775 . 2004 , a&a , 426 , 695 zuckerman , b. 2001 , annu .
astrophys . , 39 , 549 zuckerman , b. , rhee , j. h. , song , i. , & bessell , m. s. 2011 , the astrophysical journal , 732 , 61 zuckerman , b. , & song , i. 2004 , annu .
astrophys . , 42 , 685 zuckerman , b. , song , i. , bessell , m. s. , & webb , r. a. 2001 , the astrophysical journal , 562 , l87 zurlo , a. , vigan , a. , hagelberg , j. , et al .
2013 , a&a , 554 , a21 zurlo , a. , vigan , a. , galicher , r. , et al .
2016 , a&a , 587 , a57 lcccccccccc 51 eri b & 2 @xmath38 1 & 5.6 @xmath38 0.2 & 23 @xmath38 3 & 0.45 & 13 & t4.5t6 & yes & s & 1.75 & 1 , 2 , 3 + hd 95086 b & 5 @xmath38 2 & 4.96 @xmath38 0.10 & 17 @xmath38 4 & 0.6 & 56 & l / t : & yes & s & 1.6 & 47 + hr 8799 b & 5 @xmath38 1 & 5.1 @xmath38 0.1 & 40 @xmath38 5 & 1.7 & 68 & @xmath4l / tpec & yes & s & 1.5 & 811 + lkca 15 b & 6 @xmath38 4 & @xmath89 & 2 @xmath38 1 & 0.08 & 20 & @xmath89 & yes & s & 1.0 & 1215 + hr 8799 c & 7 @xmath38 2 & 4.7 @xmath38 0.1 & 40 @xmath38 5 & 0.95 & 38 & @xmath4l / tpec & yes & s & 1.5 & 811 + hr 8799 d & 7 @xmath38 2 & 4.7 @xmath38 0.2 & 40 @xmath38 5 & 0.62 & 24 & @xmath4l7pec & yes & s & 1.5 & 8 , 10 , 11 + hr 8799 e & 7 @xmath38 2 & 4.7 @xmath38 0.2 & 40 @xmath38 5 & 0.38 & 14 & @xmath4l7pec & yes & s & 1.5 & 10 , 11 , 16 + @xmath20 pic b & 12.7 @xmath38 0.3 & 3.78 @xmath38 0.03 & 23 @xmath38 3 & 0.4 & 9 & l1 & yes & s & 1.6 & 1720 + wd
0806 - 661 b & 7.5 @xmath38 1.5 & @xmath89 & 2000 @xmath38 500 & 130 & 2500 & y ? & no & s & 2.0 & 2123 + ross 458 c & 9 @xmath38 3 & 5.62 @xmath38 0.03 & 150800 & 102 & 1190 & t8.5pec & no & b & 0.6 , 0.09 & 2428 + roxs
42b b & 10 @xmath38 4 & 3.07 @xmath38 0.07 & 3 @xmath38 2 & 1.2 & 140 & l1 & yes & b & 0.89 , 0.36 & 2933 + hd 106906 b & 11 @xmath38 2 & 3.64 @xmath38 0.08 & 13 @xmath38 2 & 7.1 & 650 & l2.5 & no & b & 1.5 & 34 , 35 + gu psc b & 11 @xmath38 2 & 4.75 @xmath38 0.15 & 120 @xmath38 10 & 42 & 2000 & t3.5 & no & s & 0.30 & 36 + chxr 73 b & 13 @xmath38 6 & 2.85 @xmath38 0.14 & 2 @xmath38 1 & 1.3 & 210 & @xmath90m9.5 & no & s & 0.30 & 37 + sr12 c & 13 @xmath38 2 & 2.87 @xmath38 0.20 & 3 @xmath38 2 & 8.7 & 1100 & m9.0 & no & b & 1.0 , 0.5 & 31 , 38 + tyc 9486 - 927 - 1 b & 1215 & @xmath89 & 1045 & 217 & 6900 & l3 & no & s & 0.4 & 39 , 40 + 2m12073932 b & 5 @xmath38 2 & 4.68 @xmath38 0.05 & 10 @xmath38 3 & 0.8 & 41 & l3 & no & s & 0.024 & 4144 , 11 + 2m0441 + 2301 bb & 10 @xmath38 2 & 3.03 @xmath38 0.09 & 2 @xmath38 1 & 12/0.1 & 1800/15 & l1 & yes & b / s & 0.2 , 0.018 & 4547 + 1rxs j16092105 b & 14 @xmath38 2 & 3.36 @xmath38 0.09 & 11 @xmath38 2 & 2.2 & 330 & l2 & no & s & 0.85 & 4851 + 2m01035515 b & 1335 & 3.49 @xmath38 0.11 & 45 @xmath38 4 & 1.7 & 84 & @xmath89 & yes & b & 0.19 , 0.17 & 52 , 53 , 11 + 2m01222439 b & 1227 & 4.19 @xmath38 0.10 & 120 @xmath38 10 & 1.4 & 52 & l4 & no & s & 0.4 & 53 , 54 + 2m02193925 b & 14 @xmath38 1 & 3.84 @xmath38 0.05 & 45 @xmath38 4 & 4.0 & 156 & l4 & no & s & 0.11 & 55 + ab pic b & 1330 & 3.7 @xmath38 0.2 & 45 @xmath38 4 & 5.5 & 250 & l0 & no & s & 0.95 & 56 , 57 + cfbdsir j1458 + 1013 b & 520 & 6.74 @xmath38 0.19 & 10005000 & 0.1 & 2.6 & y0 : & yes & s & 0.010.04 & 58 , 59 + dh tau b & 822 & 2.71 @xmath38 0.12 & 2 @xmath38 1 & 2.3 & 340 & m9.25 & no & s & 0.5 & 60 , 37 , 15 + fomalhaut b & @xmath282 & @xmath89 & 440 @xmath38 40 & 13 & 119 & @xmath89 & yes & s & 1.92 & 6164 + fu tau b & @xmath416 & 2.60 & 2 @xmath38 1 & 5.7 & 800 & m9.25 & no & s & 0.05 & 65 + fw tau b & @xmath410100 & @xmath89 & 2 @xmath38 1 & 2.3 & 330 & pec & no & b & 0.3 , 0.3 & 29 , 31 , 66 + g196 - 3 b & 1225 & 3.8 @xmath38 0.2 & 2085 & 16 & 400 & l3 & no & s & 0.43 & 6769 , 53 , 44 + gj 504 b & 330 & 6.13 @xmath38 0.03 & 1006500 & 2.5 & 44 & t : & yes & s & 1.16 & 7073 + gj 758 b & 1040 & 6.1 @xmath38 0.2 & 10006000 & 1.9 & 29 & t8 : & yes & s & 1.0 & 7477 + gsc 6214 - 210 b & 15 @xmath38 2 & 3.1 @xmath38 0.1 & 11 @xmath38 2 & 2.2 & 320 & m9.5 & no & s & 0.9 & 50 , 31 , 78 , 79 + hd 100546 b & @xmath410 @xmath38 5 & @xmath89 & 510 & 0.48 & 53 & @xmath89 & no & s & 2.4 & 8082 + hd 100546 c & @xmath320 & @xmath89 & 510 & 0.13 & 13 & @xmath89 & no & s & 2.4 & 83 + hd 203030 b & 1230 & 4.64 @xmath38 0.07 & 130400 & 12 & 490 & l7.5 & yes & s & 0.95 & 84 , 85 + hn peg b & 1231 & 4.77 @xmath38 0.03 & 300 @xmath38 200 & 43 & 800 & t2.5 & no & s & 1.07 & 85 , 86 + @xmath21 and b & 1266 & 3.76 @xmath38 0.06 & 40300 & 1.1 & 55 & l1 & no & s & 2.8 & 8790 + lkca 15 c & @xmath310 & @xmath89 & 2 @xmath38 1 & 0.08 & 15 & @xmath89 & yes & s & 1.0 & 14 , 15 + lkca 15 d & @xmath310 & @xmath89 & 2 @xmath38 1 & 0.09 & 18 & @xmath89 & yes & s & 1.0 & 14 , 15 + lp 261 - 75 b & 1226 & 4.43 @xmath38 0.09 & 100200 & 14 & 450 & l4.5 & no & s & 0.22 & 91 , 53 + roxs12 b & 16 @xmath38 4 & @xmath89 & 8 @xmath38 3 & 1.8 & 210 & @xmath89 & yes & s & 0.9 & 29 , 33 + sdss2249 + 0044 a & 1260 & 3.9 @xmath38 0.3 & 20300 & 0.3/49 & 17/2600 & l3 & no & s / s & @xmath89 & 92 + sdss2249 + 0044 b & 852 & 4.2 @xmath38 0.3 & 20300 & 0.3 & 17 & l5 & no & s & 0.03 & 92 + vhs12561257 b & 1021 & 5.05 @xmath38 0.22 & 150300 & 8.1 & 102 & l7 & no & b & 0.07 , 0.07 & 93 , 94 + wise j0146 + 4234 b & 416 & 7.01 @xmath38 0.22 & 1000 - 10000 & 0.09 & 1 & y0 & yes & s & 0.0050.016 & 95 + wise j1217 + 1626 b & 520 & 6.79 @xmath38 0.18 & 10005000 & 0.08 & 8 & y0 & no & s & 0.010.04 & 96 + lcccccccc hd 19467 b & 1.65 & 51 & bd & 0.050 & t5.5 & no & 1.37 @xmath38 0.09 & 1 , 2 + hd 4747 b & 0.60 & 11.3 & bd & 0.057 & @xmath89 & no & 30.7 @xmath38 0.6 & 3 + hr 7672 b & 0.5 & 18 & bd & 0.066 & l4.5 & no & 24 @xmath38 0.6 & 4 , 5 + hip 71898 b & 3.0 & 30 & lms & 0.074 & l0 & no & 8.6 @xmath38 0.4 & 6 , 7 , 8 + hd 5608 b & 0.6 & 34 & lms & 0.10 & @xmath89 & yes & 5.51 @xmath38 0.45 & 9 + hd 68017 b & 0.59 & 13 & lms & 0.15 & [ m5 ] & no & 16.3 @xmath38 0.9 & 10 + hd 7449 b & 0.54 & 18 & lms & 0.2 & [ m4.5 ] & yes & @xmath89 & 11 + gl 15 b & 41 & 146 & lms & 0.2 & m3.5 & yes & 0.27 @xmath38 0.09 & 12 + hd 104304 b & 1.0 & 13 & lms & 0.21 & [ m4 ] & no & @xmath89 & 13 + hd 109272 b & 1.2 & 59 & lms & 0.28 & @xmath89 & no & 2.4 @xmath38 0.2 & 9 + hd 71881 b & 0.85 & 35 & lms & 0.29 & [ m3 ] & no & 10.3 @xmath38 0.2 & 10 + kepler-444 bc & 1.8 & 66 & lms & 0.29 + 0.25 & [ m ] & yes & 7.8 @xmath38 0.5 & 14 + hd 164509 b & 0.75 & 39 & lms & 0.33 & [ m ] & yes & 3.3 @xmath38 0.5 & 15 + hd 126614 b & 0.49 & 33 & lms & 0.33 & [ m ] & yes & 16.2 @xmath38 0.2 & 16 + hd 197037 & 3.7 & 24 & lms & 0.34 & [ m ] & yes & 1.9 @xmath38 0.3 & 17 + hat - p-10 b & 0.34 & 42 & lms & 0.36 & [ m ] & yes & 5.1 @xmath38 1.4 & 18 , 19 + hd 41004 a & 0.5 & 22 & lms & 0.4 & k1 & yes & 105 @xmath38 10 & 20 , 21 , 22 + hd 77407 b & 1.7 & 50 & lms & 0.4 & [ m2 ] & no & @xmath89 & 23 + @xmath91 cep b & 0.88 & 20 & lms & 0.41 & [ m4 ] & yes & @xmath89 & 24 + hd 8375 c & 0.3 & 17 & lms & 0.5 & [ m1 ] & no & 52.9 @xmath38 1.9 & 25 + hd 196885 b & 0.7 & 23 & lms & 0.5 & m1 & yes & @xmath89 & 26 , 27 + hd 114174 b & 0.69 & 18 & wd & 0.54 & @xmath89 & no & 61.1 @xmath38 0.1 & 28 , 29 + hd 8049 b & 1.5 & 50 & wd & 0.56 & @xmath89 & no & @xmath89 & 30 + hd 195109 b & 2.4 & 92 & lms & 0.58 & @xmath89 & yes & 1.9 & 15 + gl 86 b & 2.3 & 28 & wd & 0.59 & dq6 & yes & 131 & 3136 + procyon b & 4.3 & 15 & wd & 0.59 & dqz & no & @xmath89 & 37 , 38 , 39 + @xmath91 hya b & 1.6 & 66 & lms & 0.61 & @xmath89 & no & 4.1 @xmath38 0.2 & 9 + hd 53665 b & 0.14 & 103 & lms & 0.7 & [ k7 ] & no & 5.3 @xmath38 0.3 & 40 + lccccc ba & 110 & 7.7@xmath92% & 2.8@xmath2% & 3.5@xmath93% & @xmath36.4% + fgk & 155 & @xmath36.8% & @xmath34.1% & @xmath35.8% & @xmath35.1% + m & 118 & @xmath34.2% & @xmath33.9% & @xmath35.4% & @xmath37.3% + all stars & 384 & 0.8@xmath94% & 0.6@xmath1% & 0.8@xmath95% & @xmath32.1%
lccccccc 1998 oct 30 & 1998.828 & @xmath26/nicmos & 1.721 @xmath38 0.012 & 55.1 @xmath38 0.4 & 1.411 @xmath38 0.009 & 0.986 @xmath38 0.009 & @xcite + 1998 oct 30 & 1998.828 & @xmath26/nicmos & 1.738 @xmath38 0.018 & 54.7 @xmath38 0.4 & [ 1.418 @xmath38 0.016 ] & [ 1.00 @xmath38 0.014 ] & @xcite + 2002 jul 19 & 2002.545 & subaru / ciao & [ 1.74 @xmath38 0.03 ] & [ 58.2 @xmath38 0.6 ] & 1.481 @xmath38 0.023 & 0.919 @xmath38 0.017 & @xcite + 2004 jul 14 & 2004.534 & keck / nirc2 & [ 1.716 @xmath38 0.007 ] & [ 59.0 @xmath38 0.17 ] & 1.471 @xmath38 0.005 & 0.884 @xmath38 0.005 & @xcite + 2005 jul 15 & 2005.534 & keck / nirc2 & [ 1.724 @xmath38 0.007 ] & [ 60.2 @xmath38 0.17 ] & 1.496 @xmath38 0.005 & 0.856 @xmath38 0.005 & @xcite + 2007 aug 02 & 2007.583 & keck / nirc2 & [ 1.726 @xmath38 0.004 ] & [ 61.8 @xmath38 0.1 ] & 1.522 @xmath38 0.003 & 0.815 @xmath38 0.003 & @xcite + 2007 aug 02 & 2007.583 & keck / nirc2 & [ 1.721 @xmath38 0.004 ] & [ 60.90 @xmath38 0.10 ] & 1.504 @xmath38 0.003 & 0.837 @xmath38 0.003 & @xcite + 2007 oct 25 & 2007.813 & gemini - n / niri & [ 1.713 @xmath38 0.007 ] & [ 62.0 @xmath38 0.17 ] & 1.512 @xmath38 0.005 & 0.805 @xmath38 0.005 & @xcite + 2007 oct 25 & 2007.813 & keck / nirc2 & [ 1.717 @xmath38 0.010 ] & [ 60.9 @xmath38 0.2 ] & 1.500 @xmath38 0.007 & 0.836 @xmath38 0.007 & @xcite + 2008 jul 11 & 2008.523 & keck / nirc2 & [ 1.723 @xmath38 0.006 ] & [ 62.4 @xmath38 0.13 ] & 1.527 @xmath38 0.004 & 0.799 @xmath38 0.004 & @xcite + 2008 aug 12 & 2008.613 & keck / nirc2 & [ 1.724 @xmath38 0.003 ] & [ 62.3 @xmath38 0.07 ] & 1.527 @xmath38 0.002 & 0.801 @xmath38 0.002 & @xcite + 2008 sep 18 & 2008.715 & keck / nirc2 & [ 1.724 @xmath38 0.004 ] & [ 62.4 @xmath38 0.10 ] & 1.528 @xmath38 0.003 & 0.798 @xmath38 0.003 & @xcite + 2008 sep 18 & 2008.715 & keck / nirc2 & [ 1.723 @xmath38 0.006 ] & [ 61.65 @xmath38 0.13 ] & 1.516 @xmath38 0.004 & 0.818 @xmath38 0.004 & @xcite + 2008 nov 21 & 2008.890 & mmt / clio & [ 1.728 @xmath38 0.017 ] & [ 63.2 @xmath38 0.4 ] & 1.54 @xmath38 0.01 & 0.780 @xmath38 0.014 & @xcite + 2008 nov 21 & 2008.890 & mmt / clio & [ 1.73 @xmath38 0.03 ] & [ 62.5 @xmath38 0.7 ] & 1.532 @xmath38 0.02 & 0.796 @xmath38 0.02 & @xcite + 2009 aug 15 & 2009.619 & subaru / ircs & [ 1.725 @xmath38 0.014 ] & [ 62.9 @xmath38 0.3 ] & 1.536 @xmath38 0.01 & 0.785 @xmath38 0.01 & @xcite + 2009 sep 12 & 2009.695 & mmt / clio & [ 1.72 @xmath38 0.04 ] & [ 63.2 @xmath38 1.0 ] & 1.538 @xmath38 0.03 & 0.777 @xmath38 0.03 & @xcite + 2009 oct 06 & 2009.761 & vlt / naco & [ 1.74 @xmath38 0.03 ] & [ 62.0 @xmath38 0.7 ] & 1.535 @xmath38 0.02 & 0.816 @xmath38 0.02 & @xcite + 2009 oct 08 & 2009.767 & vlt / naco & [ 1.720 @xmath38 0.010 ] & [ 62.9 @xmath38 0.2 ] & 1.532 @xmath38 0.007 & 0.783 @xmath38 0.007 & @xcite + 2009 jul 11 & 2009.523 & palomar / wcs & 1.73 @xmath38 0.10 & 62.9 @xmath38 1.5 & 1.54 @xmath38 0.04 & 0.79 @xmath38 0.06 & @xcite + 2009 jul 30 & 2009.575 & keck / nirc2 & [ 1.721 @xmath38 0.006 ] & [ 62.42 @xmath38 0.13 ] & 1.526 @xmath38 0.004 & 0.797 @xmath38 0.004 & @xcite + 2009 aug 01 & 2009.580 & keck / nirc2 & [ 1.725 @xmath38 0.010 ] & [ 62.6 @xmath38 0.2 ] & 1.531 @xmath38 0.007 & 0.794 @xmath38 0.007 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 1.74 @xmath38 0.03 ] & [ 62.5 @xmath38 0.6 ] & 1.54 @xmath38 0.019 & 0.80 @xmath38 0.019 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 1.719 @xmath38 0.014 ] & [ 62.5 @xmath38 0.3 ] & 1.524 @xmath38 0.010 & 0.795 @xmath38 0.010 & @xcite + 2010 jul 13 & 2010.528 & keck / nirc2 & [ 1.720 @xmath38 0.007 ] & [ 62.93 @xmath38 0.17 ] & 1.532 @xmath38 0.005 & 0.783 @xmath38 0.005 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 1.716 @xmath38 0.021 ] & [ 63.5 @xmath38 0.5 ] & 1.535 @xmath38 0.015 & 0.766 @xmath38 0.015 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 1.717 @xmath38 0.007 ] & [ 64.2 @xmath38 0.17 ] & 1.546 @xmath38 0.005 & 0.748 @xmath38 0.005 & @xcite + 2011 jul 21 & 2011.550 & keck / nirc2 & [ 1.719 @xmath38 0.007 ] & [ 63.69 @xmath38 0.17 ] & 1.541 @xmath38 0.005 & 0.762 @xmath38 0.005 & @xcite + 2011 oct 16 & 2011.789 & lbt / pisces & [ 1.741 @xmath38 0.016 ] & [ 65.1 @xmath38 0.4 ] & 1.579 @xmath38 0.011 & 0.734 @xmath38 0.011 & @xcite + 2011 nov 09 & 2011.854 & lbt / pisces & [ 1.708 @xmath38 0.016 ] & [ 64.9 @xmath38 0.4 ] & 1.546 @xmath38 0.011 & 0.725 @xmath38 0.011 & @xcite + 2012 jun 14 & 2012.452 & palomar / p1640 & [ 1.715 @xmath38 0.007 ] & [ 65.7 @xmath38 0.17 ] & 1.563 @xmath38 0.005 & 0.706 @xmath38 0.005 & @xcite + 2012 jul 22 & 2012.556 & keck / nirc2 & [ 1.716 @xmath38 0.007 ] & [ 64.20 @xmath38 0.17 ] & 1.545 @xmath38 0.005 & 0.747 @xmath38 0.005 & @xcite + 2012 oct 26 & 2012.819 & keck / nirc2 & [ 1.718 @xmath38 0.006 ] & [ 64.37 @xmath38 0.13 ] & 1.549 @xmath38 0.004 & 0.743 @xmath38 0.004 & @xcite + 2013 oct 16 & 2013.789 & keck / nirc2 & [ 1.706 @xmath38 0.031 ] & [ 64.9 @xmath38 0.7 ] & 1.545 @xmath38 0.022 & 0.724 @xmath38 0.022 & @xcite + 2013 oct 21 & 2013.802 & lbt / lmircam & 1.717 @xmath38 0.013 & 65.46 @xmath38 0.44 & 1.562 @xmath38 0.008 & 0.713 @xmath38 0.013 & @xcite + 2014 jul 13 & 2014.528 & vlt / sphere & [ 1.722 @xmath38 0.008 ] & [ 65.8 @xmath38 1.7 ] & 1.570 @xmath38 0.006 & 0.707 @xmath38 0.006 & @xcite + 2014 jul 17 & 2014.539 & keck / nirc2 & [ 1.720 @xmath38 0.018 ] & [ 65.1 @xmath38 0.4 ] & 1.560 @xmath38 0.013 & 0.725 @xmath38 0.013 & @xcite + 2014 dec 0408 & 2014.93 & vlt / sphere & [ 1.724 @xmath38 0.007 ] & [ 65.9 @xmath38 0.17 ] & 1.574 @xmath38 0.005 & 0.703 @xmath38 0.005 & @xcite + 2014 dec 0509 & 2014.93 & vlt / sphere & 1.729 @xmath38 0.007 & 65.93 @xmath38 0.15 & [ 1.579 @xmath38 0.007 ] & [ 0.705 @xmath38 0.007 ] & @xcite + 1998 oct 30 & 1998.828 & @xmath26/nicmos & 0.966 @xmath38 0.022 & 300 @xmath38 0.8 & [ 0.84 @xmath38 0.02 ] & [ 0.483 @xmath38 0.016 ] & @xcite + 2004 jul 14 & 2004.534 & keck / nirc2 & [ 0.960 @xmath38 0.007 ] & [ 309.6 @xmath38 0.3 ] & 0.739 @xmath38 0.005 & 0.612 @xmath38 0.005 & @xcite + 2005 jul 15 & 2005.534 & keck / nirc2 & [ 0.951 @xmath38 0.007 ] & [ 311.5 @xmath38 0.3 ] & 0.713 @xmath38 0.005 & 0.630 @xmath38 0.005 & @xcite + 2007 aug 02 & 2007.583 & keck / nirc2 & [ 0.952 @xmath38 0.007 ] & [ 315.1 @xmath38 0.3 ] & 0.672 @xmath38 0.005 & 0.674 @xmath38 0.005 & @xcite + 2007 aug 02 & 2007.583 & keck / nirc2 & [ 0.957 @xmath38 0.006 ] & [ 314.5 @xmath38 0.2 ] & 0.683 @xmath38 0.004 & 0.671 @xmath38 0.004 & @xcite + 2007 oct 25 & 2007.813 & gemini - n / niri & [ 0.958 @xmath38 0.007 ] & [ 315.3 @xmath38 0.3 ] & 0.674 @xmath38 0.005 & 0.681 @xmath38 0.005 & @xcite + 2007 oct 25 & 2007.813 & keck / nirc2 & [ 0.957 @xmath38 0.010 ] & [ 314.9 @xmath38 0.4 ] & 0.678 @xmath38 0.007 & 0.676 @xmath38 0.007 & @xcite + 2008 jul 11 & 2008.523 & keck / nirc2 & [ 0.961 @xmath38 0.006 ] & [ 316.8 @xmath38 0.2 ] & 0.658 @xmath38 0.004 & 0.701 @xmath38 0.004 & @xcite + 2008 aug 12 & 2008.613 & keck / nirc2 & [ 0.964 @xmath38 0.003 ] & [ 317.06 @xmath38 0.12 ] & 0.657 @xmath38 0.002 & 0.706 @xmath38 0.002 & @xcite + 2008 sep 18 & 2008.715 & keck / nirc2 & [ 0.964 @xmath38 0.004 ] & [ 317.06 @xmath38 0.18 ] & 0.657 @xmath38 0.003 & 0.706 @xmath38 0.003 & @xcite + 2008 sep 18 & 2008.715 & keck / nirc2 & [ 0.959 @xmath38 0.004 ] & [ 316.27 @xmath38 0.18 ] & 0.663 @xmath38 0.003 & 0.693 @xmath38 0.003 & @xcite + 2008 nov 21 & 2008.890 & mmt / clio & [ 0.921 @xmath38 0.03 ] & [ 316.8 @xmath38 1.1 ] & 0.631 @xmath38 0.015 & 0.671 @xmath38 0.02 & @xcite + 2008 nov 21 & 2008.890 & mmt / clio & [ 0.958 @xmath38 0.03 ] & [ 316.9 @xmath38 1.2 ] & 0.654 @xmath38 0.02 & 0.700 @xmath38 0.02 & @xcite + 2009 jan 09 & 2009.022 & mmt / clio & [ 0.90 @xmath38 0.04 ] & [ 317.4 @xmath38 1.9 ] & 0.612 @xmath38 0.03 & 0.665 @xmath38 0.03 & @xcite + 2009 jul 11 & 2009.523 & palomar / wcs & 1.00 @xmath38 0.06 & 315.9 @xmath38 1.5 & 0.70 @xmath38 0.05 & 0.72 @xmath38 0.05 & @xcite + 2009 jul 30 & 2009.575 & keck / nirc2 & [ 0.957 @xmath38 0.006 ] & [ 318.1 @xmath38 0.2 ] & 0.639 @xmath38 0.004 & 0.712 @xmath38 0.004 & @xcite + 2009 aug 01 & 2009.580 & keck / nirc2 & [ 0.962 @xmath38 0.013 ] & [ 318.7 @xmath38 0.5 ] & 0.635 @xmath38 0.009 & 0.722 @xmath38 0.009 & @xcite + 2009 sep 12 & 2009.695 & mmt / clio & [ 0.96 @xmath38 0.03 ] & [ 319.2 @xmath38 1.2 ] & 0.625 @xmath38 0.02 & 0.725 @xmath38 0.02 & @xcite + 2009 sep 12 & 2009.695 & mmt / clio & [ 0.942 @xmath38 0.04 ] & [ 317.7 @xmath38 1.8 ] & 0.634 @xmath38 0.03 & 0.697 @xmath38 0.03 & @xcite + 2009 oct 06 & 2009.761 & vlt / naco & [ 0.940 @xmath38 0.06 ] & [ 317 @xmath38 2 ] & 0.636 @xmath38 0.04 & 0.692 @xmath38 0.04 & @xcite + 2009 oct 08 & 2009.767 & vlt / naco & [ 0.952 @xmath38 0.010 ] & [ 318.8 @xmath38 0.4 ] & 0.627 @xmath38 0.007 & 0.716 @xmath38 0.007 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 0.957 @xmath38 0.018 ] & [ 318.8 @xmath38 0.8 ] & 0.63 @xmath38 0.013 & 0.72 @xmath38 0.013 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 0.961 @xmath38 0.013 ] & [ 318.5 @xmath38 0.5 ] & 0.636 @xmath38 0.009 & 0.720 @xmath38 0.009 & @xcite + 2010 jul 13 & 2010.528 & keck / nirc2 & [ 0.956 @xmath38 0.006 ] & [ 319.6 @xmath38 0.2 ] & 0.619 @xmath38 0.004 & 0.728 @xmath38 0.004 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 0.960 @xmath38 0.017 ] & [ 320.8 @xmath38 0.7 ] & 0.607 @xmath38 0.012 & 0.744 @xmath38 0.012 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 0.949 @xmath38 0.007 ] & [ 320.9 @xmath38 0.3 ] & 0.598 @xmath38 0.005 & 0.737 @xmath38 0.005 & @xcite + 2011 jul 21 & 2011.550 & keck / nirc2 & [ 0.955 @xmath38 0.006 ] & [ 321.5 @xmath38 0.2 ] & 0.595 @xmath38 0.004 & 0.747 @xmath38 0.004 & @xcite + 2011 oct 16 & 2011.789 & lbt / pisces & [ 0.938 @xmath38 0.014 ] & [ 323.3 @xmath38 0.6 ] & 0.561 @xmath38
0.010 & 0.752 @xmath38 0.010 & @xcite + 2011 nov 09 & 2011.854 & lbt / pisces & [ 0.960 @xmath38 0.014 ] & [ 323.0 @xmath38 0.6 ] & 0.578 @xmath38 0.010 & 0.767
@xmath38 0.010 & @xcite + 2012 jun 14 & 2012.452 & palomar / p1640 & [ 0.947 @xmath38 0.006 ] & [ 323.9 @xmath38 0.2 ] & 0.558 @xmath38 0.004 & 0.765 @xmath38 0.004 & @xcite + 2012 jul 22 & 2012.556 & keck / nirc2 & [ 0.956 @xmath38
0.007 ] & [ 322.8 @xmath38 0.3 ] & 0.578 @xmath38 0.005 & 0.761 @xmath38 0.005 & @xcite + 2012 oct 26 & 2012.819 & keck / nirc2 & [ 0.958 @xmath38 0.004 ] & [ 323.32 @xmath38 0.18 ] & 0.572 @xmath38 0.003 & 0.768 @xmath38 0.003 & @xcite + 2013 oct 16 & 2013.789 & keck / nirc2 & [ 0.953 @xmath38 0.031 ] & [ 325.3 @xmath38 1.3 ] & 0.542 @xmath38 0.022 & 0.784 @xmath38 0.022 & @xcite + 2013 oct 21 & 2013.802 & lbt / lmircam & 0.951 @xmath38 0.014 & 325.5 @xmath38 0.9 & 0.538 @xmath38 0.006 & 0.784 @xmath38 0.013 & @xcite + 2014 jul 13 & 2014.528 & vlt / sphere & [ 0.948 @xmath38 0.006 ] & [ 326.6 @xmath38 0.2 ] & 0.522 @xmath38 0.004 & 0.791 @xmath38 0.004 & @xcite + 2014 jul 17 & 2014.539 & keck / nirc2 & [ 0.964 @xmath38 0.018 ] & [ 325.9 @xmath38 0.8 ] & 0.540 @xmath38 0.013 & 0.799 @xmath38 0.013 & @xcite + 2014 dec 0408 & 2014.93 & vlt / sphere & [ 0.951 @xmath38 0.006 ] & [ 327.0 @xmath38 0.2 ] & 0.518 @xmath38 0.004 & 0.797 @xmath38 0.004 & @xcite + 2014 dec 0509 & 2014.93 & vlt / sphere & 0.947 @xmath38 0.007 & 327.8 @xmath38 0.4 & [ 0.504 @xmath38 0.007 ] & [ 0.802 @xmath38 0.007 ] & @xcite + 1998 oct 30 & 1998.828 & @xmath26/nicmos & 0.549 @xmath38 0.028 & 166 @xmath38 3 & [ 0.13 @xmath38 0.03 ] & [ 0.53 @xmath38 0.03 ] & @xcite + 2005 jul 15 & 2005.534 & keck / nirc2 & [ 0.585 @xmath38 0.014 ] & [ 188.6 @xmath38 1.0 ] & 0.087 @xmath38 0.010 & 0.578 @xmath38 0.010 & @xcite + 2007 aug 02 & 2007.583 & keck / nirc2 & [ 0.613 @xmath38 0.011 ] & [ 196.1 @xmath38 0.7 ] & 0.170 @xmath38 0.008 & 0.589 @xmath38 0.008 & @xcite + 2007 aug 02 & 2007.583 & keck / nirc2 & [ 0.615 @xmath38 0.007 ] & [ 196.9 @xmath38 0.5 ] & 0.179 @xmath38 0.005 &
0.588 @xmath38 0.005 & @xcite + 2007 oct 25 & 2007.813 & keck / nirc2 & [ 0.614 @xmath38 0.014 ] & [ 196.5 @xmath38 0.9 ] & 0.175 @xmath38 0.010 & 0.589 @xmath38 0.010 & @xcite + 2008 jul 11 & 2008.523 & keck / nirc2 & [ 0.618 @xmath38 0.006 ] & [ 199.7 @xmath38 0.4 ] & 0.208 @xmath38 0.004 & 0.582
@xmath38 0.004 & @xcite + 2008 aug 12 & 2008.613 & keck / nirc2 & [ 0.621 @xmath38 0.003 ] & [ 200.36 @xmath38 0.18 ] & 0.216 @xmath38 0.002 & 0.582 @xmath38 0.002 & @xcite + 2008 sep 18 & 2008.715 & keck / nirc2 & [ 0.621 @xmath38 0.004 ] & [ 200.4 @xmath38 0.3 ] & 0.216 @xmath38 0.003 & 0.582 @xmath38 0.003 & @xcite + 2008 sep 18 & 2008.715 & keck / nirc2 & [ 0.622 @xmath38 0.006 ] & [ 199.0 @xmath38 0.4 ] & 0.202 @xmath38 0.004 & 0.588 @xmath38 0.004 & @xcite + 2008 nov 21 & 2008.890 & mmt / clio & [ 0.68 @xmath38 0.02 ] &
[ 198.9 @xmath38 1.7 ] & 0.22 @xmath38 0.021 & 0.644 @xmath38 0.013 & @xcite + 2008 nov 21 & 2008.890 & mmt / clio & [ 0.646 @xmath38 0.03 ] & [ 199.6 @xmath38 1.8 ] & 0.217 @xmath38 0.02 & 0.608 @xmath38 0.02 & @xcite + 2009 jul 11 & 2009.523 & palomar / wcs & 0.64 @xmath38 0.04 & 197.5 @xmath38 1.5 & 0.19 @xmath38 0.07 & 0.61 @xmath38 0.03 & @xcite + 2009 jul 30 & 2009.575 & keck / nirc2 & [ 0.624 @xmath38 0.004 ] & [ 202.3 @xmath38 0.3 ] & 0.237 @xmath38 0.003 & 0.577 @xmath38 0.003 & @xcite + 2009 aug 01 & 2009.580 & keck / nirc2 & [ 0.622 @xmath38 0.010 ] & [ 203.7 @xmath38 0.6 ] & 0.250 @xmath38 0.007 & 0.570 @xmath38 0.007 & @xcite + 2009 sep 12 & 2009.695 & mmt / clio & [ 0.65 @xmath38 0.04 ] & [ 205 @xmath38 3 ] & 0.28 @xmath38 0.03 & 0.59 @xmath38 0.03 & @xcite + 2009 oct 06 & 2009.761 & vlt / naco & [ 0.66 @xmath38 0.10 ] & [ 204 @xmath38 6 ] & 0.270 @xmath38 0.07 & 0.600 @xmath38 0.07 & @xcite + 2009 oct 08 & 2009.767 & vlt / naco & [ 0.634 @xmath38 0.010 ] & [ 202.4 @xmath38 0.6 ] & 0.241 @xmath38 0.007 & 0.586 @xmath38 0.007 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 0.63 @xmath38 0.02 ] & [ 202.5 @xmath38 1.3 ] & 0.24 @xmath38 0.014 & 0.58 @xmath38 0.014 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 0.626 @xmath38 0.010 ] & [ 203.7 @xmath38 0.6 ] & 0.251 @xmath38 0.007 & 0.573 @xmath38 0.007 & @xcite + 2010 jul 13 & 2010.528 & keck / nirc2 & [ 0.634 @xmath38 0.006 ] & [ 204.7 @xmath38 0.4 ] & 0.265 @xmath38 0.004 & 0.576 @xmath38 0.004 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 0.634 @xmath38 0.018 ] & [ 207.8 @xmath38 1.2 ] & 0.296 @xmath38 0.013 & 0.561 @xmath38 0.013 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 0.634 @xmath38 0.007 ] & [ 206.5 @xmath38 0.6 ] & 0.283 @xmath38 0.005 & 0.567 @xmath38 0.005 & @xcite + 2011 jul 21 & 2011.550 & keck / nirc2 & [ 0.638 @xmath38 0.007 ] & [ 208.3 @xmath38 0.4 ] & 0.303 @xmath38 0.005 & 0.562 @xmath38 0.005 & @xcite + 2011 oct 16 & 2011.789 & lbt / pisces & [ 0.637 @xmath38 0.014 ] & [ 208.0 @xmath38 0.9 ] & 0.299 @xmath38 0.010 & 0.563 @xmath38 0.010 & @xcite + 2011 nov 09 & 2011.854 & lbt / pisces & [ 0.635 @xmath38 0.014 ] & [ 210.2 @xmath38 0.9 ] & 0.320 @xmath38 0.010 & 0.549 @xmath38 0.010 & @xcite + 2012 jun 14 & 2012.452 & palomar / p1640 & [ 0.620 @xmath38 0.008 ] & [ 211.4 @xmath38 0.6 ] & 0.323 @xmath38 0.006 & 0.529 @xmath38 0.006 & @xcite + 2012 jul 22 & 2012.556 & keck / nirc2 & [ 0.650 @xmath38 0.007 ] & [ 211.4 @xmath38 0.4 ] & 0.339 @xmath38 0.005 & 0.555 @xmath38 0.005 & @xcite + 2013 oct 16 & 2013.789 & keck / nirc2 & [ 0.647 @xmath38 0.023 ] & [ 216.2 @xmath38 1.4 ] & 0.382 @xmath38 0.016 & 0.522 @xmath38 0.016 & @xcite + 2013 oct 21 & 2013.802 & lbt / lmircam & 0.657 @xmath38 0.013 & 215.0 @xmath38 1.2 & 0.377 @xmath38 0.007 & 0.538 @xmath38 0.011 & @xcite + 2012 oct 26 & 2012.819 & keck / nirc2 & [ 0.648 @xmath38 0.006 ] & [ 212.3 @xmath38 0.4 ] & 0.346 @xmath38 0.004 & 0.548 @xmath38 0.004 & @xcite + 2014 jul 13 & 2014.528 & vlt / sphere & [ 0.658 @xmath38 0.008 ] & [ 216.3 @xmath38 0.5 ] & 0.390 @xmath38 0.005 & 0.530 @xmath38 0.006 & @xcite + 2014 jul 17 & 2014.539 & keck / nirc2 & [ 0.667 @xmath38 0.016 ] & [ 216.8 @xmath38 0.9 ] & 0.400 @xmath38 0.011 & 0.534 @xmath38 0.011 & @xcite + 2014 aug 12 & 2014.611 & vlt / sphere & [ 0.658 @xmath38 0.006 ] & [ 216.5 @xmath38 0.3 ] & 0.391 @xmath38 0.004 & 0.529 @xmath38 0.004 & @xcite + 2014 dec 0408 & 2014.93 & vlt / sphere & [ 0.660 @xmath38 0.006 ] & [ 217.5 @xmath38 0.3 ] & 0.402 @xmath38 0.004 & 0.523 @xmath38 0.004 & @xcite + 2014 dec 0509 & 2014.93 & vlt / sphere & 0.654 @xmath38 0.007 & 217.6 @xmath38 0.8 & [ 0.400 @xmath38 0.009 ] & [ 0.518 @xmath38 0.008 ] & @xcite + 2009 jul 30 & 2009.575 & keck / nirc2 & [ 0.372 @xmath38 0.010 ] & [ 235.4 @xmath38 1.1 ] & 0.306 @xmath38 0.007 & 0.211 @xmath38 0.007 & @xcite + 2009 jul 31 & 2009.578 & keck / nirc2 & [ 0.369 @xmath38 0.03 ] & [ 234 @xmath38 3 ] & 0.299 @xmath38 0.019 & 0.217 @xmath38 0.019 & @xcite + 2009 aug 01 & 2009.580 & keck / nirc2 & [ 0.368 @xmath38 0.018 ] & [ 235 @xmath38 2 ] & 0.303 @xmath38 0.013 & 0.209 @xmath38 0.013 & @xcite + 2009 aug 01 & 2009.580 & keck / nirc2 & [ 0.373 @xmath38 0.014 ] & [ 238.5 @xmath38 1.5 ] & 0.318 @xmath38 0.010 & 0.195 @xmath38 0.010 & @xcite + 2009 oct 08 & 2009.767 & vlt / naco & [ 0.375 @xmath38 0.010 ] & [ 234.7 @xmath38 1.1 ] & 0.306 @xmath38 0.007 & 0.217 @xmath38 0.007 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 0.362 @xmath38 0.014 ] & [ 237.2 @xmath38 1.6 ] & 0.304 @xmath38 0.010 & 0.196 @xmath38 0.010 & @xcite + 2009 nov 01 & 2009.832 & keck / nirc2 & [ 0.362 @xmath38 0.013 ] & [ 238.9 @xmath38 1.4 ] & 0.310 @xmath38 0.009 & 0.187 @xmath38 0.009 & @xcite + 2010 jul 13 & 2010.528 & keck / nirc2 & [ 0.368 @xmath38 0.011 ] & [ 242.0 @xmath38 1.3 ] & 0.325 @xmath38 0.008 & 0.173 @xmath38 0.008 & @xcite + 2010 jul 13 & 2010.528 & keck / nirc2 & [ 0.363 @xmath38 0.008 ] & [ 242.8 @xmath38 1.0 ] & 0.323 @xmath38 0.006 & 0.166 @xmath38 0.006 & @xcite + 2010 jul 21 & 2010.550 & keck / nirc2 & [ 0.368 @xmath38 0.016 ] & [ 241.6 @xmath38 1.7 ] & 0.324 @xmath38 0.011 & 0.175 @xmath38 0.011 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 0.371 @xmath38 0.014 ] & [ 244.1 @xmath38 1.5 ] & 0.334 @xmath38 0.010 & 0.162 @xmath38 0.010 & @xcite + 2010 oct 30 & 2010.827 & keck / nirc2 & [ 0.370 @xmath38 0.027 ] & [ 247.2 @xmath38 2.5 ] & 0.341 @xmath38 0.016 & 0.143 @xmath38 0.016 & @xcite + 2011 jul 21 & 2011.550 & keck / nirc2 & [ 0.375 @xmath38 0.011 ] & [ 249.7 @xmath38 1.2 ] & 0.352 @xmath38 0.008 & 0.130 @xmath38 0.008 & @xcite + 2011 oct 16 & 2011.789 & lbt / pisces & [ 0.347 @xmath38 0.016 ] & [ 249.9 @xmath38 1.8 ] & 0.326 @xmath38 0.011 & 0.119 @xmath38 0.011 & @xcite + 2011 nov 09 & 2011.854 & lbt / pisces & [ 0.403 @xmath38 0.016 ] & [ 251.6 @xmath38 1.6 ] & 0.382 @xmath38 0.011 & 0.127 @xmath38 0.011 & @xcite + 2012 jun 14 & 2012.452 & palomar / p1640 & [ 0.377 @xmath38 0.008 ] & [ 256.2 @xmath38 0.9 ] & 0.366 @xmath38 0.006 & 0.090 @xmath38 0.006 & @xcite + 2012 jul 22 & 2012.556 & keck / nirc2 & [ 0.382 @xmath38 0.011 ] & [ 257.3 @xmath38 1.2 ] & 0.373 @xmath38 0.008 & 0.084 @xmath38 0.008 & @xcite + 2012 oct 26 & 2012.819 & keck / nirc2 & [ 0.378 @xmath38 0.013 ] & [ 258.4 @xmath38 1.4 ] & 0.370 @xmath38 0.009 & 0.076 @xmath38 0.009 & @xcite + 2013 oct 16 & 2013.789 & keck / nirc2 & [ 0.373 @xmath38 0.018 ] & [ 267.4 @xmath38 2.0 ] & 0.373 @xmath38 0.013 & 0.017 @xmath38 0.013 & @xcite + 2013 oct 21 & 2013.802 & lbt / lmircam & 0.395 @xmath38 0.012 & 264.8 @xmath38 1.7 & 0.394 @xmath38 0.011 & 0.036 @xmath38 0.017 & @xcite + 2014 jul 13 & 2014.528 & vlt / sphere & [ 0.386 @xmath38 0.012 ] & [ 268.8 @xmath38 1.3 ] & 0.386 @xmath38 0.009 & 0.008 @xmath38 0.009 & @xcite + 2014 jul 17 & 2014.539 & keck / nirc2 & [ 0.387 @xmath38 0.016 ] & [ 270.4 @xmath38 1.7 ] & 0.387 @xmath38 0.011 & 0.003 @xmath38 0.011 & @xcite + 2014 aug 12 & 2014.611 & vlt / sphere & [ 0.384 @xmath38 0.003 ] & [ 269.3 @xmath38 0.3 ] & 0.384 @xmath38 0.002 & 0.005 @xmath38 0.002 & @xcite + 2014 dec 0408 & 2014.93 & vlt / sphere & [ 0.384 @xmath38 0.014 ] & [ 272.1 @xmath38 1.5 ] & 0.384 @xmath38 0.010 & 0.014 @xmath38 0.010 & @xcite + 2014 dec 0509 & 2014.93 & vlt / sphere & 0.381 @xmath38 0.007 & 272.5 @xmath38 0.7 & [ 0.381 @xmath38 0.007 ] & [ 0.016 @xmath38 0.005 ] & @xcite + 2003 nov 10 & 2003.857 & vlt / naco & 0.411 @xmath38 0.008 & 31.8 @xmath38 1.3 & [ 0.217 @xmath38 0.009 ] & [ 0.349 @xmath38 0.008 ] & @xcite + 2003 nov 10 & 2003.857 & vlt / naco & 0.411 @xmath38 0.008 & 31.5 @xmath38 1.3 & [ 0.215 @xmath38 0.009 ] & [ 0.350 @xmath38 0.008 ] & @xcite + 2003 nov 10 & 2003.857 &
vlt / naco & 0.413 @xmath38 0.022 & 34.4 @xmath38 3.5 & [ 0.233 @xmath38 0.022 ] & 0.341 @xmath38 0.022 & @xcite + 2003 nov 13 & 2003.865 & vlt / naco & 0.401 @xmath38 0.008 & 32.1 @xmath38 1.4 & [ 0.213 @xmath38 0.009 ] &
[ 0.340 @xmath38 0.009 ] & @xcite + 2008 nov 11 & 2008.862 & vlt / naco & 0.210 @xmath38 0.027 & 211.5 @xmath38 1.9 & [ 0.110 @xmath38 0.015 ] & [ 0.179 @xmath38 0.023 ] & @xcite + 2009 oct 25 & 2009.813 & vlt / naco & 0.298 @xmath38 0.016 & 210.6 @xmath38 3.6 & [ 0.152 @xmath38 0.018 ] & [ 0.257 @xmath38 0.017 ] & @xcite + 2009 oct 25 & 2009.813 & vlt / naco & 0.299 @xmath38 0.014 & 210.7 @xmath38 2.9 & 0.153 @xmath38 0.014 & 0.257 @xmath38 0.014 & @xcite + 2009 nov dec & 2009.9 & vlt / naco & 0.299 @xmath38 0.016 & 209.4 @xmath38 3.5 & [ 0.147 @xmath38 0.018 ] & [ 0.260 @xmath38 0.017 ] & @xcite + 2009 dec 03 & 2009.920 & gemini - s / nici & 0.323 @xmath38 0.010 & 209.3 @xmath38 1.8 & [ 0.158 @xmath38 0.010 ] & [ 0.282 @xmath38 0.010 ] & @xcite + 2009 dec 03 & 2009.920 & gemini - s / nici & 0.339 @xmath38 0.010 & 209.2 @xmath38 1.7 & [ 0.165 @xmath38 0.010 ] & [ 0.296 @xmath38 0.010 ] & @xcite + 2009
dec 26 & 2009.983 & vlt / naco & 0.302 @xmath38 0.016 & 212.8 @xmath38 3.7 & [ 0.164 @xmath38 0.019 ] & [ 0.254 @xmath38 0.017 ] & @xcite + 2009 dec 26 & 2009.983 & vlt / naco & 0.299 @xmath38 0.013 & 211.1 @xmath38 1.7 & [ 0.154 @xmath38 0.010 ] & [ 0.256 @xmath38 0.012 ] & @xcite + 2009
dec 29 & 2009.991 & vlt / naco & 0.314 @xmath38 0.016 & 211.5 @xmath38 3.5 & [ 0.164 @xmath38 0.018 ] & [ 0.268 @xmath38 0.017 ] & @xcite + 2009 dec 29 & 2009.991 & vlt / naco & 0.326 @xmath38 0.013 & 210.6 @xmath38 1.2 & [ 0.166 @xmath38 0.009 ] & [ 0.280 @xmath38 0.012 ] & @xcite + 2009 dec 29 & 2009.991 & vlt / naco & 0.306 @xmath38 0.009 & 212.1 @xmath38 1.7 & 0.163 @xmath38 0.009 ] & [ 0.260
@xmath38 0.008 & @xcite + 2010 mar 20 & 2010.214 & vlt / naco & 0.345 @xmath38 0.007 & 209.4 @xmath38 0.3 & [ 0.169 @xmath38 0.004 ] & [ 0.301 @xmath38 0.006 ] & @xcite + 2010 apr 03 & 2010.252 & vlt / naco+app & 0.354 @xmath38
0.012 & 209.1 @xmath38 2.1 & [ 0.172 @xmath38 0.013 ] & [ 0.309 @xmath38 0.012 ] & @xcite + 2010 apr 10 & 2010.271 & vlt / naco & 0.355 @xmath38 0.002 & 208.4 @xmath38 1.1 & [ 0.169 @xmath38 0.006 ] & [ 0.312 @xmath38 0.004 ] & @xcite + 2010 apr 10 & 2010.271 & vlt / naco & 0.346 @xmath38 0.007 & 209.9 @xmath38 1.2 & 0.173 @xmath38 0.007 & 0.300 @xmath38 0.007 & @xcite + 2010 sep 28 & 2010.739 & vlt / naco & 0.383 @xmath38 0.011 & 210.3 @xmath38 1.7 & 0.193 @xmath38 0.011 & 0.331 @xmath38 0.011 & @xcite + 2010 sep 28 & 2010.739 & vlt / naco & 0.385 @xmath38
0.011 & 210.1 @xmath38 1.5 & [ 0.193 @xmath38 0.010 ] & [ 0.333 @xmath38 0.011 ] & @xcite + 2010 nov 16 & 2010.873 & vlt / naco & 0.387 @xmath38 0.008 & 212.4 @xmath38 1.4 & 0.207 @xmath38 0.008 & 0.326 @xmath38 0.010 & @xcite + 2010 nov 17 & 2010.876 & vlt / naco & 0.390
@xmath38 0.013 & 212.3 @xmath38 2.1 & 0.209 @xmath38 0.013 & 0.330 @xmath38 0.014 & @xcite + 2010 dec 25 & 2010.980 & gemini - s / nici & 0.404 @xmath38 0.010 & 212.1 @xmath38 0.7 & [ 0.215 @xmath38 0.007 ] & [ 0.342 @xmath38 0.009 ] & @xcite + 2010 dec 25 & 2010.980 & gemini - s / nici & 0.407 @xmath38 0.005 & 212.9 @xmath38 1.4 & [ 0.221 @xmath38 0.009 ] & [ 0.342 @xmath38 0.007 ] & @xcite + 2011 jan 02 & 2011.003 & vlt / naco & 0.408 @xmath38 0.009 & 211.1 @xmath38 1.5 & 0.211 @xmath38 0.009 & 0.350 @xmath38 0.010 & @xcite + 2011 mar 26 & 2011.230 & vlt / naco & 0.426 @xmath38 0.013 & 210.1 @xmath38 1.8 & 0.214 @xmath38 0.012 & 0.367 @xmath38 0.014 & @xcite + 2011 oct 12 & 2011.778 & vlt / naco & 0.439 @xmath38 0.005 & 212.9 @xmath38 0.5 & [ 0.238 @xmath38 0.003 ] & [ 0.369 @xmath38 0.003 ] & @xcite + 2011 oct 20 & 2011.800 & gemini - s / nici & 0.455 @xmath38 0.003 & 211.9 @xmath38 0.4 & [ 0.240 @xmath38 0.003 ] & [ 0.386 @xmath38 0.003 ] & @xcite + 2011 oct 20 & 2011.800 & gemini - s / nici & 0.452 @xmath38 0.005 & 211.6 @xmath38 0.6 & [ 0.237 @xmath38 0.005 ] & [ 0.385 @xmath38 0.005 ] & @xcite + 2011 dec 11 & 2011.942 & vlt / naco & 0.441 @xmath38 0.003 & 212.9 @xmath38 0.3 & [ 0.240 @xmath38 0.003 ] & [ 0.370 @xmath38 0.003 ] & @xcite + 2012 mar 29 & 2012.341 & gemini - s / nici & 0.447 @xmath38 0.003 & 210.8 @xmath38 0.4 & [ 0.229 @xmath38 0.003 ] & [ 0.384 @xmath38 0.003 ] & @xcite + 2012 mar 29 & 2012.341 & gemini - s / nici & 0.448 @xmath38 0.005 & 211.8 @xmath38 0.6 & [ 0.236 @xmath38 0.005 ] & [ 0.381 @xmath38 0.005 ] & @xcite + 2012 nov 01 & 2012.835 & vlt / naco & 0.456 @xmath38 0.011 & 211.9 @xmath38 2.3 & 0.241 @xmath38 0.013 & 0.387 @xmath38 0.014 & @xcite + 2012 dec 04 & 2012.925 & magellan / magao & 0.470 @xmath38 0.010 & 212.0 @xmath38 1.2 & [ 0.249 @xmath38 0.010 ] & [ 0.399 @xmath38 0.010 ] & @xcite + 2012 dec 0107 & 2012.92 & magellan / magao & 0.461 @xmath38 0.014 & 211.9 @xmath38 1.2 & [ 0.244 @xmath38 0.011 ] & [ 0.391 @xmath38 0.013 ] & @xcite + 2012 dec 16 & 2012.958 & vlt / naco & 0.449 @xmath38 0.006 & 211.6 @xmath38 0.7 & [ 0.235 @xmath38 0.006 ] & [ 0.382 @xmath38 0.006 ] & @xcite + 2013
jan 31 & 2013.082 & vlt / naco+agpm & 0.452 @xmath38 0.010 & 211.2 @xmath38 1.3 & [ 0.234 @xmath38 0.010 ] & [ 0.387 @xmath38 0.010 ] & @xcite + 2013 jan 31 & 2013.082 & vlt / naco+agpm & 0.448 @xmath38 0.004 & 212.3 @xmath38 0.3 & [ 0.239 @xmath38 0.003 ] & [ 0.379 @xmath38 0.004 ] & @xcite + 2013 nov 16 & 2013.873 & gemini - s / gpi & 0.430 @xmath38 0.003 & 212.3 @xmath38 0.4 & [ 0.230 @xmath38 0.003 ] & [ 0.363 @xmath38 0.003 ] & @xcite + 2013 nov 16 & 2013.873 & gemini - s / gpi & 0.426 @xmath38 0.003 & 212.8 @xmath38 0.4 & [ 0.231 @xmath38 0.003 ] & [ 0.358 @xmath38 0.003 ] & @xcite + 2013 nov 18 & 2013.879 & gemini - s / gpi & 0.434 @xmath38 0.006 & 211.8 @xmath38 0.5 & [ 0.229 @xmath38 0.005 ] & [ 0.369 @xmath38 0.005 ] & @xcite + 2013 nov 18 & 2013.879 & gemini - s / gpi & 0.428 @xmath38 0.003 & 212.2 @xmath38 0.4 & [ 0.228 @xmath38 0.003 ] & [ 0.362 @xmath38 0.003 ] & @xcite + 2013 dec 10 & 2013.939 & gemini - s / gpi & 0.430 @xmath38 0.010 & 211.6 @xmath38 1.3 & [ 0.225 @xmath38 0.010 ] & [ 0.366 @xmath38 0.010 ] & @xcite + 2013 dec 10 & 2013.939 & gemini - s / gpi & 0.419 @xmath38 0.006 & 212.2 @xmath38 0.8 & [ 0.223 @xmath38 0.006 ] & [ 0.355 @xmath38 0.006 ] & @xcite + 2013 dec 10 & 2013.939 & gemini - s / gpi & 0.419 @xmath38 0.004 & 212.6 @xmath38 0.5 & [ 0.226 @xmath38 0.004 ] & [ 0.353 @xmath38 0.004 ] & @xcite + 2013 dec 11 & 2013.939 & gemini - s / gpi & 0.419 @xmath38 0.005 & 212.3 @xmath38 0.7 & [ 0.224 @xmath38 0.005 ] & [ 0.354 @xmath38 0.005 ] & @xcite + 2013 dec 12 & 2013.945 & gemini - s / gpi & 0.427 @xmath38 0.007 & 211.8 @xmath38 0.7 & [ 0.225 @xmath38 0.006 ] & [ 0.363 @xmath38 0.007 ] & @xcite + 2014 mar 23 & 2014.222 & gemini - s / gpi & 0.413 @xmath38 0.003 & 212.1 @xmath38 0.4 & [ 0.219 @xmath38 0.003 ] & [ 0.350 @xmath38 0.003 ] & @xcite + 2014 nov 08 & 2014.852 & gemini - s / gpi & 0.363 @xmath38 0.004 & 212.2 @xmath38 0.7 & [ 0.193 @xmath38 0.004 ] & [ 0.307 @xmath38 0.004 ] & @xcite + 2015 jan 24 & 2015.063 & gemini - s / gpi & 0.348 @xmath38 0.005 & 212.2 @xmath38 0.7 & [ 0.185 @xmath38 0.004 ] & [ 0.294 @xmath38 0.005 ] & @xcite +
|
high - contrast adaptive optics imaging is a powerful technique to probe the architectures of planetary systems from the outside - in and survey the atmospheres of self - luminous giant planets .
direct imaging has rapidly matured over the past decade and especially the last few years with the advent of high - order adaptive optics systems , dedicated planet - finding instruments with specialized coronagraphs , and innovative observing and post - processing strategies to suppress speckle noise .
this review summarizes recent progress in high - contrast imaging with particular emphasis on observational results , discoveries near and below the deuterium - burning limit , and a practical overview of large - scale surveys and dedicated instruments .
i conclude with a statistical meta - analysis of deep imaging surveys in the literature .
based on observations of 384 unique and single young ( @xmath05300 myr ) stars spanning stellar masses between 0.13.0 , the overall occurrence rate of 513 companions at orbital distances of 30300 au is 0.6@xmath1% assuming hot - start evolutionary models . the most massive giant planets regularly accessible to direct imaging are about as rare as hot jupiters are around sun - like stars .
dividing this sample into individual stellar mass bins does not reveal any statistically - significant trend in planet frequency with host mass : giant planets are found around 2.8@xmath2% of ba stars , @xmath34.1% of fgk stars , and @xmath33.9% of m dwarfs . looking forward , extreme adaptive optics systems and the next generation of ground- and space - based telescopes with smaller inner working angles and deeper detection limits will increase the pace of discovery to ultimately map the demographics , composition , evolution , and origin of planets spanning a broad range of masses and ages .
|
in this work we tackle the problem of _ unconstrained text recognition _ recognising text in natural images without restricting the words to a fixed lexicon or dictionary .
usually this problem is decomposed into a word detection stage followed by a word recognition stage .
the word detection stage generates bounding boxes around words in an image , while the word recognition stage takes the content of these bounding boxes and recognises the text within .
this paper focuses on the text recognition stage , developing a model based on deep convolutional neural networks ( cnns ) ( @xcite ) .
previous methods using cnns for word recognition ( discussed in more detail in section section [ sec : related ] ) has either constrained ( @xcite ) or heavily weighted ( @xcite ) the recognition results to be from a dictionary of known words .
this works very well when training and testing are limited to a fixed vocabulary , but does not generalise to where previously unseen or non - language based text must be recognised for example for generic alpha - numeric strings such as number plates or phone numbers . the shift of focus towards a model which performs accurately without a fixed dictionary increases the complexity of the text recognition problem .
to solve this , we propose a novel cnn architecture ( figure [ fig : path ] ) employing a _ conditional random field _
( crf ) whose unary terms are outputs of a cnn character predictor , which are position - dependent , and whose higher order terms are outputs of a cnn n - gram predictor , which are position - independent .
the recognition result is then obtained by finding the character sequence that maximises the crf score , enforcing the consistency of the individual predictions .
the crf model builds on our previous work where we explored dictionary - based recognition ( @xcite ) for two scenarios : the first was to train a different cnn character classifier for each position in the word being recognised , using the whole image of the word as input to each classifier ( an idea also expored by @xcite ) ; the second was to construct a cnn predictor to detect the n - grams contained in the word , effectively encoding the text as a bag - of - n - grams .
the dictionary - free joint model proposed here is trained by defining a structured output learning problem , and back - propagating the corresponding _ structured output loss_. this formulation results in multi - task learning of both the character and n - gram predictors , and additionally learns how to combine their representations in the crf , resulting in more accurate text recognition .
the result is a highly flexible text recognition system that achieves excellent unconstrained text recognition performance as well as state - of - the - art recognition performance when using standard dictionary constraints .
while performance is measured on real images as contained in standard text recognition benchmarks , all results are obtained by training the model _ purely on synthetic data_. the model is evaluated on this synthetic data as well in order to study its performance under different scenarios .
section [ sec : related ] outlines work related to ours .
section [ sec : characters ] reviews the character sequence model and section [ sec : ngrams ] the bag - of - n - grams model .
section [ sec : joint ] shows how these predictors can be combined to form a joint crf model and formulates the training of the latter as structured - output learning .
section [ sec : eval ] evaluates these models extensively and section [ sec : conclusion ] summarises our findings .
we concentrate here on text recognition methods , recognising from a cropped image of a single word , rather than the text detection stages of scene text recognition ( ` text spotting ' ) that generate the word detections .
traditional text recognition methods are based on sequential character classification , finding characters by sliding window methods ( @xcite , after which a word prediction is made by integrating character classifier predictions in a left - to - right manner .
the character classifiers include random ferns ( @xcite ) in @xcite , and cnns in @xcite . both @xcite and @xcite use a small fixed lexicon as a language model to constrain word recognition .
more recent works such as @xcite make use of over - segmentation methods , guided by a supervised classifier , to generate candidate character proposals in a single - word image , which are subsequently classified as true or false positives .
for example , photoocr ( @xcite ) uses binarization and a sliding window classifier to generate candidate character regions , with words recognised through a beam search driven by classifier scores and static n - gram language model , followed by a re - ranking using a dictionary of 100k words .
@xcite uses the convolutional nature of cnns to generate response maps for characters and bigrams which are integrated to score lexicon words .
in contrast to these approaches based on character classification , the work by @xcite instead uses the notion of holistic word recognition .
@xcite still rely on explicit character classifiers , but construct a graph to infer the word , pooling together the full word evidence .
@xcite use aggregated fisher vectors ( @xcite ) and a structured svm framework to create a joint word - image and text embedding . @xcite and more recently @xcite also formluate joint embedding spaces , achieving impressive results with minimal training data .
@xcite use whole word - image features to recognize words by comparing to simple black - and - white font - renderings of lexicon words . in our own previous work ( @xcite )
we use large cnns acting on the full word image region to perform 90k - way classification to a dictionary word .
it should be noted that all the methods make use of strong static language models , either relying on a constrained dictionary or re - ranking mechanism .
@xcite had great success using a cnn with multiple position - sensitive character classifier outputs ( closely related to the character sequence model in section [ sec : characters ] ) to perform street number recognition .
this model was extended to captcha sequences ( up to 8 characters long ) where they demonstrated impressive performance using synthetic training data for a synthetic problem ( where the generative model is known ) , but we show that synthetic training data can be used for a real - world data problem ( where the generative model is unknown ) .
there have been previous uses of graphical models with back - propagated loss functions for neural networks , such as the early text recognition work of @xcite to combine character classifier results on image segmentations .
another example is the recent work of @xcite for human pose estimation , where an mrf - like model over the distribution of spatial locations for each body part is constructed , incorporating a single round of message - passing .
we now review the component cnn models , originally presented in our tech report @xcite , that form the basis of our joint model in section [ sec : joint ] . [ cols="^,^ " , ]
in this paper we have introduced a new formulation for word recognition , designed to be used identically in language and non - language scenarios . by modelling character positions and the presence of common n - grams ,
we can define a joint graphical model .
this can be trained effectively by back propagating structured output loss , and results in a more accurate word recognition system than predicting characters alone .
we show impressive results for unconstrained text recognition with the ability to generalise recognition to previously unseen words , and match state - of - the - art accuracy when comparing in lexicon constrained scenarios .
|
we develop a representation suitable for the unconstrained recognition of words in natural images , where unconstrained means that there is no fixed lexicon and words have unknown length . to this end
we propose a convolutional neural network ( cnn ) based architecture which incorporates a conditional random field ( crf ) graphical model , taking the whole word image as a single input .
the unaries of the crf are provided by a cnn that predicts characters at each position of the output , while higher order terms are provided by another cnn that detects the presence of n - grams .
we show that this entire model ( crf , character predictor , n - gram predictor ) can be jointly optimised by back - propagating the structured output loss , essentially requiring the system to perform multi - task learning , and training requires only synthetically generated data .
the resulting model is a more accurate system on standard real - world text recognition benchmarks than character prediction alone , setting a benchmark for systems that have not been trained on a particular lexicon .
in addition , our model achieves state - of - the - art accuracy in lexicon - constrained scenarios , without being specifically modelled for constrained recognition . to test the generalisation of our model
, we also perform experiments with random alpha - numeric strings to evaluate the method when no visual language model is applicable .
|
giant molecular clouds ( gmcs ) are the major reservoirs of molecular gas and the sites of most star formation in our galaxy and other galaxies .
their properties set the initial conditions for protostellar collapse , and may play a role in determining the stellar initial mass function @xcite .
moreover , because gmcs provide the bulk of the material for forming new stars their creation may be the limiting mechanism that regulates the rate of star formation in galaxies . therefore
, increasing our understanding of their properties and distribution throughout the different environments of external galaxies is likely to provide further insights into gmc and stellar formation processes .
there is a limited amount of information , however , that can be gained from studies that resolve the general distribution of molecular gas but not the individual molecular clouds . resolving gmcs to measure their sizes , velocity dispersions , and luminosities is a critical step in understanding the processes that ultimately drive galaxy evolution .
studies of resolved molecular clouds in the milky way find that gmcs are in approximate virial equilibrium and obey scaling relations , commonly known as larson laws @xcite , that have their origin in the character of the turbulence in the interstellar medium @xcite .
large scale surveys of the milky way show that gmcs follow uniform scaling relations @xcite , with few differences present between those located in the inner and the outer disk @xcite .
@xcite established that velocity dispersion , size , and luminosity are correlated in milky way gmcs .
observations indicate that gmc line widths increase as a power of their radius @xcite , such that @xmath1 where @xmath2 is the one - dimensional velocity dispersion of the gmc , and @xmath3 is its radius measured in parsecs
. equation [ size_linewidth_eq ] is known as the _ size - line width _ relation , and it has been shown to hold even within gmcs down to very small scales @xcite .
the _ size - line width _ relation is an expression of the equilibrium turbulence conditions in the molecular ism .
detailed modeling of gmc line profiles shows that the emission is macroturbulent ( i.e. , the scale size of the turbulence is larger than the photon mean free path ) , corresponding to many optically thick clumps that have a clump - to - clump velocity dispersion similar to the observed line width @xcite .
it is known that the observed line widths in all but a few very compact , quiescent clouds are too large to be thermal ; at the typical gmc temperatures of @xmath4 k the thermal co velocity dispersion would be @xmath5 .
they are rather due to supersonic turbulence within the clouds .
the dominant sources of turbulence in the molecular ism remain somewhat controversial . for example
, it is not yet known whether turbulence in molecular clouds is primarily internally or externally driven ( e.g. , * ? ? ?
contributors must include star formation in the form of jets and winds for low mass stars , and winds and expanding hii regions for massive stars and star destruction in the form of supernovae shocks and expanding superbubbles .
global processes that couple the large reservoirs of thermal or rotational energy to the turbulent cascade such as spiral shocks and several types of instabilities that involve shear , magnetic fields , and self gravity are important on the largest spatial scales .
the fact that turbulence is present on scales @xmath6 pc ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and that the dissipation lifetime of the largest eddies is @xmath7 yr @xcite , point to constant ( or at least frequent ) energy injection on large scales .
this in turn suggests that massive stellar death and gobal mechanisms are predominantly responsible for the bulk of turbulent energy injection in the milky way , although their respective dominance is a matter of debate @xcite . the last two larson relations , the _ luminosity - line width _ and _ luminosity - size _ relations , describe correlations between cloud luminosity , @xmath8 , and either velocity dispersion or size ( note , as is frequently pointed out , that only two of the three larson relations are independent ) .
they are @xcite @xmath9 and @xmath10 with @xmath2 and @xmath3 expressed in their corresponding units of and pc .
thus , the observations show that the virial mass , @xmath11 , relates to the co luminosity ( which is proportional to the inferred luminous mass @xmath12 ) in the following manner @xcite , @xmath13 where @xmath8 is expressed in
. we will call this equation the galactic _ mass - luminosity _ relation .
note that the exponent of @xmath8 is somewhat uncertain , and we take here the value reported by @xcite .
although the relation may not be precisely linear it is almost so , and simple models of molecular clouds can explain the observed linearity @xcite .
this approximate proportionality between @xmath14 and @xmath11 within the galaxy allows the definition of an empirical co - to - h@xmath0 conversion factor , , that relates the intensity of a gmc in the @xmath15co @xmath16 transition to its h@xmath0 column density ( @xmath17 ) and ultimately its mass .
for gmcs around a median mass of @xmath18 m@xmath19 @xcite obtained a co - to - h@xmath0 factor that , when adjusted to the currently accepted distance of 8.5 kpc to the galactic center , yields @xmath20 @xcite . in the milky way ,
this `` virial mass '' approach yields conversion factors similar to those obtained from non - dynamical measurements , strongly suggesting that clouds are in a dynamical state intermediate between marginal gravitational binding and virial equilibrium ( e.g. , @xmath21-ray results and modeling of the galactic plane dust continuum ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , with magnetic energy introducing a small but non - negligible correction on the small scales ( @xmath22 ; * ? ? ?
throughout this paper we adopt @xmath23 as the galactic conversion factor .
note , however , that some nearby lines - of - sight at high galactic latitudes may exhibit significant larger values of this coefficient ( e.g. , * ? ? ?
* ; * ? ? ?
therefore , mass estimates of gmcs show that they are gravitationally bound and in approximate virial equilibrium , which requires @xmath24 . as a consequence of equation [ size_linewidth_eq ] and virial equilibrium ,
all gmcs have approximately the same mean column density : @xmath25 @xmath26 , or an equivalent mass surface density @xmath27 m@xmath19 pc@xmath28 ( after applying a 40% correction for the mass contribution of elements heavier than hydrogen ) .
an equivalent result is that all milky way gmcs have approximately the same mean visual extinction from edge to edge , @xmath29 , using the standard dust - to - gas ratio @xcite .
@xcite argues that equation [ lco_mvir_eq ] is a consequence of the interplay between co chemistry , the milky way radiation field , the dust - to - gas ratio , and the local pressure of the ism . as a result ,
h@xmath0 clouds observable in co are at least marginally bound .
furthermore , @xcite argues that the constancy of @xmath30 stems from the fact that clouds with lower column density would not be bright in co , while clouds with higher column density would rapidly collapse .
another consequence of the larson relations is that galactic gmcs have low volume - averaged densities , @xmath31 @xmath32 .
therefore their clumpiness must be high in order to excite the rotational transitions of co and of other commonly observed high - density tracer molecules such as hco@xmath33 , which require volume densities in excess of 10@xmath34 @xmath32 .
this clumpiness is probably induced by supersonic turbulence @xcite .
further , the average density of a gmc decreases with increasing mass , and as a result the free fall time in the cloud increases .
this led @xcite to predict that larger gmcs will form collapsing cores and stars less rapidly than smaller ones .
the purpose of this study is to determine , _ using a consistent methodology _ , the relations between size , velocity dispersion , and luminosity for extragalactic gmcs that are encapsulated in the milky way by the larson laws .
we do so by studying gmcs in dwarf galaxies and comparing them with those in spiral galaxies .
dwarf galaxies are both numerous and nearby , and host physical conditions that fall outside those explored by surveys of the milky way .
the range of parameter space thus probed provides an ideal laboratory to test theories of cloud structure and formation .
particularly , dwarf galaxies have low metallicities and their lack of internal extinction results in intense radiation fields , reminiscent of the conditions in primitive galaxies .
they display irregular morphologies and are not dominated by spiral structure .
dwarfs often have slowly rising , nearly solid - body , rotation curves ; this translates into small rotational shear and perhaps changes the sources and scales of turbulence in their ism .
the stellar potential wells of dwarf galaxies are weaker than those in large spiral galaxies and the ambient ism may be correspondingly less dense .
these environmental changes may influence the equilibrium conditions of gmcs , and affect their larson relations .
the integrated properties of dwarf galaxies already show clear evidence that environment affects at least the co content . dwarf irregular galaxies are fainter in co than large spiral galaxies .
galaxies with metallicity @xmath35 are seldom detected in co emission @xcite .
dwarf galaxies of higher metallicity detected in co , however , show similar co luminosity relative to either their stellar or atomic gas content as larger galaxies @xcite . here
we investigate whether the resolved properties of gmcs in these galaxies also reflect their changing environments .
the study of resolved gmc properties in other galaxies remains technically challenging .
a typical size for a galactic gmc is @xmath36 pc , which at the @xmath37 mpc distance of the m 81 group translates to an angular size of @xmath38 .
achieving such a resolution requires an antenna 250 meters in diameter operating at @xmath39 mm , the wavelength of the ground transition of @xmath15co .
this is clearly an impractical proposition , but aperture synthesis techniques used by millimeter - wave interferometers routinely attain this resolution .
they do so , however , at the price of reduced surface brightness sensitivity with respect to their filled aperture equivalents , and studies such as the one presented here are ultimately limited by sensitivity .
until the advent of the atacama large millimeter array ( alma ) these observations will remain challenging .
because of these difficulties , the resolved properties of gmcs in galaxies other than the milky way remain a largely unexplored domain . because of its proximity , located at a distance where even a modest telescope can attain a resolution of @xmath40 pc , the first galaxy where observations capable of clearly resolving individual gmcs were possible was the large magellanic cloud @xcite .
the first resolved study of a gmc beyond the immediate vicinity of the milky way was performed in m 31 @xcite , while the first large scale interferometric survey of gmcs in another galaxy was done on m 33 @xcite .
the first comprehensive analyses in the magellanic clouds were performed on the small magellanic cloud @xcite .
@xcite recently presented a study of gmcs in the local group , analyzing observations from co surveys complete down to a known mass limit over a significant fraction of each galaxy .
this study concluded that the larson relations in the milky way disk were approximately those observed throughout the local group , although some offsets ( possibly methodological ) were present .
they also concluded that gmcs within a particular galaxy have roughly constant surface density , @xmath41 , and that once corrections for the local co - to - h@xmath0 conversion factor are applied the sample as a whole has a scatter of only a factor of 2 in @xmath41 .
finally , they calculated a typical conversion factor @xmath42 , accurate to within 50% for most of the sample except the small magellanic cloud , for which they obtain a value @xmath43 times larger .
this paper proceeds as follows : in [ data ] we describe the data sets and and galaxies studied . in [ method ]
we summarize the method used to consistently derive gmc properties across the different datasets . in [ results ]
we discuss the _ size - line width _
, _ luminosity - line width _ , and _ luminosity - size _ relations in our sample of extragalactic gmcs , and compare them with the galaxy . in
[ discussion ] we discuss the implications of these results on the equilibrium of clouds in the small magellanic cloud , the photoionization - regulated star formation theory , the co - to - h@xmath0 conversion factor , and the brightness temperature of extragalactic clouds . in each case
we contrast our results in dwarf galaxies with those obtained in the milky way and the other local group spirals , m 31 and m 33 . in [ summary ]
we summarize this study and present our conclusions .
we use data from four telescopes : the berkeley - illinois - maryland array ( bima ; * ? ? ?
* ) , caltech s owens valley radio observatory ( ovro ) millimeter array , the institut de radioastronomie millimtrique plateau de bure interferometer ( pdbi ) , and the swedish - eso submillimeter telescope ( sest ) .
we measure gmc properties in @xmath44 dwarf galaxies : the large and small magellanic clouds , ic 10 , ngc 185 , ngc 205 , ngc 1569 , ngc 2976 , ngc 3077 , ngc 4214 , ngc 4449 , and ngc 4605 .
we compare gmc properties derived in dwarfs to those found in the milky way ( data obtained by the five colleges radio astronomy observatory , fcrao ; * ? ? ?
* ) and the other two local group spiral galaxies , m 31 and m 33 . the latter two act as key `` control '' data sets because they are obtained by the same telescopes at similar angular and spatial resolutions to the dwarf galaxy gmc data .
table [ datatable ] gives the source , telescope , line observed , resolutions , and reference for each data set . when we analyze observations of the co @xmath45 transition , we assume @xmath46 ( i.e. , thermalized optically thick emission ) . for the gmc complex in the n83 region of the small magellanic cloud , this is within @xmath47 of the observed ratio @xcite . in the milky way
the observed ratio varies from unity to @xmath48 ( e.g. , * ? ? ?
several data sets appear here for the first time .
notably , gmcs in ngc 3077 , ngc 4214 , and ngc 4449 where mapped at bima as part of a large project studying the molecular ism in dwarf galaxies .
the bima maps of ngc 2976 and ngc 4605 were obtained as part of the same project and have appeared previously in kinematic analyses @xcite .
the observation and reduction procedure for these data follow the descriptions in those papers .
table [ galaxytable ] lists the properties of the galaxies that we study .
most have late type morphologies ; the exceptions are ngc 185 and ngc 205 , local group dwarf ellipticals mapped by @xcite using bima .
all of our sources are @xmath49 mpc away ; the most distant galaxy is ngc 4605 at 4.26 mpc .
we indicate the absolute blue magnitude of each galaxy corrected by galactic extinction to illustrate their luminosity .
the galaxies span the range @xmath50 magnitudes .
we give metallicities for each galaxy , with the corresponding references for metallicity and distance .
lccccc ic 10 & ovro & @xmath51 & 19 & 0.7 & 5 + ngc 185 & bima & @xmath51 & 15 & 4 & 1 + ngc 205 & bima & @xmath51 & 28 & 3 & 1 + smc n83 & sest & @xmath52,@xmath51 & 11,17 & 0.25 & 9 + smc lirs36 & sest & @xmath52 & 7 & 0.05 & 10 + smc lirs49 & sest & @xmath52 & 7 & 0.05 & 10 + lmc n159 & sest & @xmath52 & 6 & 0.25 & 11 + ngc 1569 & pdbi & @xmath52,@xmath51 & 22,45 & 1.6 & 8 + ngc 2976 & bima & @xmath51 & 94 & 3 & 2 + ngc 3077 & bima & @xmath51 & 41 & 3 & 3 + ngc 3077 & ovro & @xmath51 & 41 & 1.3 & 6 + ngc 4214 & bima & @xmath51 & 80 & 3 & 3 + ngc 4214 & ovro & @xmath51 & 64 & 1.3 & 7 + ngc 4449 & bima & @xmath51 & 117 & 3 & 3 + ngc 4605 & bima & @xmath51 & 109 & 3 & 4 + milky way & fcrao & @xmath53 & @xmath54 & 0.65 & 12 + m 31 & bima & @xmath55 & 27 & 2 & 14 + m 33 & bima & @xmath55 & 30 & 2 & 13 + lccccc ic 10 & irr / bcd & 0.95 & @xmath56 & 8.2 & 6,19 + ngc 185 & dsph / de3 & 0.63 & @xmath57 & 8.2 & 1,17 + ngc 205 & e5 & 0.85 & @xmath58 & 8.6 & 1,18 + smc & sm & 0.061 & @xmath56 & 8.02 & 7,20 + lmc & sm & 0.052 & @xmath59 & 8.43 & 7,20 + ngc 1569 & irr & 2.2 & @xmath60 & 8.19 & 2,16 + ngc 2976 & sc & 3.45 & @xmath61 & 8.7 & 3,13 + ngc 3077 & irr & 3.9 & @xmath62 & 8.85 & 3,11 + ngc 4214 & irr & 2.94 & @xmath63 & 8.23 & 4,12 + ngc 4449 & irr & 3.9 & @xmath59 & 8.32 & 5,15 + ngc 4605 & sc & 4.26 & @xmath64 & 8.69 & 3,14 + milky way & sb & 0.008 & @xmath65 & 8.7 & 10 + m31 & sb & 0.79 & @xmath66 & 8.7 & 8,21 + m33 & scd & 0.84 & @xmath67 & 8.4 & 9,21
+ in this section we briefly summarize the sources and datasets used . for more in depth descriptions we point the reader to the original references .
@xcite used the bima interferometer to map co @xmath16 emission in the local group dwarf elliptical galaxies ngc 185 and ngc 205 .
these galaxies , which are satellites of m 31 , are unusual early - type dwarfs with co emission , while the rest of our sample consists of late - type dwarf spiral and irregular galaxies .
@xcite presented a kinematic study based on bima observations of co @xmath16 emission from ngc 2976 , a dwarf spiral galaxy in the m 81 group .
these observations are considerably deeper than those from the bima survey of nearby galaxies ( song ; * ? ? ?
the metallicity for this galaxy is uncertain , but the polycyclic aromatic hydrocarbon abundance and dust - to - gas ratios derived by @xcite suggests it is similar to galactic .
@xcite mapped co @xmath16 emission from ngc 3077 , a member of the m 81 group currently undergoing a dramatic interaction with m 81 and m 82 , using the ovro millimeter interferometer .
this galaxy was also observed by bima during the same period , and those observations appear here for the first time .
we present gmc property measurements from both data sets .
we derive the metallicity for this galaxy using the integrated nebular fluxes of the bright oxygen lines by @xcite , employing the calibration of @xcite as parametrized by @xcite .
the metallicity is consistent with that derived using the nitrogen calibration by @xcite .
@xcite mapped co @xmath16 emission from ngc 4214 , a nearby dwarf irregular galaxy , using the ovro millimeter interferometer .
this galaxy was also observed by bima , and here we use those data for the first time .
bima mapped co @xmath16 emission from the nearby dwarf irregular ngc 4449 .
the data appear here for the first time .
@xcite presented co @xmath16 observations of ngc 4605 , a nearby isolated dwarf galaxy , and used them to study the kinematics of this galaxy .
the metallicity for this galaxy was derived using the same procedure as for ngc 3077 .
@xcite mapped co @xmath16 emission from most of the molecular clouds in the local group dwarf irregular ic 10 using the ovro millimeter interferometer .
these data were analyzed in detail by @xcite , where maps of individual gmcs are presented .
ic 10 is a local group irregular comparable in mass to the small magellanic cloud but presently undergoing a burst of vigorous star formation .
@xcite used the pdbi to map both co @xmath45 and @xmath16 emission from the nearby dwarf irregular ngc 1569 .
this galaxy , which is about the mass of the small magellanic cloud , experienced a starburst period that ended @xmath68 myr ago .
@xcite presented high signal - to - noise mapping of the n83 region in the small magellanic cloud ( smc ) using the sest telescope .
they observed both the co @xmath45 and @xmath16 transitions .
n83/n84 is a bright star forming region in the eastern wing of the small magellanic cloud .
@xcite used sest to map co @xmath45 emission in the n159 region of the large magellanic cloud ( lmc ) .
this region lies just south of the bright star forming region 30 doradus , and comprises clouds in different stages of evolution . in this study
we use gmcs properties measured in the spiral disk galaxies of the local group as comparison and control datasets . as discussed in the introduction ,
several studies have established the scalings for galactic molecular clouds @xcite .
recent interferometric work has shown that the gmcs in the other disk galaxies in the local group ( m 31 and m 33 ) follow the same scalings @xcite . the observations of gmcs in m 33 employed here were made using the bima millimeter interferometer and are presented by @xcite .
that study analyzed 20 fields in that galaxy with a resolution @xmath69 , targeting sources identified in the all - disk survey of @xcite .
the m 31 gmcs belong to a region along the northwest spiral arm of this galaxy , also observed by bima @xcite .
this region was chosen for its favorable observing geometry , strong dust lanes and active star formation .
both studies concluded that the gmcs in these disks were indistinguishable in their aggregate properties from the clouds in the inner milky way .
these observations constitute the most complete data sets for these galaxies , have sensitivity and resolution comparable to the dwarf galaxy data , and have been analyzed in an identical fashion . for the milky way we employ the catalog of gmc properties by @xcite .
this catalog consists of size , line width , and luminosity measurements for 273 galactic gmcs and constitutes a representative data set of milky way gmc properties . like us ( see [ method ] ) , @xcite use moments to measure cloud properties and isosurface - based methods to identify individual gmcs .
their study , however , does not make an explicit correction for the effects of sensitivity ( they use boundary isosurfaces of brightness temperature t@xmath70 to 7 k ) .
it also uses an outdated galactocentric distance for the sun ( 10 kpc ) .
it is unclear how to best correct correct their catalogued cloud properties to bring them into closer agreement with our methodology .
we use here the original numbers , and rely on the m 31 and m 33 data as a consistency check . as is apparent in figures [ size_linewidth ] , [ lumlw ] , and [ lumsize ] ,
these data are in good agreement with the standard larson relations derived for the milky way sample .
the challenge of any study that uses different instruments to study galaxies with a range of intrinsic properties and distances is to control the systematic biases introduced by the differences in the datasets .
we measure cloud properties using the algorithm described in detail by @xcite and summarized here .
this method measures luminosities , line widths , and sizes of gmcs in a manner that minimizes the biases introduced by signal - to - noise and resolution .
these biases are the achilles heel of comparisons of gmc measurements across different galaxies obtained by different telescopes .
@xcite proposed that moment measurements combined with beam deconvolution and extrapolation represent a robust way to compare heterogeneous observations of molecular clouds .
the algorithm is part of a package called cprops implemented in idl and is available upon request .
the remainder of this section summarizes the approach step - by - step .
the robustness , reliability , and consistency of the cprops methodology was tested by @xcite for a globally applicable set of algorithm parameters .
they found the algorithm performed very well for data with signal - to - noise ratios ( s / n ) @xmath71 , but cautioned that measurements of the properties of gmcs observed with low s / n remain uncertain , particularly when using the default set of algorithm parameters .
unfortunately , co is faint in dwarf galaxies , which makes it difficult to attain the optimal combination of s / n and spatial resolution
. therefore , only a few of the datasets discussed here meet the s / n@xmath72 criterion .
nevertheless , the data assembled here represent the best observations of gmcs in dwarf galaxies to date . as a consequence of the low s / n data , we made minor adjustments described below to the standard cprops parameters to improve the reliability of our results .
first , we identify regions of significant emission in each data set . for most of our data
these are defined as regions that contain pixels which satisfy the condition of having 3 consecutive velocity channels above @xmath73 significance ( i.e. a high significance `` core '' ) .
we extend these regions to include all adjacent data which has at least 2 consecutive channels above @xmath74 significance .
for a few data sets this did not yield the best results and we adjusted either the core or the edge conditions : a higher edge threshold was used in the high s / n n83 and n159 data sets ; lower core thresholds were needed for the low s / n ngc 185 and ngc 205 data sets ; a few other minor adjustments were also made .
these changes yielded stable properties for clouds of low s / n , and suppressed diffuse emission blending clouds together in the extremely high s / n case .
the signal identification step , often referred to as `` masking '' , can have a large impact on derived cloud properties and is difficult to motivate physically , especially in the low s / n regime common to much of our data .
thus , by necessity this step introduces a certain amount of subjectivity in the analysis .
once signal is identified , data is apportioned into individual clouds following @xcite . in all cases ,
we use the `` physical priors '' described in their appendix , which set the parameters of the decomposition to common values motivated by the properties of galactic gmcs .
this entails decomposing emission into clouds using effective spatial and velocity resolutions of 15 pc and 2 km s@xmath75 and ignoring even high - contrast substructure below these scales .
when it is not possible to achieve these resolutions because of limitations in the data , we use the actual spatial or velocity resolution of the observations .
when two potential clouds share an isosurface , the algorithm adopts a conservative approach to separating them .
the clouds are considered separate only if all of the following conditions are met : 1 ) each is large enough to measure meaningful properties , 2 ) each shows a minimum contrast between peak and edge , and 3 ) the decision to break the isosurface into multiple clouds has a significant effect on the derived cloud properties .
@xcite showed this approach to be conservative and robust in the presence of noise , minimizing the identification of low - signal , marginally resolved ( and thus often spurious ) clouds . after identifying the individual clouds ,
we calculate their size , velocity dispersion , and co luminosity .
we determine the major and minor axis of the cloud using principal component analysis and measure the second moments of emission distribution along these axes .
the rms size of the cloud ( @xmath76 ) is the geometric mean of these two moments .
the velocity dispersion ( @xmath2 ) is measured from the second moment along the velocity axis , and the luminosity via the zeroth moment ( sum ) over the cloud . at finite
signal - to - noise the second moment will underestimate the true cloud size , introducing a sensitivity bias .
although for galactic data it is possible to measure reliable properties by adopting a fixed brightness boundary for clouds ( e.g. , * ? ? ?
* ) , it would be impossible to compare measurements of different galaxies without removing this bias .
we avoid the sensitivity bias by measuring the size , velocity , and luminosity as a function of intensity isosurface and extrapolating this to the case of infinite signal - to - noise .
the size and line width of the cloud are extrapolated linearly , the luminosity is extrapolated quadratically ( see * ? ? ? * for a justification of the extrapolation orders ) .
we discussed in [ disks ] the corrections that we apply to the galactic sample to bring it into agreement with our methodology .
it is often the case for extragalactic observations that the size and line width of a gmc are comparable to the angular and velocity resolution of the telescope .
we correct for this by `` deconvolving '' both the telescope beam and the velocity channel profile .
we do so by subtracting their values from the extrapolated moment measurements in quadrature .
clearly this step does not come for free in signal - to - noise .
size measurements of clouds that are only marginally resolved suffer from substantial uncertainty , and this is accounted for in the error estimates . from the moment measurements we convert to physical units
radius , luminous mass , virial mass .
the velocity dispersion is given directly by the moment .
we convert the rms size to the spherical radius of the cloud using the factor @xmath77 , so that radii in this paper follow the @xcite definition : @xmath78 .
this factor can be motivated by considering a constant density spherical cloud and comparing its rms size to its radius .
we compute luminous masses , @xmath12 , using the formula @xmath79 where @xmath80 is the integrated @xmath15co flux of the molecular cloud measured in jy , @xmath81 is the distance to the source in mpc , and the coefficient corresponds to our adopted conversion factor @xmath82 , and includes the contribution of the cosmological he abundance to the total mass .
equivalently , for a co luminosity @xmath8 expressed in pc@xmath83 , @xmath84 we compute the virial masses under the assumption that each cloud is spherical and virialized with a density profile of the form @xmath85 .
this assumption is made for consistency with previous work , and we can not assess its validity with the data in hand .
thus the virial mass is given by the formula @xcite @xmath86 where @xmath2 is the cloud velocity dispersion in , @xmath3 is the spherical cloud radius in pc , and @xmath11 is the cloud virial mass in m@xmath19 .
implicit in this equation is the assumption that the virial parameter equals one .
other works , in particular those focused on milky way clouds , have treated clouds as ellipsoids and used the corresponding form of the virial theorem ( e.g. * ? ? ?
* ) , instead of the simplified equation [ virialmass ] .
because of the low resolution and s / n of extragalactic observations , this approach is untenable
. under these conditions , deconvolution becomes very difficult and the precise shape of the source is hard , if not impossible , to recover .
we note that the more complex measurements of cloud dimensions reduce to the @xcite rms size parameter for use in the @xcite formulation of the the virial theorem .
to relate the mass surface density , @xmath41 , to the coefficient of the _ size - line width _ relation , @xmath87 , we simply divide equation [ virialmass ] by the area of the cloud , @xmath88 , to obtain @xmath89 where the coefficient @xmath90 is in units of pc@xmath91 and the surface density is expressed in m@xmath19 pc@xmath28 .
the errors quoted in this paper are derived from bootstrapping the data in each cloud and re - deriving its properties .
this yields a realistic assessment of the uncertainty _ once signal has been apportioned into clouds_. we also include a 25% gain uncertainty in the luminosity , typical of the uncertainty in the flux calibration at millimeter wavelengths . in high
s / n cases , e.g. our n 83 and n 159 data sets , bootstrapping yields very low uncertainties . in these cases ,
virtually all of the uncertainty in the derived cloud properties rests with the apportionment of emission into clouds .
this uncertainty is not included in our results .
table [ bimatab ] presents our measurements of gmc properties in dwarf galaxies ( available in full in the electronic edition ) . in order the columns list : ( 1 ) the galaxy that contains the gmc ; ( 2 ) a number identifying the cloud ; ( 3 ) the transition , co @xmath45 or @xmath92 , being observed ; ( 4 ) the right ascension and ( 5 ) declination of the intensity - weighted cloud center ; ( 6 ) the mean lsr velocity of the cloud in km s@xmath75 ; ( 7 ) the log of the cloud radius , in parsecs ; ( 8) the log of the line of sight velocity dispersion of the cloud , in km s@xmath75 ; ( 9 ) the log of the luminous mass from equation [ mlum ] ; ( 10 ) the log of the virial mass from equation [ virialmass ] ; and ( 11 ) the peak brightness temperature in the cloud , in kelvin .
uncertainties from bootstrapping are listed with each measurement .
table [ restab ] summarizes some of our results on a per galaxy basis . in order the columns list : ( 1 ) galaxy name ; ( 2 ) number of measured gmc sizes in either co transition ; ( @xmath93 ) average and range of gmc masses , in units of 10@xmath94 m@xmath19 ; ( @xmath95 ) error - weighted average and range of the logarithm of , in units of our adopted galactic value ; ( @xmath96 ) error - weighted average and range of the coefficient of the _ size - line width _ relation , @xmath97 ; and ( 9 ) corresponding average gmc surface density from equation [ surfacedensity ] .
we checked for correlations between parameters in table [ restab ] and distance or resolution , and found none obvious .
the major bias present , which is unavoidable , is that there is a correlation between our minimum virial mass and distance that is simply due to sensitivity . in that sense
we sample different ranges of gmc mass in different galaxies .
considerations such as the nonlinearity of the _ mass - luminosity _ relation , for example , become important in the magellanic clouds where the cloud masses are small .
of the three larson relations perhaps the most fundamental one is that between cloud size and line width ( equation [ size_linewidth_eq ] ) , as it is mostly ( but not entirely ) independent of the details of the excitation of the tracer molecule .
the initial studies already pointed out the connection between the _ size - line width _ relation and expectations from the kolmogorov theory for subsonic turbulence in incompressible fluids @xcite .
later measurements in the inner galaxy allowed a better determination of the relationship , showing that the exponent is not 1/3 as it would be expected for a pure kolmogorov cascade , but rather the steeper @xmath98 used in equation [ size_linewidth_eq ] @xcite .
an index of @xmath99 is consistent with expectations for compressible supersonic magnetohydrodynamic turbulence ( also known as burger s turbulence , * ? ? ?
* ; * ? ? ?
@xcite show that this relationship extends also within gmcs , which is consistent with the multi - scale turbulence interpretation of the observed line widths .
the authors argue that because the _ size - line width _ relation is well - defined it implies the invariance and universality of the turbulent scaling relations . for clouds in virial equilibrium @xmath100 , thus the observed _ size - line width _ relation implies @xmath101 , or equivalently that all clouds have similar surface densities .
recent studies of cloud properties in the outer galaxy have shown that the _ size - line width _ relation breaks down for very small clouds . in these objects
the velocity dispersion @xmath2 becomes a constant independent of their size @xcite .
these authors find that outer disk clouds smaller than @xmath102 pc in radius have an almost constant velocity dispersion @xmath103 , while clouds larger than @xmath102 pc follow the usual scaling .
@xcite interpret this fact in terms of the dynamical equilibrium state of the small clouds .
these clouds have a virial parameter @xmath104 , representing their ratio of kinetic to gravitational energy , that is @xmath105 for masses below @xmath106 m@xmath19 .
thus these small clouds have an excess of kinetic energy over their self - gravity , are not in virial equilibrium , and in order to be bound they have to be confined by an external pressure of order @xmath107 k. the precise location of the break in the _ size - line width _ relation is then likely a function of the local pressure of the cloud environment .
similarly , analysis of the @xmath108co emission from gmcs in the inner 3 kpc of the milky way reveals a less steep _ size - line width _
relation than that obtained from @xmath15co studies , @xmath109 @xcite .
these authors conclude that although cloud complexes are bound , many of the clumps identified in their study are not gravitationally bound , probably due in part to the methodology of their clump identification algorithm . finally , @xcite measured the gmc properties near the galactic center , finding that those clouds have line widths a factor of 3 to 5 times larger than similar clouds in the disk and suggesting that this is due to the large pressure in this region .
these studies show that there are observed departures from the traditional _ size - line width _ relation in the milky way , in the direction of larger line widths for a given size .
figure [ size_linewidth ] shows our results for the ensemble of extragalactic clouds
. remarkably , dwarf galaxy cloud measurements are broadly compatible with the milky way _ size - line width _ relation , represented by the gray area and the small dots from the sample of @xcite .
detailed inspection of the data reveals , however , that there is a tendency for gmcs in dwarf galaxies to fall preferentially under the galactic relationship .
this is most apparent in a few clouds in the lmc , ic 10 , and most clouds in the smc , which are significantly inconsistent with the milky way .
note that these departures occur in the direction opposite from those described in the previous paragraph . while we may suspect
that cloud measurements yielding sizes @xmath110 pc are affected by blending ( such as those on the right side of the diagram ) , we have no reason to think that there are problems with the lmc , smc , or ic 10 , where we have some of the best spatial resolutions in the dataset . furthermore , the control sample of m 31 and m 33 gmcs is entirely consistent with the galactic _ size - line width _ relation and does not suffer from the same offsets . we discuss the particular situation of the smc , which exhibits the most convincing departures , in [ smc ] .
note that @xcite arrived at a different conclusion in their study of the smc . because the gmc samples do not substantially overlap it is difficult to assess the source of the discrepancy .
we attribute it mostly to the differences in the analysis methodologies , albeit it is possible that the samples are intrinsically different ( perhaps due to different radiation environments ) .
although it is not entirely correct to derive a unique _ size - line width _ relation for the ensemble of extragalactic data , we have determined a best fit relation . in order to avoid driving the result with the small formal error bars in the magellanic clouds , and to better represent the scatter in the data , we increase the errors in quadrature in both variables until we obtain a best fit @xmath111 value of unity .
the resulting relation , illustrated by the dashed blue line in figure [ size_linewidth ] , is @xmath112 where the data are fit in the log log plane with simultaneous errors in both variables , and the error bars are derived using a bootstrapping technique .
if we only use the dwarf galaxy data , the coefficient and the exponent are @xmath113 and @xmath114 respectively .
if the _ size - line width _ relation is interpreted in terms of virialized self - gravitating clouds supported mainly by turbulence , the compatibility between the galactic relation and our measurements implies that most gmcs in dwarf galaxies have surface densities not very different from galactic gmcs , for which @xmath115 .
extragalactic gmcs that fall under the galactic _ size - line width _ relation would have lower surface and volume densities than corresponding clouds in the milky way . the small - size , low velocity dispersion end of the dash - dotted line in figure [ size_linewidth ] corresponds to a drop of approximately a factor of 2 in the velocity dispersion , which translates to a factor of 4 decrease in surface density with respect to the galactic value .
this surface density may be appropriate for the most extreme galaxies in the sample ( c.f . , table [ restab ] ) .
if we fit a model that is proportional to @xmath116 , the sample of dwarf galaxies is consistent with an average surface density @xmath117 .
an alternative possibility is that the virial parameter , a measure of the ratio of kinetic to gravitational energy of the cloud @xcite , is different for these objects .
the virial parameter , @xmath118 , can be expressed in terms of the velocity dispersion following equation 27 in @xcite to obtain @xmath119 .
thus , a low coefficient in the _ size - line width _ relation can be attributed to a lower than galactic virial parameter . bound clouds in the milky way
have @xmath120 , while pressure - confined clouds exhibit @xmath121 a cloud in equilibrium with @xmath122 requires some additional type of support , presumably provided by magnetic fields . on the spatial scales studied here ,
such support is usually unimportant in the milky way .
figures [ lumlw ] and [ lumsize ] show the relations between @xmath15co luminosities and either velocity dispersion or cloud radius .
these luminosities can be related to the corresponding luminous mass @xmath12 for a galactic using eqs .
[ mlum ] and [ mlum2 ] .
the steep dependency on @xmath2 in equation [ lum_lw_eq ] is a direct consequence of the observed _ size - line width _ relation in combination with the assumption of virial equilibrium .
the coefficient in front of this equation reflects a combination of the area filling fraction by co clumps within the resolved cloud , and the kinetic temperature of the gas itself . inspection of figures [ lumlw ] and [ lumsize ] reveals that , despite the variety of environments sampled and the consistently lower - than - galactic metallicity of our sources , there is good agreement between the properties of extragalactic and galactic gmcs .
nonetheless , it is clear that there is an overall tendency for clouds in dwarf galaxies to be underluminous for their size , and clouds in the smc tend to be also overluminous for their velocity dispersion . in other words , they tend to be larger and ( in the case of the smc ) more quiescent ( as traced by their velocity dispersion ) than milky way gmcs of similar luminosity .
the fit to all the extragalactic data , conducted as in [ sizelw ] , shows that the _ luminosity - line width _ relation is @xmath123 where the shallower than galactic slope is driven mostly by the small smc clouds . a similar analysis but including only the dwarf galaxy data yields a coefficient @xmath124 and an exponent @xmath125 , thus very consistent with the results when m 31 and m 33 are included . the corresponding extragalactic _ luminosity - size _
relation is @xmath126 where the corresponding coefficient and exponent when fitting only the dwarf galaxy data are @xmath127 and @xmath128 .
this shows that the data are completely consistent with the galactic slope , but the relation is noticeably displaced toward lower luminosities for a given size .
figure [ lumsize ] suggests that , despite our efforts to correct for them , there may be a remaining methodological offset in the cloud size determination between the sample of @xcite and this study , of order 30% toward larger @xmath3 ( equivalent to a factor of @xmath99 toward lower luminosities ) .
unfortunately , establishing the origin of these offset would require a full reanalysis of the original galactic survey . with the quality of the data at hand it is difficult to assert whether methodological or real offsets are driving the observed shift toward larger cloud sizes ,
although it seems that the likely methodological offsets are too small to explain the level of discrepancy in many sources most notably the smc .
figure [ lco_mvir ] illustrates the behavior of luminosity and virial mass described by equation [ lco_mvir_eq ] for our sample of extragalactic gmcs . @xcite
discuss this empirical relation in some detail , showing that it is a natural outcome of the combination of the _ size - line width _ relation with the assumption of virial equilibrium when all clouds share similar kinetic temperatures and area filling fractions in their resolved co emission . despite the aforementioned tendency for clouds in dwarf galaxies to be underluminous for their sizes , their virial masses are , on average ,
only slightly underpredicted by their co luminosity . in the particular case of gmcs in the smc , because the observed departures from the standard _ luminosity - line width _ and _ luminosity - size _ relations occur in opposite directions , their larger sizes and smaller velocity dispersions for a given luminosity approximately cancel out in the virial mass calculation .
thus most reliable extragalactic clouds in our sample are remarkably compatible with equation [ lco_mvir_eq ] .
the fit to the entire extragalactic sample yields @xmath129 while the fit to the dwarf galaxies only has a coefficient of @xmath130 , and an exponent @xmath131 .
thus the extragalactic and galactic _ mass - luminosity _ relations are in excellent agreement for gmcs in the mass range @xmath132 m@xmath19 .
this suggests that , by comparison with equation [ mlum2 ] , is on average a factor of @xmath133 to @xmath134 larger than our adopted galactic value of @xmath135 ( although with large error bars ) .
we will discuss the consequences for and the presence of any metallicity trends in the co - to - h@xmath0 conversion factor in
[ cotoh2 ] .
we outlined in [ introresol ] some of the conclusions of the recent study of the local group by @xcite .
we employed here the same datasets for m 31 and m 33 , analyzing them with a later version of the same cprops algorithm . while @xcite emphasized completion to study the cloud statistics , the present analysis made use of the available data with the best possible combination of sensitivity and resolution , thus employing different datasets for the magellanic clouds and ic 10 .
furthermore , this study incorporates a large number of galaxies in and beyond the local group that were not analyzed in @xcite .
many of the results , however , are reassuringly similar : extragalactic gmcs broadly share the same properties observed in milky way gmcs , although the clouds in some galaxies ( notably the smc ) appear to be somewhat different from the remainder of the sample .
we will argue later ( [ smc ] ) that this is probably caused by the low metallicity of the parent galaxy .
lcccccccccc ngc185 & 1 & @xmath92 & @xmath136 & @xmath137 & -194.1 & & @xmath138 & @xmath139 & & @xmath140 + ngc205 & 1 & @xmath92 & @xmath141 & @xmath142 & -248.9 & @xmath143 & @xmath144 & @xmath145 & @xmath146 & @xmath147 + ngc2976 & 1 & @xmath92 & @xmath148 & @xmath149 & -31.3 & @xmath150 & @xmath151 & @xmath152 & @xmath153 & @xmath154 + ngc2976 & 2 & @xmath92 & @xmath155 & @xmath156 & -24.8 & & @xmath157 & @xmath158 & & @xmath154 + ngc2976 & 3 & @xmath92 & @xmath159 & @xmath160 & -19.5 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 + ngc2976 & 4 & @xmath92 & @xmath166 & @xmath167 & -6.2 & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath154 + ngc2976 & 5 & @xmath92 & @xmath172 & @xmath173 & 5.5 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath154 + ngc2976 & 6 & @xmath92 & @xmath178 & @xmath179 & 15.4 & @xmath180 & @xmath181 & @xmath182 & @xmath183 & @xmath165 + ngc2976 & 7 & @xmath92 & @xmath178 & @xmath184 & 20.8 & & @xmath185 & @xmath186 & & @xmath154 + ngc2976 & 8 & @xmath92 & @xmath187 & @xmath188 & -2.6 & & @xmath189 & @xmath190 & & @xmath154 + lcccccccr@@xmath191l ic 10 & 9 & @xmath192 & @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 & 108 & 31 + ngc 205 & 1 & @xmath198 & & @xmath199 & & @xmath200 & & 447 & 191 + smc & 11 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & 45 & 21 + lmc & 7 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & 123 & 39 + ngc 1569 & 1 & @xmath213 & & @xmath214 & & @xmath215 & & 170 & 209 + ngc 2976 & 7 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & 106 & 21 + ngc 3077 & 4 & @xmath222 & @xmath223 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & 166 & 39 + ngc 4214 & 3 & @xmath228 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & 102 & 44 + ngc 4449 & 1 & @xmath234 & & @xmath235 & & @xmath236 & & 105 & 33 + ngc 4605 & 3 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath205 & @xmath241 & 44 & 18 + m 31 & 44 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & 172 & 15 + m 33 & 14 & @xmath134 & @xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & 160 & 31 + @xcite discusses a self - regulated theory of star formation , where photoionization caused by the interstellar radiation field controls the rate at which clouds contract and form new stars . in this theory
clouds undergo contraction until the energy input by the newly formed stars balances the gravitational collapse .
support against contraction at the level of individual clumps is provided by their magnetic field .
this contraction happens at the ambipolar diffusion rate , as the neutrals diffuse through the ions which are anchored by the magnetic field .
the rate of diffusion is determined by the ionization fraction of the cloud , which in its outer regions ( which comprise most of its mass ) is set by the ultraviolet interstellar radiation field ( isrf ) , and in its inner regions is due to cosmic ray ionization .
the total rate of star formation of a cloud thus depends critically on its extinction and radiation environment .
clumps immersed in weaker isrfs require less extinction to collapse , and vice versa .
these ideas have considerable implications for the structure of gmcs as a function of environment , and it is of great interest to test their predictions . in the theory outlined by @xcite ,
extinction is the key parameter that regulates the equilibrium of clouds , not total column density .
accordingly , @xcite proposed as one of the key tests of the theory that clouds in galaxies with different metallicities and dust - to - gas ratios should obey different _ size - line width _ relations , as @xmath253 this expression is the analog of equation [ size_linewidth_eq ] , where the numerical coefficient now incorporates the cloud mean extinction , @xmath254 , and the ratio of extinction per hydrogen nucleus relative to the milky way , @xmath255 , so that for galactic clouds we recover the standard _ size - line width _ relation .
the prediction from photoionization - regulated star formation theory is that equilibrium molecular clouds in galaxies with low values of @xmath255 should have high velocity dispersions for a given size . in other words , clouds in low metallicity environments
need to be more massive to attain the characteristic extinction in their inner regions for a given isrf .
figure [ met_beta ] shows the average coefficient of the _ size - line width _
relation for each galaxy in our sample plotted as a function of metallicity , which is a good proxy for @xmath255 in equation [ photosf ] .
neither the placement of individual gmcs in the _ size - line width _ relation ( fig .
[ size_linewidth ] ) , nor fig .
[ met_beta ] show evidence of the systematic behavior predicted by the theory . if anything , clouds in ic 10 ( @xmath256 ) and the smc ( @xmath257 ) , which should have higher - than - galactic velocity dispersions by factors of @xmath258 , appear to have systematically lower - than - galactic values of velocity dispersion for a given size .
even if we discard the clouds that fall under the galactic _ size - line width _ relation as potentially out of equilibrium or perhaps magnetically supported , the remaining clouds do not show larger than galactic velocity dispersion .
it is important to realize that several of these gmcs exist in environments of larger - than - galactic isrf ( e.g. , * ? ? ?
* ) , so that the observed lack of a trend in @xmath259 is not directly caused by a cancellation between the effects of isrf and metallicity .
early support for photoionization regulated star formation was found by @xcite , who used iras 60 and 100 @xmath260 m data , a model of pdr regions , and h@xmath0 and [ cii ] measurements to derive the structure of molecular clouds in the magellanic clouds .
their determination of column densities , however , hinges critically on the estimate of the radiation field , which is derived from iras measurements . in the smc they model dust temperatures in the range of @xmath261 k , higher than the temperatures in the @xcite map which employs the same data .
we use the iris data to check the temperature towards lirs36 , one of their target regions , and find it to be 33 k instead of the 45 k derived by @xcite .
because the estimated radiation field is very strongly dependent on the assumed dust temperature , this alone suggests a @xmath68 times lower field than they adopt .
this is enough to bring their result into approximate agreement with our own , and makes the radiation field in this region intermediate between galactic and 30 doradus in the lmc .
a further concern with this analysis stems from the use of 60 @xmath260 m data to establish a temperature , as it is known that there is an important contribution from stochastically heated small grains at this wavelength ( particularly in the magellanic clouds , * ? ? ?
studies combining 100 @xmath260 m data with @xmath262 @xmath260 m @xcite determine typical temperatures of @xmath263 k for these smc regions .
while our analysis does not directly support photoionization - regulated star formation theory , there are several caveats that need to be considered before rejecting it . first , strictly speaking the increased column density prediction from photoionization - regulated star formation theory applies to the structures that form stars .
thus , it is possible that these observations do not sample the scales at which the column density enhancements occur .
second , in addition to assuming clouds in virial equilibrium , this prediction assumes that the ratio of mean clump to cloud extinction is similar to galactic clouds , and that the characteristic density ( at which cosmic ray ionization is @xmath264 ) is a similar fraction of the cloud density as in galactic clouds .
unfortunately , these parameters are almost completely unconstrained for extragalactic gmcs .
third , metallicity may not be a good linear proxy for @xmath255 .
this parameter folds in the effects of dust - to - gas ratio and extinction , and although dust - to - gas ratio likely tracks metallicity the dependence of extinction on metallicity may be important . in the smc @xmath265
@xmath26 mag@xmath75 , a factor of 6.5 larger than @xcite s adopted galactic value and compatible with the ratio of metallicities between the smc and the milky way @xcite . at ultraviolet wavelengths
, however , @xcite find that @xmath266 is larger than galactic by factors of @xmath267 . as a result of this , and assuming that the ultraviolet extinction and not is the parameter that matters at setting the cloud ionization level , we may probe a smaller range of @xmath255 than that assumed in fig .
[ met_beta ] .
note that although this consideration diminishes the discrepancy with the theoretical predictions it does not eliminate it .
fourth , equation [ photosf ] requires that co is present throughout the molecular material so that its velocity dispersion accurately samples the potential of the gmc .
this may not be the case in low metallicity environments if there are large h@xmath0 envelopes faint in co surrounding the co - bright portions of the gmc , and it may constitute the biggest limitation in our test of photoionization - regulated star formation . nonetheless , the result stands that there is no observed trend of increasing velocity dispersion in the _ size - line width _
relation for decreasing metallicity on the scales sampled _ clouds show strikingly similar properties in all galaxies_. using @xmath15co observations to measure the amount of molecular hydrogen requires assuming a co - to- conversion factor @xmath268 @xcite .
this factor incorporates the effects of abundance , excitation , and cloud structure averaged over a large area .
the optically thick @xmath51 transition can be used to trace the molecular mass of a cloud because it arises from the surfaces of clump ensembles within the telescope beam : by measuring we are counting clumps , and under very general conditions the co intensity will be proportional to the total molecular mass of the clump ensemble .
this is the essence of the co `` mist '' model discussed by @xcite . because encompassing all the relevant physics of the problem into ab initio calculations is extremely difficult and not well constrained , astronomers rely on empirical calibrations of obtained using a variety of techniques ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is reasonable to expect that will depend on the local properties of the interstellar medium ( ism ) ; volume density , temperature , radiation field , and metallicity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
observations throughout the galactic disk , however , constrain within a narrow range @xmath269 , with excursions of up to a factor of @xmath270 over this value , particularly at high latitudes @xcite .
this suggests that the local variations due to volume density , temperature , and radiation field are unimportant when cloud properties are averaged over a large area , which is the case for most extragalactic observations . in cases where the average conditions are very different from those in the galaxy , however
, differences in become apparent .
for example , mass estimates of molecular clouds subjected to extreme radiation fields , such as those in starburst environments , suggest that they have an conversion factor 4 to 20 times smaller than in the galaxy ( e.g. , * ? ? ? * ) . this can be understood in terms of a larger emissivity per co molecule in starburst galaxies , stemming from a higher excitation state due to larger physical temperatures and volume densities in these environments ( e.g. ,
another possibility is that a key assumption underlying the co - to- proportionality , that the co line width traces the mass , breaks down in very co - rich systems @xcite .
similarly , conversion factors a few times lower than that of the galactic disk are commonly observed in the centers of spiral galaxies , including our own @xcite .
a property of the ism that is expected to have a dramatic impact on the value of is its heavy - element abundance @xcite . unlike density or radiation field fluctuations , metallicity does not average out over the large areas sampled by a telescope beam .
in fact , dwarf galaxies , which tend to have subsolar metallicities , show little evidence for metallicity gradients at least in the examples where such gradients could be well measured ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
metallicity affects cloud structure in two ways : 1 ) directly , as smaller abundances of c and o translate into lower co yields , and 2 ) indirectly , as the dust - to - gas ratio is lowered , which in turn diminishes the formation rates and the shielding of molecular gas from the photodissociating effects of ultraviolet radiation .
the observational effects expected from these changes can be summarized as follows : as metallicity decreases , decreasing dust shielding pushes the c@xmath33 to co transition further into the molecular gas . in terms of a spherical clump ,
the radius of the @xmath271 surface for co becomes smaller as the metallicity decreases .
because at high - enough extinction most of the carbon in the gas phase turns into , which in turn becomes rapidly optically thick , the actual c / h ratio of the gas has only a small effect on the position of @xmath271 surface . for a typical galactic cloud , for example
, the c@xmath33/co transition occurs at @xmath272 while the @xmath271 surface of co occurs at column densities @xmath273 @xmath26 ( ) @xmath75 which translate into visual extinction of @xmath274 for a typical gas phase carbon abundance .
thus , almost immediately after becoming the dominant form of carbon , turns optically thick .
because of its large abundance , is self shielding in milky way molecular clouds @xcite .
co , however , is only mildly self shielding , although it also cross shields with at column densities @xmath275 @xmath26 @xcite .
if the conditions are such that self - shielding dominates over dust shielding at lower metallicities , the atomic to molecular /transition will not move into the cloud for decreasing dust - to - gas ratios . as a result ,
a molecular clump will not change its size with decreasing metallicity but its emitting core will diminish , yielding an increasing conversion factor ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
there are several observational calibrations of with metallicity , @xmath276 , in the literature , showing a range of behaviors .
most of the calibrations find an increasing with decreasing @xmath276 although the rate of increase varies greatly depending on the technique and the spatial resolution of the observations used .
a key datum in these calibrations is the measurement in the small magellanic cloud , as this is the galaxy with the lowest metallicity where co is reliably detected ( @xmath277 ; * ? ? ?
* ; * ? ? ?
in the first systematic study of this galaxy , @xcite used single - dish co observations and the assumption of clouds in virial equilibrium to determine that @xmath278 at a resolution of @xmath279 pc , while finding @xmath280 on scales of @xmath281 pc .
@xcite revisited the estimate of in the smc with new single
dish observations at a resolution of @xmath282 pc and found @xmath283 , in rough agreement with the previous results .
one important issue with these determinations is that it is unclear whether the large scale structures observed in co are self - gravitating , virialized molecular clouds .
the applicability of the virial theorem on scales of 100 pc and larger and the ability to accurately measure the size of the co structures , both necessary to apply equation [ virialmass ] , are problematic .
another issue with these determinations is that they do not account for the finite angular resolution of the telescope , which is in many cases comparable to the sizes of the clouds .
this will systematically bias these measurements in the direction of larger virial masses and larger .
subsequently @xcite used millimeter - wave continuum observations to obtain the mass of the quiescent cloud smcb1 - 1 and compare it with its co emission , reaching the conclusion that the dust - derived mass is 7 to 20 times larger than the cloud virial mass .
they suggest that this is located in an envelope of the cloud that does not emit brightly in co. very recently , @xcite used new far - infrared ( fir ) images of the smc obtained by the _
space telescope to model the dust emission and locally calibrate its dust - to - gas ratio , obtaining a map of in this galaxy .
comparison with the co emission indicates that the clouds are on average @xmath284 more extended than the co - emitting regions , so that co clouds are immersed in an extended molecular envelope of .
furthermore , over the volume occupied by co , @xcite find @xmath278 , while the overall for the entire smc is approximately twice that . among the calibrations for with metallicity , a number of them rely on the assumption of virial equilibrium to obtain the molecular mass of gmcs .
@xcite used interferometric co observations of m 33 and dwarf galaxies in conjunction with the aforementioned @xcite results to establish that @xmath285 , where @xmath276 is measured using the oxygen abundance , as o / h .
@xcite used several observations in the literature to determine that @xmath286 . employing some of the same interferometric observations analyzed here , @xcite determined a galactic value for in ngc 4214 , a galaxy where @xmath287 . a result of 60% of the galactic value
was obtained in ngc 3077 @xcite , a galaxy with approximately galactic metallicity .
similarly , @xcite used interferometric observations of the entire disk of m 33 and found no dependence of on metallicity over a range of 0.8 dex in @xmath276 ( a factor of 6 ) .
all these studies are based on virial mass techniques .
calibrations tend to be even more discrepant when other methods are employed .
besides the aforementioned results by @xcite , @xcite used and infrared astronomy satellite ( iras ) observations of several galaxies , estimating @xmath288 in the smc and finding a metallicity dependence @xmath289 , or @xmath290 when taking into account the local interstellar radiation field .
@xcite used measurements of the fir ( @xmath291p@xmath292p@xmath293 ) transition to estimate @xmath294 in some regions of the low metallicity dwarf galaxy ic 10 , and over 100 times the galactic value overall ( @xmath295 ) .
@xcite used stellar extinction in the lmc to measure @xmath296 , toward the low end of virial mass estimates for this source .
finally , @xcite found a milder @xmath286 dependence using a combination of virial and dust continuum methods . in summary , there are large discrepancies between the different authors and techniques , and in general estimates based on fir observations find stronger dependencies on metallicity than those based on virial arguments , although typically they also probe larger scales .
a problem with the available studies has been the lack of uniformity in the datasets , methods , and analysis techniques .
unexpectedly , the most uniform calibration in the literature ( that of * ? ? ?
* ) , where all the clouds are at the same distance , the spatial resolution is good ( 20 pc ) , and identical analysis is applied to the data finds that is independent of @xmath276 .
very recent observations , however , cast doubts on the magnitude of the metallicity gradient in this galaxy .
@xcite find that the metallicity gradient of m 33 is a factor of 3 shallower than previously accepted , which implies that the rosolowsky et al .
( 2003 ) gmc data probes a considerably smaller range of @xmath276 than previously thought .
thus mild metallicity dependencies of within this galaxy may be masked by the internal scatter of the measurements .
another potential issue is that similar gradients in the radiation field and metallicity with galactocentric distance have opposite effects and may conspire to produce an almost constant @xcite .
the fact that the overwhelming majority of our extragalactic gmcs with reliable property determinations are compatible with the empirical galactic relationship between @xmath8 and @xmath11 described by equation [ lco_mvir_eq ] ( [ luminosity_relations ] ) implies that we do not observe extreme departures from a galactic .
indeed , figure [ metxco ] shows our results for the extragalactic calibration of with metallicity , using virial gmc masses .
the gray region illustrates the approximate range of in the milky way found by @xcite , a factor of 2 around the nominal @xmath135 .
the thick lines show the dispersion of values for the gmcs in the @xcite sample with masses larger than @xmath297 m@xmath19 .
the symbols show the average of @xmath298 for all gmcs in each galaxy in our sample , with the corresponding vertical bars indicating the error and range of @xmath298 within a galaxy . the galaxy that shows the largest departure from galactic is ngc 205 , but with only one identified cloud we lack significant statistics .
the situation is similar for ngc 1569 .
note that although the gmcs in the smc show an average three times galactic , they are not unusual compared with galactic clouds of the same mass ( recall that equation [ lco_mvir_eq ] is not a linear relation , and galactic clouds with masses similar to our smc clouds have similar ratios of virial to luminous mass ; c.f . ,
figure [ lco_mvir ] ) .
thus , gmcs in dwarf galaxies are remarkably compatible with a galactic _ independent of their metallicity _ , and any metallicity trends appear to be much weaker than the dispersion of the measurements . to further quantify this statement
we have carried out a number of fits to the data .
we do not consider in the fits ngc 2976 ( because of its uncertain metallicity ) , and the milky way ( which may suffer from a systematic methodological offset ) .
a least - squares bivariate fit @xmath299 with simultaneous errors in both axes to the measurements for all remaining galaxies yields a slope @xmath300 and @xmath301 , with errors estimated using bootstrapping . increasing all errors proportionally to obtain @xmath302 yields @xmath303 .
similar fits to the data for the galaxies with the largest number of gmc measurements ( m 31 , m 33 , ic 10 , lmc , and smc ) yield @xmath304 ( @xmath305 ) and @xmath306 ( @xmath302 ) .
these results , plus the considerations in the paragraph above , confirm that there is no measurable trend in the resolved with metallicity present in the data . as we discuss below , however , this does not imply that there is no metallicity trend in the global ratio of co - to - h@xmath0 in galaxies . what about the possibility of clouds being more massive than the simple virial mass estimate
? we already discussed the observation that several clouds in ic 10 and the smc are not quite compatible with the galactic _ size - line width _ relation , in the direction of too small a line width for a given size .
we will argue in [ smc ] that a possibility is that these clouds are transient structures not in equilibrium that will collapse in a few myr due to the lack of turbulent support , or that they may be supported by magnetic fields in larger proportion than the typical galactic gmc . in either case
, such clouds could be removed from the sample since their line width may underestimate their true mass .
doing so slightly changes the averages for the smc and ic 10 , but does not significantly alter the results .
thus , a consistent and uniform analysis of the available data for resolved gmcs shows no evidence for an increasing with decreasing metallicity .
we note that this is unlike the scenario mentioned before , where similar gradients in radiation field and metallicity conspire to keep approximately constant in the disks of the milky way or m 33 . here
we have galaxies ( and particular gmc complexes ) that simultaneously have a higher - than - galactic radiation field and a lower - than - galactic metallicity , yet a very similar for their resolved gmcs ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we emphasize that these results correspond to the resolved . as we discussed in
[ calibrations ] , fir observations show that in galaxies such as the smc , co - bright cores are likely embedded in extended envelopes of that do not emit in co @xcite .
the global molecular gas content of a galaxy relative to its co luminosity may well steeply scale with its dust - to - gas ratio or metallicity , as it is strongly suggested by the available data on a few of these objects @xcite .
nevertheless , it appears that the structures that we are able to identify as individual gmcs by means of their co emission have co - to- conversion factors ( as well as larson relations ) that are approximately galactic , independent of their nebular metallicities .
we can understand the joint results from the co and dust - continuum observations in the following terms : in a low - metallicity environment such as the smc , co - bright clouds are the innermost portions of considerably larger h@xmath0 structures mostly devoid of co emission .
this scenario was suggested by @xcite and @xcite for the smc and is consistent with calculations of the effect of metallicity on the placement of the photodissociation fronts @xcite .
it also has been suggested as the situation in outer galaxy disks ( * ? ? ?
* but see also @xcite for a recent determination of h@xmath0 formation rates in diffuse gas ) .
it appears that ( as described for milky way gmcs by * ? ? ?
* ) these inner portions approximately follow the larson relations and , to co observers capable of resolving them , they appear very similar to galactic gmcs .
thus an approximately galactic factor correctly estimates the mass of the co - bright core , but underestimates considerably the mass of h@xmath0 in the entire gmc . in this scenario
the bulk of the mass has to be located in an envelope , surrounding the co emission .
this requires that the radiation field responsible for the photodissociation of co be external to the cloud , and that be effective at self - shielding even in an environment with low dust - to - gas ratio and consequently reduced formation rate .
column ( 11 ) in table [ bimatab ] gives brightness temperature at the cloud peak , @xmath307 , for each gmc , a quantity set by the excitation temperature , optical depth , and filling factor of co. although not part of the larson relations , this quantity is accessible in extragalactic gmcs and may yield insight into the physical state of clouds . for galactic clouds
the average brightness temperature is @xmath308 k @xcite .
many of our data exhibit notably lower @xmath307 than this .
this might indicate either lower excitation temperatures or a different optical depth of co , due to e.g. a lower cloudlet filling factor or other geometry well below our resolution .
however , the simplest explanation is that we observe gmcs with a ( spatially ) large beam and that beam dillution , i.e. a low co filling factor within the beam , lowers @xmath307 in more distant systems .
figure [ fig_tpeak ] shows that beam dilution indeed explains many of the the low observed @xmath307 .
we plot the mean and full range of @xmath307 for each galaxy as a function of the spatial resolution of the data .
we also show the expected @xmath307 for the case of a uniform brightness cloud with fwhm size 40 pc ( dashed line ) .
the dashed line is consistent with most of our data , including both local group galaxies ( the lmc , ic 10 , m 33 , and m 31 ) and more distant dwarfs ( ngc 2976 , ngc 4214 , ngc 4605 ) . for the more distant dwarfs ,
this highlights that while we resolve large ( @xmath309 pc ) structures in the co , these are likely blends of several gmcs .
a few extreme values of @xmath307 can not be immediately explained by beam dilution .
ngc 3077 appears unusually bright in co when observed with both bima and ovro , perhaps a sign the that ongoing starburst at the heart of this galaxy leads to higher excitation temperatures .
ngc 4449 shows a similarly high @xmath307 . in ngc 185 , ngc 205 ,
ngc 1569 , and the smc ( n83 ) @xmath307 is low despite excellent spatial resolution . in these galaxies ,
molecular clouds are physically smaller , have lower area filling fractions of co emission , and/or have lower kinetic temperatures than in large galaxies .
the first two might be caused by the low abundance of co or by diminished shielding from dust as discussed in previous sections .
a lower kinetic temperature remains a possibility , but is not supported by existing studies at least in the magellanic clouds @xcite .
clouds in the smc are indeed physically smaller than in most other galaxies , but that effect does not completely account for the observed differences ( see dotted line in figure [ fig_tpeak ] ) .
this suggests that the co emission has a lower area filling fraction : clouds in some of these galaxies are porous structures , where co arises only from high @xmath310 clumps .
we discussed in [ sizelw ] the peculiar situation of several of the least massive clouds in this study , mostly those belonging to the smc , that systematically exhibit velocity dispersions that are too small for their sizes according to the galactic _ size - line width _ relation .
we noted that , despite their small sizes , these objects display exactly the opposite behavior from that observed in the outer galaxy , where small clouds are confined by the external pressure @xcite . under the assumption of turbulence - supported clouds in virial equilibrium following a @xmath311 relation ,
the fact that many smc and a few ic 10 clouds lie a factor of two in @xmath2 below the _ size - line width _ relation implies that their surface densities are four times lower than galactic clouds , yielding @xmath312 .
recall that , for milky way clouds , the observed surface density translates into a visual extinction @xmath29 through the cloud . in low - metallicity galaxies such as the smc and ic 10 , where the dust - to - gas ratio is lower than in the milky way by at least a factor similar to their heavy - element deficit
, a reduction by a factor of 4 in surface density is compounded by another factor of @xmath37 in the dust - to - gas ratio implying that we would expect extinctions of order @xmath313 through the cloud , or @xmath314 at the cloud center .
we do not expect bright @xmath15co emission at such low extinction , where most co molecules would be photodissociated .
it is possible , however , that gmcs in the smc have a considerably lower co area filling fraction than their galactic counterparts .
thus , although the average surface densities and corresponding extinctions are are too low to allow the formation of co , it still exists in small , well shielded clumps within these structures .
we have discussed in the previous section that there is some evidence along these lines , since clouds in the smc indeed have lower brightness temperatures suggestive of smaller co area filling fractions . alternatively ,
if the hypothesis of virialized clouds supported chiefly by turbulent motions is not valid , the smc and ic 10 gmcs that fall under the _ size - line width _ relation show a deficit of turbulent kinetic energy with respect to similar size milky way clouds .
since galactic clouds are supported against collapse by a combination of turbulence and magnetic fields with turbulence lending the main support on the large scales @xcite , this apparent deficit of turbulent energy would translate into rapid collapse and subsequent star formation in a free - fall timescale ( @xmath315 myr , where @xmath316 is the mean volume density of the cloud ; @xcite ) unless there is significant cloud support provided by a magnetic field that is proportionally larger ( or better coupled to the cloud material , perhaps due to larger cloud ionization fractions ) than in otherwise similar galactic structures . to test this hypothesis
it is necessary to obtain cloud mass determinations independent of virial assumptions .
this is usually attained by employing the factor , modeling molecular line emission , or using measurements of the dust continuum and assuming a dust - to - gas ratio and grain emissivity .
the latter method was used by @xcite , who determined masses for several molecular clouds in the sw region of the smc .
these authors found the dust - derived masses for these clouds to be , on average , twice as large as their virial masses . under the assumption of long - lived clouds
, they attributed this fact to additional magnetic support by a @xmath317 @xmath260 g field .
this field is similar to that present in smaller structures ( a few parsecs in size with masses @xmath318 m@xmath19 ) in the milky way @xcite . along similar lines ,
@xcite modeling of the far - infrared emission in the smc to derive molecular gas surface densities found a mean surface density of molecular gas @xmath319 on 46 pc scales , very similar to that observed in milky way gmcs .
these observations suggest that surface densities in the smc are higher than what would be implied by cloud line widths under the assumption of turbulence - supported virialized clouds .
we argued in [ ourxco ] that the joint co and dust continuum data in the smc is best understood in terms of extended co - faint envelopes surrounding co - bright cores . in this interpretation ,
part ( possibly a large part ) of the mass or surface density obtained by dust continuum modeling is contributed by the extended envelope , and the mass of this envelope may not be accurately reflected in the kinematics of the co - bright core .
consequently , the high dust - derived surface densities and low observed co velocity dispersions may not necessarily require an enhanced magnetic support .
we expect that further joint analysis of high spatial resolution fir and co observations will shed light on the existence of extended envelopes @xcite .
we present and discuss a comprehensive set of high spatial resolution observations of co in low mass galaxies , obtained by a combination of interferometer and single - dish instruments .
although the data are heterogeneous , we analyze them in a consistent manner to obtain gmc sizes , velocity dispersions , and luminosities .
the analysis is performed using the algorithm described by @xcite , which does a good job at removing the biases due to dissimilar resolution and signal - to - noise .
we compare this uniform dataset of resolved extragalactic molecular cloud properties against those of gmcs in the three disk galaxies in the local group .
to do so , we analyze the interferometric maps of m 31 @xcite , m 33 @xcite , and the sample of galactic gmcs discussed by @xcite .
the main result of this study is that , remarkably , _ there are only small differences between co - bright gmcs in the milky way and gmcs in galaxies with metallicities as low as @xmath320 subject to a variety of physical conditions_. our sample of extragalactic gmcs follows approximately the same _ size - line width _ , _ luminosity - size _ , and _ luminosity - line width _ relations as galactic clouds
. such uniformity may be in part responsible for the observed invariance of the stellar initial mass function in galaxies . in any case
, this result underscores that the galactic larson relations provide a remarkably good description of co - bright giant molecular clouds independent of their environment , at least in the range of environments explored by this study .
although the larson relations are approximately galactic there are some significant departures .
gmcs in dwarf galaxies tend to be slightly larger than gmcs in the milky way , m 31 , and m 33 for a given co luminosity or velocity dispersion .
the largest departures occur in the smc , the galaxy with the lowest metallicity of the sample .
a possible interpretation , viable for most of our objects , is that gmcs in small galaxies have on average a surface density of @xmath321 rather than the canonical @xmath322 m@xmath323 pc@xmath28 observed in the galaxy . in the case of the smc , however , the implied surface densities , and consequently the central extinction in the gmcs , would be too low to expect bright co emission .
we explore three possibilities : these clouds are transient structures not supported by turbulence and will collapse in a few myr , or else they are supported by magnetic fields in a larger proportion than similar galactic clouds , or maybe the kinematics of the co cores do not reflect the presence of massive envelopes .
analysis of the properties of our sample of extragalactic gmcs shows that they do not accord with simple predictions from the theory of photoionization - regulated star formation @xcite . in the framework of this theory , clouds in lower metallicity environments
will have larger surface densities to attain similar extinction in their central regions , which can then decouple from the magnetic field and collapse .
such a trend would be evident in the coefficient of the _ size - line width _ relation , as expressed by equation [ photosf ] . as fig .
[ met_beta ] illustrates , we see no evidence for such a trend in our data .
we point out four possible caveats with these results : 1 ) it is possible that our observations do not probe the spatial scales on which these enhancements may occur .
2 ) perhaps other parameters that enter in the theory ( such as the cosmic ray ionization rate , for example ) change from galaxy to galaxy in a manner that conspires to keep the coefficient of the _ size - line width _ relation approximately constant .
3 ) maybe our identification of @xmath255 with metallicity is incorrect and we sample a smaller range of conditions than we think we do . or , 4 ) the co kinematics do not trace the full potential of the cloud . in any case
, we find a @xmath324 product that is approximately constant or even decreasing for decreasing metallicity , contrary to expectations .
finally , we address the matter of the dependency of the co - to - conversion factor , , on metallicity .
we find that the extragalactic gmcs in our analysis agree very well with the galactic _ luminosity - virial mass _ relation .
consequently , in our study of resolved gmcs we find no measurable change of : over a factor of 5 in metallicity most of our galaxies are compatible with a galactic @xmath23 within a factor of two , and there is no discernible trend with metallicity .
we emphasize , however , that this measurement is relevant on the scales of the individual , co - bright gmcs .
studies in the fir and at millimeter - waves in low metallicity environments ( particularly the smc ) suggest that these clouds are embedded in larger envelopes that are not traced by co emission but contain an appreciable mass of molecular gas , at least in the case of the smc @xcite .
these fir results are approximately consistent with virial estimates obtained on larger scales than the ones discussed in this paper @xcite .
we suggest that the fir and co observations can be simultaneously understood in the following terms : in low metallicity gas bright co emission is relegated to the density peaks .
observations that resolve those density peaks show properties that are similar to those of milky way gmcs that is the result of this study , and it is supported by studies of the structure of milky way gmcs @xcite . on larger scales ,
the cloud - to - cloud velocity dispersion as well as the dust opacity suggest that there is a large mass of likely molecular material that is not accounted for by the co intensity .
such regions would be co - bright in objects of higher metallicity .
studying the resolved properties of extragalactic gmcs is a challenging undertaking with the current generation of instruments , particularly for clouds in co - faint dwarf galaxies . in the near future
the deployment of the alma interferometer will make it possible to obtain more precise measurements on larger samples at farther distances .
studies such as the one presented here suffer from an important bias : we can only perform them on the brightest co peaks .
the detection of clouds as faint as those in the smc in co , for example , is currently impossible beyond the immediate neighborhood of the local group .
these improvements in instrumentation will soon provide powerful tests of theories of star and molecular cloud formation beyond our own galaxy .
we thank the referee , christine wilson , for her thoughtful and constructive comments .
we wish to thank lisa young , christopher l. taylor , mnica rubio , and frank israel for making the data on ngc 185 , ngc 205 , ngc 1569 , lirs36 , and lirs49 available for this study .
we also want to thank christopher mckee , eve ostriker , mark wolfire , mark heyer , mark krumholz , mnica rubio , stuart vogel , and andrew harris for valuable discussions and comments on drafts of this manuscript .
alberto bolatto wishes to acknowledge partial support from the national science foundation grant ast-0540450 .
|
we use high spatial resolution observations of co to systematically measure the resolved _ size - line width _ , _ luminosity - line width _ , _ luminosity - size _ , and the _ mass - luminosity _ relations of giant molecular clouds ( gmcs ) in a variety of extragalactic systems .
although the data are heterogeneous we analyze them in a consistent manner to remove the biases introduced by limited sensitivity and resolution , thus obtaining reliable sizes , velocity dispersions , and luminosities .
we compare the results obtained in dwarf galaxies with those from the local group spiral galaxies .
we find that extragalactic gmc properties measured across a wide range of environments are very much compatible with those in the galaxy .
the property that shows the largest variability is their resolved brightness temperature , although even that is similar to the average galactic value in most sources .
we use these results to investigate metallicity trends in the cloud average column density and virial co - to - h@xmath0 factor .
we find that these measurements do not accord with simple predictions from photoionization - regulated star formation theory , although this could be due to the fact that we do not sample small enough spatial scales or the full gravitational potential of the molecular cloud .
we also find that the virial co - to - h@xmath0 conversion factor in co - bright gmcs is very similar to galactic , and that the excursions do not show a measurable metallicity trend .
we contrast these results with estimates of molecular mass based on far - infrared measurements obtained for the small magellanic cloud , which systematically yield larger masses , and interpret this discrepancy as arising from large h@xmath0 envelopes that surround the co - bright cores .
we conclude that gmcs identified on the basis of their co emission are a unique class of object that exhibit a remarkably uniform set of properties from galaxy to galaxy .
|
single - transverse spin asymmetries ( ssa ) in hadronic processes have a long history , starting from the 1970s and 1980s when surprisingly large ssas were observed in @xmath1 @xcite and @xmath2 @xcite at forward rapidities of the produced hadron .
they have again attracted much interest in recent years from both experimental and theoretical sides @xcite . in particular , first measurements by the star , phenix , and brahms collaborations at rhic
have now become available @xcite , which extend the ssa observations from the fixed - target energy range to the collider regime . again
, large asymmetries were found in @xmath3 at forward rapidities of the produced pion .
meanwhile , experimental studies in deep inelastic scattering ( dis ) by the hermes collaboration at desy , by smc at cern , and by clas at the jefferson laboratory also show remarkably large ssas in semi - inclusive hadron production , @xmath4 @xcite .
data from compass for scattering off deuterons have been published as well @xcite , which show no large asymmetry . on the theoretical side ,
there are several approaches to understanding ssas within quantum chromodynamics ( qcd ) @xcite .
recent interest focuses on the role of partonic transverse momentum in creating the observed asymmetries .
transverse - momentum - dependent ( tmd ) parton distributions and fragmentation functions , and their relevance for semi - inclusive dis ( sidis ) , the drell - yan process , and single - inclusive hadron production at hadron colliders have been investigated in @xcite . compared to the normal integrated distributions ,
the tmd distributions provide much more information ; for example , some of them contain information on orbital angular momenta of partons in the nucleon and have also been linked to spatial distributions of partons @xcite .
the sivers function @xcite is one of these interesting tmd parton distributions .
it represents a distribution of unpolarized quarks in a transversely polarized nucleon , through a correlation between the quark s transverse momentum @xmath5 and the nucleon polarization vector @xmath6 .
the existence of the sivers function requires final / initial - state interactions , and an interference between different helicity fock states of the nucleon . in the absence of interactions ,
the sivers function would vanish by time - reversal invariance of qcd , hence it is often referred to as a `` naively time - reversal - odd '' distribution . as was shown in @xcite , the interactions are represented in a natural way by the gauge link that is required for a gauge - invariant definition of a tmd parton distribution .
interference between different helicity fock states implies nonzero orbital angular momentum @xcite .
both these properties motivate the study of this function .
the sivers function will contribute to the target ssa in semi - inclusive dis , but also to ssas in polarized @xmath0 scattering processes such as the drell - yan process and di - jet or jet - photon correlations .
we will discuss all these asymmetries in this paper .
the collins function is another `` naively time - reversal - odd '' function .
it is a transverse - momentum dependent fragmentation function and was introduced in @xcite .
it represents a correlation between the transverse spin of a fragmenting quark and the transverse momentum of the hadron relative to the `` jet axis '' in the fragmentation process . like the sivers function
, it vanishes when integrated over all transverse momentum .
indications of a nonvanishing collins effect have been found in semi - inclusive dis @xcite .
very recently results for measurements in @xmath7 annihilation to two hadrons have been reported , which give direct evidence for the collins effect @xcite .
the formulation and study of tmd functions is really useful only when they appear in processes for which qcd factorization at small transverse momentum can be established .
the processes , therefore , also need to be characterized by a large momentum scale , and there has to be additionally a small measured transverse momentum .
rigorous theoretical analyses of such reactions started from collins and soper s seminal paper @xcite , in which they proved factorization for di - hadron semi - inclusive processes in @xmath7 annihilation .
non - perturbative tmd fragmentation functions were defined and then further studied along with tmd parton distributions in @xcite .
the approach was extended to drell - yan dimuon production at hadron colliders @xcite .
more recently , these factorization theorems and the one for semi - inclusive dis have been carefully ( re-)examined in the context of the gauge - invariant definitions of the tmd parton distributions and fragmentation functions @xcite , paying in particular attention to the `` naively time - reversal - odd '' functions . in summary ,
qcd factorization has been established for three classes of semi - inclusive processes : di - hadron production in @xmath7 annihilation , semi - inclusive dis , and the drell - yan process .
it still remains to be seen whether factorization holds for more complicated processes in hadronic scattering , such as for di - jet ( di - hadron ) azimuthal angular correlations @xcite .
these processes , too , are characterized by a large scale ( the individual jet or hadron transverse momenta ) , and by an additional small transverse momentum related , for example , to the _ pair _ transverse momentum , or to the deviation of the two jets from being `` back - to - back '' .
note that this is in contrast to _ single - inclusive _ processes at hadron colliders like @xmath8 .
the spin asymmetries for such reactions are power - suppressed ( `` higher twist '' ) , and the theoretical description should be based on the methods developed in @xcite , where factorization in terms of higher - twist correlation functions was established . in this paper , we will use the factorization approaches at small transverse momentum discussed above to study the single spin asymmetries in semi - inclusive dis .
we will focus on the sivers and collins asymmetries which presently are the most interesting ones .
we realize that at the current stage it is difficult to apply the full factorization formulas developed in the literature in fitting the data and making predictions . instead
, we will make some approximations , neglecting higher order terms in the hard and soft factors . in this way we of course introduce some theoretical shortcoming , which we hope can be overcome in future studies .
our purpose is to look for the `` main effects '' , that is , to provide a quantitative description of the spin effects now studied experimentally , and to draw our conclusions on the collins and sivers functions from these .
another goal of this paper is to use the information gathered from semi - inclusive dis to make predictions for processes at rhic , which is now taking data in transversely polarized @xmath0 scattering . with the fitted parameterizations for the sivers functions that describe the hermes data very well
, we will calculate the sivers asymmetries for processes at rhic , including drell - yan dimuon production and di - jet and jet - plus - photon correlations .
we will demonstrate that these asymmetries are expected to be large at rhic and should therefore be closely investigated in the future polarized @xmath0 runs .
this would then provide further tests of the physical picture behind the ssa , and of our theoretical understanding .
to predict the ssas at rhic from the distributions fitted in dis relies on the factorization for the @xmath0 processes which , as discussed above , is so far only established for the drell - yan reaction .
it also relies on the universality of the tmd distributions .
this issue has been addressed in detail in @xcite .
it was found , for example , that the sivers functions for the drell - yan process will have an opposite sign compared to those for sidis , as a result of the behavior of the gauge - links in the functions under the time - reversal operation .
we will use this additional sign in our prediction for the drell - yan ssa at rhic based on our sivers function fitted to the dis data .
however , recent work has shown @xcite that the issue of universality appears to be much more complicated for the case of di - jet correlations , where the more involved color structure has profound consequences on the gauge links . as a result ,
the sivers functions for this reaction will differ from those in dis by more than just a sign .
this reservation notwithstanding , in order to obtain an order of magnitude estimate we will assume in this paper that the sivers functions to be used for di - jet correlations have the same sizes as those for the drell - yan processes , and opposite signs with respect to the dis sivers functions . in all calculations of cross sections and asymmetries below , we will use the grv lo parameterizations for the unpolarized quark distributions @xcite , and the kretzer set of unpolarized quark fragmentation functions @xcite .
these will also serve as starting points for our parameterizations of the sivers and collins functions .
the rest of the paper is organized as follows . in sec .
ii , we will review the basic formulas for the ssas in sidis , and make model parameterizations for the sivers and collins functions .
we then fit our parameterizations to the hermes data .
we will also compare our fit with the recent compass data on the sivers and collins asymmetries .
iii , we will calculate the sivers asymmetries for the drell - yan process and for di - jet and jet - photon correlations at rhic , using the fitted parameterizations from sec .
we summarize in sec . iv .
in this section , we will study the ssa in the sidis processes @xmath9 and @xmath10 , where @xmath11 represents a hadron observed in the final state
. we will compare the theoretical calculations of the asymmetries with the hermes measurements .
we will use some simple parameterizations for the sivers functions and the collins fragmentation functions , and fit these to the experimental data .
a comparison of our fit with the compass measurements will also be presented .
similar phenomenological studies of these asymmetries have also been performed in @xcite for the sivers case and in @xcite for the collins asymmetry , using the earlier hermes data .
we will start by briefly recalling the factorization formulas for the sidis process . for details ,
we refer the reader to ref .
as discussed in the introduction , we will make some simplifying approximations , in order to sharpen the constraints on the sivers and collins functions . the differential cross section for sidis , including the unpolarized part and the sivers and collins asymmetry contributions , may be written in the following form : @xmath12 \ , \end{aligned}\ ] ] where @xmath13 ( @xmath14 ) is the angle between the lepton plane and the @xmath15-hadron - plane ( and the transverse target spin ) , @xmath16 is the fraction of the incident lepton energy carried by the photon , and @xmath17 is the ( measured ) transverse momentum of the hadron . in order to compare with the experimental data , in the above formula and the following calculations , the azimuthal angles ( @xmath14 and @xmath13 ) are defined in the so - call virtual photon frame where the virtual photon is moving in the @xmath18 direction .
these definitions are different from those in @xcite where a hadron frame has been chosen to define these angles .
this difference has led to the different signs in the above formula , compared to that in @xcite .
the structure functions @xmath19 and @xmath20 will depend in general on @xmath17 , and on the invariant mass @xmath21 of the virtual photon , the bjorken variable @xmath22 , and on the fraction @xmath23 of the photon longitudinal momentum carried by the hadron observed in the final state . according to the factorization formula of @xcite , the structure functions can be factorized into the tmd parton distributions and fragmentation functions , and soft and hard factors . for example , for the unpolarized structure function , we will have @xcite @xmath24 this form is valid at low transverse momentum @xmath25 and is accurate at the leading power of @xmath26 .
as seen from the @xmath27-function expressing transverse - momentum conservation , the observed hadron s transverse momentum is generated by three contributions : the transverse momentum @xmath5 of partons in the nucleon , ( described by the tmd distribution @xmath28 ) , the transverse momentum @xmath29 acquired in the fragmentation process ( as expressed by the tmd fragmentation function @xmath30 ) , and the combined transverse momenta @xmath31 of ( large - angle ) soft - gluon radiation , embodied in the soft factor @xmath32 .
each of these transverse momenta is integrated in eq .
( [ crsecfac ] ) , but leaves its imprint in the distribution in @xmath33 of the observed hadron .
in contrast to the tmd functions in ( [ crsecfac ] ) , @xmath34 is a hard factor that depends solely on the large scale @xmath35 .
furthermore , @xmath36 is a renormalization scale , @xmath37 a gluon rapidity cut - off parameter , and @xmath38 is defined as @xmath39 , with @xmath40 the target hadron momentum , taken in the `` plus''-light - cone direction , @xmath41 , and @xmath42 a time - like vector conjugate to @xmath40 , i.e. , with only a `` minus''-light - cone component . for details , see @xcite , where also the related definition of @xmath43 is given .
similar factorization formulas as ( [ crsecfac ] ) can be written down for the single - transversely polarized structure functions @xmath44 and @xmath45 . for simplicity , in the following numerical calculations
, we will use the leading order expressions for the hard scattering and the soft factors , for which we have @xmath46 . at this order
, we may also neglect the @xmath47 dependences in the parton distributions and fragmentation functions .
all this brings us to the parton model picture for semi - inclusive dis @xcite .
however , we stress that higher order effects can be systematically and consistently studied only within the complete factorization framework . with the above approximations ,
the structure functions can be simplified to @xcite @xmath48 where a sum over all quark and antiquark flavors , weighted with the squared quark electric charge , is implicitly understood from now on .
@xmath49 denotes a unit vector in direction of @xmath50 . in the above equations , @xmath28 and @xmath51 represent the unpolarized quark distribution and fragmentation functions , respectively , @xmath52 the sivers functions , @xmath53 the collins functions , and @xmath54 the transversity distribution functions .
the definitions of the above distributions and fragmentation functions are consistent with the so - called `` trento conventions '' @xcite , while opposite in sign with respect to that used in @xcite for the sivers function . to optimize statistics , the experimental measurements of the asymmetries
are normally presented after integrating over the modulus of the hadron s transverse momentum @xmath33 .
after integration , the cross section can be written as @xmath55 where the various differential cross sections will depend on @xmath22 , @xmath23 and @xmath16 .
the dependence of the cross sections on the azimuthal angles results in the azimuthal asymmetries measured in experiment .
the unpolarized cross section is given by @xmath56 where @xmath57 and @xmath58 are the integrated parton distribution and fragmentation functions .
here we assume that we can obtain the integrated parton distribution by integrating over the transverse momentum in the corresponding tmd parton distribution .
this assumption will of course need to be modified if higher - order corrections are considered @xcite .
similarly , we can calculate the polarized cross sections . in these calculations , we further assume that the final hadron s transverse momentum is entirely related to the transverse - momentum dependence in the sivers and collins functions . the transverse momentum contributed by the other factors in the factorized formula ( [ crsecfac ] )
will give some smearing effects which may be viewed as `` sub - dominant '' .
after this approximation , we can write down the polarized cross sections as @xmath59 where @xmath54 is the integrated transversity distribution function . @xmath60 and
@xmath61 are defined as @xmath62 the above formulas ( [ eq4])-([eq7 ] ) will be used in the following calculations to study the experimental data for the asymmetries as functions of @xmath22 and @xmath23 . before doing so , we need to set up models for the collins and sivers functions .
we would like to add one more comment before we proceed . in the derivation of eq .
( [ eq6a ] ) , we have omitted the transverse momentum dependence of the fragmentation function , which we referred to as `` sub - dominant '' .
this `` sub - dominant '' contribution could become important at small @xmath23 where we can not neglect the influence of the transverse momentum in the fragmentation process .
for example , for a typical transverse momentum of the final state hadron of @xmath63 mev , the quark transverse momentum could be as large as @xmath64 gev at @xmath65 , much bigger than the typical value of the intrinsic quark transverse momentum for the sivers function , which is of order of a few hundred mev .
this means that at small @xmath23 our approximation will break down , and the transverse momentum in the fragmentation function will be important .
this effect will smear out the polarized cross section and suppress the asymmetry .
indeed , when assuming a gaussian transverse momentum dependence for both the distribution and the fragmentation functions , an additional factor of @xmath23 appears , suppressing the polarized cross section at small @xmath23 @xcite . as a consequence , we should be cautious to apply eq .
( [ eq6a ] ) at small @xmath23 . on the other hand , the above drawback does not apply to the case of the polarized cross section for the collins contribution , eq .
( [ eq6b ] ) , where we omit the transverse momentum dependence in the distribution .
this is because there is no kinematic enhancement associated with the intrinsic transverse momentum in the parton distribution , compared to the fragmentation case .
there exist by now quite a few model calculations for the quark sivers functions in the nucleon @xcite .
the results of these vary rather widely . here
, we will instead adopt simple parameterizations for the sivers functions and fit these to the hermes data .
we choose a form that has only a single free parameter for each flavor ; the present data probably do not yet warrant a more complex form .
our parameterization is as follows : @xmath66 where in both equations @xmath67 is the unpolarized @xmath68-quark distribution .
we assume that only the quark sivers functions are non - zero , and that the antiquark ones vanish .
this assumption will of course likely need to be modified at small @xmath69 . in the above parameterization , the factor @xmath69 on the right - hand - side represents the valence nature of the sivers function , whereas the factor @xmath70 denotes an expected suppressed behavior of the function at large @xmath69 could actually be as strong as @xmath71 @xcite .
in our fit , most data are in the intermediate range of @xmath69 , and the @xmath72 limit is not really reached . in any case , the power of @xmath70 will be modified by logarithms . ] for the collins asymmetry , we have two sets of unknown functions : the transversity distributions and the collins fragmentation functions . from the present experimental data on the collins asymmetries
we can not obtain constraints on both of them simultaneously .
here we adopt a parameterization for the transversity function @xcite that is based on saturation of the soffer inequality @xcite .
we note that this parameterization represents an upper bound for the transversity functions for the quarks . as for the sivers functions ,
there have also been several model calculations for the collins functions @xcite , showing rather wide variations .
again , we will just use a simple parameterization for the @xmath23 dependence of @xmath73 .
the flavor dependence of the collins functions is important since one would like to describe the asymmetries for different hadron species . from the theory side
, one could get constraints for the flavor dependence based on momentum conservation in the fragmentation process : the schfer - teryaev sum rules @xcite .
these sum rules state that the odd - moment ( for the intrinsic transverse momentum ) of the collins function vanishes when the function is summed over all hadron states .
because the sum rules only involve the integrals over @xmath23 and @xmath74 ( with a weight @xmath74 ) of the collins functions , one can not obtain from them more detailed constraints , e.g. , for the @xmath23 or @xmath74 dependences . in the following
we will motivate a simple conjecture for the collins functions , based on quark - hadron duality in the fragmentation process .
this will provide us with additional constraints .
the main result is that _ any quark collins fragmentation function is very small when summed over all final hadrons_. for example , the @xmath68 quark collins functions to all hadron final states will satisfy @xmath75 the above equation is motivated as follows .
the collins function is defined as @xmath76 where @xmath77 and @xmath78 . here
, the final state hadron has been taken to have a large light - cone `` minus '' momentum component @xmath79 [ we remind the reader that @xmath80 denotes the momentum of the observed hadron , while @xmath81 is the ( integrated ) transverse momentum that the hadron acquires in the fragmentation process relative to the fragementing parton ; see eq .
( [ crsecfac ] ) ] . finally , @xmath82 is the gauge link along the light - cone direction conjugate to @xmath80 . here , now ,
if we sum over all hadrons in the above collins fragmentation functions , the intermediate hadronic states can be replaced by a quark or a quark plus gluons ( or quark - antiquark pairs ) using quark - hadron duality arguments .
we then get the following equation : @xmath83 whose validity rests on the argument of quark - hadron duality for the fragmentation process .
duality - breaking effects will somewhat modify the above equation .
the right hand side of eq .
( [ cdual ] ) may be viewed as a quark collins fragmentation function into a quark ( or antiquark / gluon ) state .
the helicity - flip required for a non - vanishing collins function is then possible because of a finite quark mass .
thus , we approximately expect @xmath84 if we further assume that the fragmentation functions for @xmath68 and @xmath85 quarks to strange mesons are suppressed relative to those into pions , we can have even stronger constraints for the pion collins functions : @xmath86 where @xmath28 represents any flavor of @xmath68,@xmath85 quarks and their antiquarks
. further simplification of the above equation can be derived by considering isospin and charge symmetry relations between the different fragmentation functions .
for example , we will have the following relations , @xmath87 \ .\end{aligned}\ ] ] here @xmath88 and @xmath89 represent the `` favored '' ( in the sense that the leading fock state of the hadron contains the parent quark flavor ) and `` unfavored '' ( where it does not contain it ) fragmentation functions , respectively . substituting the above relations into eq .
( [ collinsconstr ] ) , we will obtain @xmath90 which means that the unfavored collins function is approximately equal to the favored one with opposite sign .
this result is counter to the usual notion in the literature that the favored fragmentation functions should be much larger than the unfavored ones .
it of course crucially depends on the validity of the approximations made in the above derivation and will be subject to some corrections .
we note , however , that this observation has also support from a string model description for the collins fragmentation function @xcite . as a test
, we will treat in the following the favored and unfavored collins functions free from the constraint ( [ ucollins ] ) , and fit them to the data to see if they naturally satisfy the above relation or not . to parameterize the collins functions , we use the following two sets of functional forms , @xmath91 the @xmath18 factor in these parameterizations represents the vanishing of the collins function at small @xmath18 , and the @xmath92 factor follows arguments made in @xcite .
the difference between these two sets is that for set i we parameterize both favored and unfavored collins functions in terms of the favored unpolarized quark fragmentation function , while for set ii we parameterize the unfavored collins function using the unfavored unpolarized quark fragmentation function .
set i is inspired by the constraint of eq .
( [ ucollins ] ) ; note that this ansatz is expected to violate the positivity constraints at very large @xmath23 . on the other hand , set ii
respects the positivity constraints , provided @xmath93 . in the following
, we will fit the hermes data with these two sets of parameterizations .
we will now calculate the collins and sivers asymmetries using the functions specified above , and fit the free parameters to the new hermes data on the collins and sivers asymmetries @xcite . here
we will use @xmath94 gev@xmath95 , which is the average for the hermes kinematics .
we choose @xmath96 in the unpolarized parton distribution and fragmentation functions .
we will then compare our fit results to the recent measurements by compass @xcite .
for the fitting , we use the cernlib minuit routine @xcite . as a function of @xmath22
, the sivers asymmetry can be calculated from the following formula : @xmath97 where the minus sign results from the minus sign in the polarized differential cross section in eq .
( [ eq4 ] ) . since the @xmath16 integral
is the same for the numerator and denominator , it cancels out .
moreover , the integral over @xmath23 can be factored out as a consequence of the approximations that led to eqs .
( [ eq5 ] ) and ( [ eq6a ] ) , and the sivers asymmetry will be proportional just to the ratio of the sivers functions over the unpolarized quark distributions , summed appropriately over flavors .
the sivers asymmetry as a function of @xmath23 can be calculated similarly .
error of the fitted parameters .
note that the data have not yet been corrected for acceptance and smearing .
[ fig1]_,height=377 ] -0.4 cm in fig .
[ fig1 ] , we show the results of our fit of the sivers asymmetries for @xmath98 and @xmath99 to the hermes data @xcite .
for the two free parameters the fit gives @xmath100 with @xmath101 . the band in each plot of fig .
1 corresponds to a 1-@xmath102 error in the determined parameters .
one can see that the @xmath68 quark sivers function appears somewhat better constrained by the data than the @xmath85 quark one .
this is readily understood from the fact that @xmath68 quarks in dis enter with the charge factor @xmath103 , so that in scattering off a proton target the @xmath68 quark distribution is particularly selected .
another feature is that the @xmath85 sivers function comes out larger ( by about a factor of two ) than the @xmath68 quark one , and with opposite sign .
this behavior is quite different from model calculations @xcite .
the result is due to the fact that the hermes sivers asymmetry for @xmath99 is much smaller than that for @xmath98 .
theoretically , however , @xmath99 production should also have a significant contribution from @xmath68 quarks , because one finds @xmath104 for the fragmentation functions when integrated over the experimentally relevant region @xmath105 . to obtain a much smaller asymmetry for @xmath99 than for @xmath98
, there then have to be fairly strong cancellations between the @xmath68 and @xmath85 quark sivers functions .
we note that the signs we find for our sivers functions are consistent with expectations in @xcite , where they were qualitatively related to the opposites of the quark contributions to the proton anomalous magnetic moment .
figure [ fig2 ] shows predictions for the @xmath106 sivers asymmetries as functions of @xmath22 and @xmath23 , based on our fits for the sivers functions .
we note that our prediction for the @xmath106 asymmetry is nearly independent of @xmath23 .
this is because the @xmath68 and @xmath85 quark fragmentation functions for @xmath106 are the same , and because in our approximation the distribution and fragmentation functions are decoupled .
however , we have to keep in mind that this decoupling might break down at small @xmath23 , as we discussed before .
this could be tested by future hermes data .
production at hermes .
[ fig2]_,height=226 ] -0.4 cm in @xcite , earlier hermes results @xcite for the sivers asymmetries were fitted .
the methods somewhat differed from ours . in @xcite , a particular transverse momentum dependence
is assumed for the sivers functions and the unpolarized quark distribution and fragmentation functions . with more free parameters
the experimental data are fitted equally well , and the @xmath68 and @xmath85 valence sivers functions are obtained from the fit . in @xcite ,
the asymmetries weighted with the transverse momentum of the hadron were used for the fit . both fits find a large @xmath85 quark sivers function with opposite sign relative to the @xmath68-quark one .
we have also checked that our fit results for the sivers functions are consistent with these fits within the current large uncertainties , where we notice that the sivers function in @xcite has an opposite sign compared to ours and to the `` trento conventions '' @xcite .
the compass collaboration also has measured the sivers asymmetry @xcite , separately for positively and negatively charged hadrons , produced off a deuteron target . to simplify the comparison with their data ,
we assume that the leading hadrons are mostly pions .
we calculate the sivers asymmetries for @xmath98 and @xmath99 in the kinematic region of the compass experiment , using the above fitted sivers functions for @xmath68 and @xmath85 quarks , and compare to their data for leading positive and negative hadrons , respectively .
we show this comparison in fig .
one can see that our calculations based on fits to the hermes data are also consistent with the compass data , within error bars .
we note that for the kinematical region of the compass experiment , our predicted sivers asymmetries for a deuteron target are very small , except in the large-@xmath69 valence region . the smallness of the sivers asymmetry is again related to cancellations between @xmath68 and @xmath85 contributions , which for deuterons enter in a different combination than for a proton target
. it will be very interesting to check these predictions with future compass data for a proton target .
thanks to the higher @xmath21 , such data would also help in confirming the leading - twist nature of the sivers and collins asymmetries . _ , height=377 ] -0.4 cm we next turn to the collins asymmetry . here
we follow a similar procedure as we did for the sivers case above . as we mentioned earlier , the situation is more complicated because of the fact the nucleon transversity densities are currently not known , and we need to resort to a model or ansatz for the latter . as described above , we will use the parameterizations for the quark transversity distributions of @xcite , which represent upper bounds for the densities .
we will fit to the hermes data using the two sets of simple parameterizations for favored and unfavored collins functions given in eq.([collinspar ] ) . for the collins asymmetries , using the set i parameterization of the collins functions .
the data @xcite and the theory curves are for the so - called lepton beam asymmetries .
note that the data have not yet been corrected for acceptance and smearing.[fig4]_,height=377 ] -0.4 cm , for the set ii parameterizations for the collins functions .
[ fig5]_,height=377 ] -0.4 cm the asymmetry as a function of @xmath22 is calculated from the formula @xmath107 where again the minus sign comes from the sign in the polarized differential cross section , eq .
( [ eq4 ] ) . @xmath108 and
@xmath109 represent the fragmentation functions integrated over the accessed region in @xmath23 .
kinematic cuts impose a correlation between @xmath22 and @xmath16 , and the integral over @xmath16 will depend on @xmath22 . in the experimental analysis ,
the data for the collins asymmetries are presented in two different ways .
one is to give results in terms of the virtual - photon asymmetry , factoring out the term @xmath110 .
the other way is to give the directly measured lepton - beam asymmetry . in our calculations
, we follow the latter way .
we neglect the contribution of longitudinal photons to the unpolarized cross section , which hermes has considered in the analysis of the virtual - photon asymmetries @xcite . in view of the overall uncertainties ,
this is a minor effect , as we have checked by comparing also to the virtual - photon asymmetries . from the fit to the lepton - beam asymmetry data , we get the two fit parameters as follows : @xmath111 with @xmath112 for the set i and set ii parameterizations , respectively .
the fit results are shown in figs .
[ fig4 ] and [ fig5 ] , compared to the hermes data .
both fits are of the same quality . in figs .
[ fig9 ] and [ fig10 ] , we plot the fitted favored and unfavored collins functions ( times @xmath18 ) for sets i and ii respectively . note that we multiply the favored ones by @xmath113 to compare their magnitudes . for comparison
, we also show the corresponding unpolarized quark fragmentation functions @xcite .
it is evident that the two sets of collins functions indeed both satisfy the positivity constraints .
the equal quality of the fits obtained for sets i and ii implies that the current experimental data neither necessarily support the constraints we derived in eq .
( [ ucollins ] ) , nor do they rule them out . however , from both fits we indeed find that in a quite large range of @xmath23 the unfavored collins function has the same size as that of the favored one with opposite sign .
a similar conclusion was obtained from a fit to this asymmetry using the transversity functions calculated in the chiral quark model @xcite .
we hope that higher - statistics data will become available in the near future that will test the relations .
favored and unfavored collins fragmentation functions . also shown are the unpolarized quark fragmentation functions from kretzer s parameterizations .
[ fig9 ] _ , height=226 ] -0.4 cm for the set ii collins fragmentation functions .
[ fig10]_,height=226 ] -0.4 cm figures [ fig6 ] and [ fig6p ] show predictions for the @xmath106 collins asymmetries as functions of @xmath22 and @xmath23 , based on our fits for the collins functions . from these plots , we find that the asymmetries are very small for both sets of the collins functions , because of strong cancellations between the contributions from favored and unfavored collins functions .
we note that preliminary data from hermes indeed indicate that the @xmath106 asymmetry is consistent with zero @xcite .
higher - statistics data on this will be highly interesting for testing the above conclusions .
production at hermes with set i. [ fig6]_,height=226 ] -0.4 cm production at hermes with set ii .
[ fig6p]_,height=226 ] -0.4 cm as before , we also compare our fit to the compass measurements @xcite for the collins asymmetries , where only the virtual - photon asymmetries are presented .
we also note that the convention for the collins asymmetry used by the compass collaboration is different from that used by hermes . in our calculations
, we have made the relevant modifications in order to compare with the compass data .
we show these comparisons in figs .
[ fig7 ] and [ fig8 ] for our set i and ii collins functions , respectively . as for the sivers case above
, there is good consistency .
the overall asymmetries are again small because of the deuteron target used , and because the assumed transversity sea quark distributions are small .
future high - statistics compass data for a proton target would be highly interesting . _
, height=377 ] -0.4 cm , but for set ii of collins fragmentation functions .
[ fig8 ] _ , height=377 ] -0.4 cm
an important issue in the study of hard - scattering processes is the universality of the nonperturbative objects : the parton distributions and fragmentation functions . in the case of single - spin asymmetries ,
if universality holds , the sivers functions obtained from , for example , sidis can be used to predict single - spin asymmetries in @xmath0 or @xmath114 scattering .
the universality between different classes of processes is a complicated and interesting issue that has attracted much interest recently @xcite .
as we described earlier , it was found that , while strict universality is violated already when going from sidis to the drell - yan process , the sivers functions for the two processes differ only by a sign .
it is therefore possible to use the fitted sivers function from the last section and to predict the sivers single - spin asymmetry for the drell - yan process at hadron colliders .
more complicated processes in hadronic scattering , such as the ssa in di - jet angular correlations @xcite , are not yet completely understood at present , as far as factorization and universality are concerned .
progress has been made recently @xcite ; it appears that similarly to the drell - yan process universality is violated only by terms that are calculable from the color structure of the partonic scattering and hence may be taken into account in phenomenological analyses .
below , we will also give estimates for sivers contributions to ssas in di - jet correlations and in jet - plus - photon correlations at rhic , using the sivers functions of section ii , and the usual unpolarized hard - scattering functions . in the light of ref .
@xcite , we expect that our estimates will likely need to be revised once these reactions will be completely understood in the context of factorization and universality , to take into account the appropriate factors embodying the nonuniversality of the sivers functions .
we shall briefly return to this point below . in this subsection
, we will calculate the sivers single - spin asymmetry for the drell - yan process at rhic , using the fit result of the last section [ see eqs .
( [ eq9 ] ) and ( [ eq19 ] ) ] . as just discussed , one has @xmath116 after integrating out the lepton angles in the rest frame of the virtual photon , we obtain the following differential cross section for the drell - yan process : @xmath117\ , \end{aligned}\ ] ] where @xmath118 is the invariant mass of the lepton pair , @xmath119 the virtual photon s transverse momentum , and @xmath16 its rapidity .
@xmath120 is the difference between the azimuthal angles of the virtual photon and the transverse polarization vector in a frame where the polarized hadron is moving in the @xmath18 direction . at low transverse momentum , @xmath121 and @xmath122
are related to the mass and rapidity through @xmath123 and @xmath124 where @xmath125 is the hadronic center - of - mass energy squared . according to the factorization theorem @xcite , the hadronic tensors
@xmath126 and @xmath127 can be factorized into the tmd parton distributions , and soft and hard - scattering factors . again , neglecting the soft factor and using the born expression for the hard part , we obtain simple expressions for the tensors : @xmath128 where @xmath129 is now the sivers function for the drell - yan process .
a further approximation can be made by integrating out the transverse momentum @xmath130 , but keeping the dependence on azimuthal angle .
the differential cross section can then be written as @xmath131\ , \ ] ] where @xmath132 and @xmath133 in fig .
[ fig11 ] , we plot the @xmath134 asymmetries as functions of the photon rapidity @xmath16 and the invariant mass @xmath118 . from this plot , we see that the sivers ssa asymmetry for the drell - yan process at rhic is expected to be sizable for large rapidity , and should be measurable at rhic if enough statistics can be accumulated in transverse - spin running .
we note that the sivers asymmetry in the drell - yan process is also a particular focus for proposed measurements in polarized @xmath114 scattering at the planned gsi - fair facility @xcite . and invariant mass @xmath118 .
[ fig11]_,height=264 ] -0.4 cm other interesting observables at hadron colliders from which one can access the intrinsic transverse - momentum dependence of parton distributions are `` back - to - back '' correlations between two jets @xcite .
more specifically , we are interested in situations in which the sum of the two jet transverse momenta , @xmath136 ( or a component or projection thereof ) , is measured , while both @xmath137 and @xmath138 individually are large .
as for the drell - yan process discussed above , this is usually a small transverse momentum , much smaller than the large scales @xmath139 characterizing the overall process , and a special factorization may apply .
let us first , however , consider the cross section integrated over @xmath140 . here
, collinear factorization applies , and the di - jet cross section has the parton - model expression @xcite , @xmath141 where @xmath142 is the differential cross section for the partonic process @xmath143 , with @xmath144 the appropriate parton distribution functions .
we have defined the transverse momentum @xmath145 denote the rapidities of the two jets .
the kinematics are as follows : @xmath146 here , @xmath147 , @xmath148 , and @xmath149 are the usual partonic mandelstam variables appearing in the partonic cross sections for the reactions @xmath143 , namely @xmath150 , @xmath151 , @xmath152 , in obvious notation of the partonic momenta .
the leading order contributions produce the di - jet pair exactly balanced , that is back - to - back in the partonic center - of - mass frame .
an imbalance in the transverse direction is generated by higher - order qcd corrections . at small but nonzero imbalance between the two jets
, the dominant contributions will come from the intrinsic transverse momenta of the initial partons . as a model
, we will generalize the above factorization formula to the case of small @xmath140 , in analogy with the sidis and drell - yan cases discussed earlier , taking into account the various contributions to the transverse momentum dependence coming from the parton distributions and soft factors : @xmath153 where @xmath154 is a soft factor for the process
@xmath155 , while @xmath156 is the hard part of the reaction , related to lowest order to @xmath142 .
we emphasize the overly simplistic character of eq .
( [ jeteq ] ) as it stands .
the detailed factorized form ( if it exists ) will likely be different ; in particular , one expects an interplay of the color structures of the soft factors and the hard parts , as found in resummation studies for jet cross sections @xcite . in a similar fashion
, we write the sivers - type contribution @xmath157 to the single - polarized cross section , @xmath158 where @xmath159 is the unit vector in the @xmath18-axis direction . in a factorized form
, we will get : @xmath160 we can further simplify the polarized cross section by evaluating the expression @xmath161 where @xmath162 and @xmath163 are the azimuthal angles of the two jets relative to the polarization vector @xmath6 , and @xmath164 the angle between the two jet transverse momenta .
all these azimuthal angles are defined in a frame that the polarized proton is moving the @xmath18 direction .
in the above derivation , we have used the approximations latexmath:[$|p_{1\perp}|\approx
|p_{2\perp}|\approx which are valid at small @xmath119 ( @xmath166 is close to @xmath167 ) . from the above result
, we can see that there are two terms contributing to the spin asymmetry : one is with @xmath168 and the other with @xmath169 .
the first term has a sign function associated , which gives a positive contribution when @xmath166 is smaller than @xmath167 and a negative one otherwise . in the above formulas
, we did not include any gluon sivers function contributions .
the gluon sivers function could dominate the asymmetry at central rapidities @xcite .
another important issue is the relevant sudakov suppressions for the asymmetries , which was found to be sizable for the di - jet correlation in the rhic energy range @xcite . in the following numerical studies , as an order of magnitude estimate for these asymmetries , we will neglect these effects , which however should be taken into account in future more detailed studies .
following the same procedure that we used for the sivers asymmetries in sidis and drell - yan processes , we can further simplify the polarized cross section by integrating out the transverse momentum , but keeping the azimuthal angle dependence explicit .
the differential cross section then can be written as @xmath170 where the unpolarized cross section has been given in eq .
( [ lojet ] ) , and the polarized ones read @xmath171 with the distribution @xmath172 defined as @xmath173 we note that the second term in the polarized cross section is power - suppressed by @xmath174 . this suppression is due to the fact that we have integrated over all intrinsic transverse momentum .
clearly , this term is beyond the approximations we have made , and we can not reliably predict it since there will be other sources of power - suppressed contributions , for example generated within the qiu - sterman mechanism @xcite . since it is anyway expected to be small , we will discard it in the following . employing the same set of sivers functions that we used for our predictions for the drell - yan process above
, we then find the results for @xmath175 shown in fig .
[ fig12 ] .
we show the asymmetries as functions of the rapidity of jet `` 1 '' , and of the jet transverse momentum @xmath176 .
one can see that the ssa for the di - jet correlation can become very large , in particular in the forward rapidity region .
asymmetries of this size should be relatively easily measurable in the future .
we also note that the asymmetry has opposite sign compared to that for drell - yan dimuon production discussed earlier .
the reason for this is that @xmath68-quark contributions dominate in drell - yan , thanks to their large electromagnetic charge , whereas for di - jets , @xmath85-quark contributions are not charge - suppressed and in fact dominate , keeping in mind that the analysis of the hermes data appears to favor a large @xmath85 quark sivers function .
the opposite signs of the sivers up and down quark distributions we found in eq .
( [ eq19 ] ) then explains the opposite signs of the spin asymmetries for drell - yan and di - jets .
we note that if the two jets are within the central rapidity region , our prediction for the asymmetry is much smaller . as mentioned above
, the gluon sivers function could dominate the asymmetry in this region @xcite .
we stress again that if factorization can be shown for the sivers ssa in di - jet production , it is likely that the structure of the resulting expression may differ from the one we use .
in particular , there will be calculable factors that represent the non - universality of the sivers functions related to the process dependence of the gauge links , leading effectively to modified partonic hard - scattering functions @xcite , at variance with our use of the standard unpolarized ones .
as a test , we have also used the modified partonic cross sections derived in @xcite .
we find relatively small changes in the results we obtain .
unfortunately , however , this is not really representative : the cross sections given in @xcite are only for the quark-(anti)quark scattering channels , whereas the dominant contribution in our calculation mostly comes from @xmath177 scattering ( with the gluon from the unpolarized proton ) .
it remains to be seen to what extent eventually our predictions will change , once the process - dependence for the sivers functions in di - jet correlations is completely understood .
we finally note that it would also be interesting in particular for measurements with the phenix detector to study correlations between hadrons in opposite jets .
such di - hadron correlations could serve as surrogates for the di - jet correlations we have discussed above . in this case there will , however , also be contributions from the collins mechanism . and transverse momentum @xmath176 . [ fig12]_,height=264 ]
-0.4 cm it is straightforward to extend the analysis of di - jet correlations discussed above to the case of jet - plus - photon correlations . we simply need to implement the cross sections for the appropriate born - level partonic scatterings @xmath179 and @xmath180 in eq .
( [ dijetsiv ] ) .
although events with a photon suffer from smaller rates than two - jet events , they would offer additional information on the sivers functions .
it is also likely that proofs of factorization are more easily obtained here , since the reactions @xmath181 and @xmath180 each have only a single color structure . figure [ fig13 ] shows results for the single - spin asymmetry for jet - photon correlations for the same kinematics as for the di - jet case in fig .
[ fig12 ] .
variables with the subscript `` 1 '' denote photon variables .
again , sizable asymmetries are seen , in particular at forward rapidities . the asymmetries are somewhat smaller than the ones we found for di - jets .
this is a result of cancellations between our sivers @xmath68 and @xmath85 functions , due to the larger weighting factor @xmath103 that the @xmath68-quark contributions now have for the prompt - photon case . and transverse momentum @xmath176 .
[ fig13]_,height=264 ] -0.4 cm
in this paper we have studied single - transverse spin asymmetries in semi - inclusive deep inelastic scattering and at hadron colliders .
we have analyzed the sivers and collins contributions to the spin asymmetry in sidis , and fitted simple parameterizations of the corresponding functions to recent data from hermes .
these fits work well and also turn out to be consistent with compass measurements of the asymmetries in dis off a deuteron target . for the sivers functions
, we found dominance of the down - quark distribution over the up - quark one .
the sivers-@xmath85 density in sidis turns out to be positive , while the @xmath68-quark distribution comes out negative . concerning the collins functions , we have found that current data do not yet pin down the relative size of `` favored '' and `` unfavored '' functions , which is also due to the fact that the transversity densities are not yet known .
we have also given theoretical arguments that the `` favored '' and `` unfavored '' collins functions could be of similar size , and of opposite sign .
we have then investigated sivers - type single - spin asymmetries at hadron colliders , focusing on the drell - yan process and on di - jet and jet - photon correlations , all in circumstances where there is a small measured transverse momentum , but the process is overall characterized by a large scale . using the sivers functions obtained from the analysis of the hermes data , we have made predictions for single - spin asymmetries for these processes .
we find relatively large asymmetries , in particular at forward rapidities of the observed final state .
such asymmetries should be measurable with dedicated efforts at rhic .
besides the additional valuable information they would give on the sivers functions and therefore on the structure of the nucleon , they would also provide a test of our theoretical understanding of `` naively time - reversal - odd '' phenomena in qcd .
the crucial issues in this are the factorization of the corresponding cross sections , and the universality of the sivers functions , on both of which further theoretical work is required .
we thank elke aschenauer , delia hasch , and gunar schnell for providing the hermes data , and for discussions regarding the data .
we are grateful to andrea bressan for sending us the compass data , and for discussions .
we thank leslie bland for stimulating discussions on jet - photon correlations .
we also thank harut avakian , daniel boer , xiangdong ji , naomi makins , and matthias grosse - perdekamp for valuable conversations .
we also thank umberto dalesio , rainer joosten , and aram kotzinian for correspondence and comments .
we are finally grateful to riken , brookhaven national laboratory and the u.s .
department of energy ( contract number de - ac02 - 98ch10886 ) for providing the facilities essential for the completion of his work .
* note added : * upon completion of this paper , we noticed the preprint @xcite where also the sivers functions were fitted to the new hermes and compass data , and predictions for ssas in the drell - yan process were made .
as far as we can see , our results are in qualitative agreement with those of @xcite , keeping in mind that their sign convention for the sivers function is opposite to ours and to that in the `` trento conventions '' .
see , for example : d. l. adams _ et al . _
[ e581 and e704 collaborations ] , phys .
b * 261 * , 201 ( 1991 ) ; d. l. adams _ et al .
_ [ fnal - e704 collaboration ] , phys .
b * 264 * , 462 ( 1991 ) ; k. krueger _
et al . _ ,
b * 459 * , 412 ( 1999 ) . for reviews , see : m. anselmino , a. efremov and e. leader , phys .
rept . * 261 * , 1 ( 1995 ) [ erratum - ibid . *
281 * , 399 ( 1997 ) ] ; v. barone , a. drago and p. g. ratcliffe , phys .
rept . * 359 * , 1 ( 2002 ) ; s. d. bass , arxiv : hep - ph/0411005 .
j. adams _ et al . _ [ star collaboration ] , phys . rev
. lett .
* 92 * , 171801 ( 2004 ) [ arxiv : hep - ex/0310058 ] .
s. s. adler [ phenix collaboration ] , arxiv : hep - ex/0507073 ; c. aidala [ phenix collaboration ] , arxiv : hep - ex/0410003 ; arxiv : hep - ex/0501054 . f. videbaek [ brahms collaboration ] , talk presented at the `` 13th international workshop on deep inelastic scattering ( dis 2005 ) '' , madison , wisconsin , april 27-may 1 , 2005 , to appear in the proceedings .
d. adams _ et al . _
[ spin muon collaboration ( smc ) ] , phys . lett .
b * 336 * , 125 ( 1994 ) ; a. bravar [ spin muon collaboration ] , nucl .
a * 666 * , 314 ( 2000 ) .
a. airapetian _ et al . _
[ hermes collaboration ] , phys .
lett . * 84 * , 4047 ( 2000 ) ; a. airapetian _ et al . _
[ hermes collaboration ] , phys .
d * 64 * , 097101 ( 2001 ) .
a. airapetian _ et al . _
[ hermes collaboration ] , phys .
lett . * 94 * , 012002 ( 2005 ) [ arxiv : hep - ex/0408013 ] .
m. diefenthaler [ hermes collaboration ] , talk presented at the `` 13th international workshop on deep inelastic scattering ( dis 2005 ) '' , madison , wisconsin , april 27-may 1 , 2005 , to appear in the proceedings , arxiv : hep - ex/0507013 .
h. avakian [ clas collaboration ] , talk presented at the rbrc workshop `` single - spin asymmetries '' , brookhaven national laboratory , upton , new york , june 1 - 3 , 2005 , to appear in the proceedings . v. y. alexakhin _ et al . _ [ compass collaboration ] , phys .
lett . * 94 * , 202002 ( 2005 ) [ arxiv : hep - ex/0503002 ] .
a. v. efremov and o. v. teryaev , sov .
j. nucl .
* 36 * , 140 ( 1982 ) [ yad .
36 * , 242 ( 1982 ) ] ; a. v. efremov and o. v. teryaev , phys .
b * 150 * , 383 ( 1985 ) .
j. qiu and g. sterman , phys .
lett . * 67 * , 2264 ( 1991 ) ; nucl .
b * 378 * , 52 ( 1992 ) ; phys .
d * 59 * , 014004 ( 1999 ) .
ralston and d. e. soper , nucl .
b * 152 * , 109 ( 1979 ) .
j. c. collins and d. e. soper , nucl .
b * 193 * , 381 ( 1981 ) [ erratum - ibid .
b * 213 * , 545 ( 1983 ) ] ; nucl .
b * 197 * , 446 ( 1982 ) .
j. c. collins and d. e. soper , nucl .
b * 194 * , 445 ( 1982 ) .
j. c. collins , d. e. soper and g. sterman , nucl .
b * 250 * , 199 ( 1985 ) .
d. w. sivers , phys .
d * 41 * , 83 ( 1990 ) ; phys .
d * 43 * , 261 ( 1991 ) .
j. c. collins , nucl .
b * 396 * , 161 ( 1993 ) [ arxiv : hep - ph/9208213 ] .
a. kotzinian , nucl .
b * 441 * , 234 ( 1995 ) [ arxiv : hep - ph/9412283 ] .
m. anselmino , m. boglione and f. murgia , phys .
b * 362 * , 164 ( 1995 ) [ arxiv : hep - ph/9503290 ] ; m. anselmino and f. murgia , phys . lett .
b * 442 * , 470 ( 1998 ) [ arxiv : hep - ph/9808426 ] ; m. anselmino , m. boglione , u. dalesio , e. leader and f. murgia , phys .
d * 71 * , 014002 ( 2005 ) .
p. j. mulders and r. d. tangerman , nucl .
b * 461 * , 197 ( 1996 ) [ erratum - ibid .
b * 484 * , 538 ( 1997 ) ] [ arxiv : hep - ph/9510301 ] ; d. boer and p. j. mulders , phys . rev .
d * 57 * , 5780 ( 1998 ) [ arxiv : hep - ph/9711485 ] .
s. j. brodsky , d. s. hwang and i. schmidt , phys .
b * 530 * , 99 ( 2002 ) ; nucl .
b * 642 * , 344 ( 2002 ) .
j. c. collins , phys .
b * 536 * , 43 ( 2002 ) .
x. ji and f. yuan , phys .
b * 543 * , 66 ( 2002 ) ; a. v. belitsky , x. ji and f. yuan , nucl .
b * 656 * , 165 ( 2003 ) .
d. boer , p. j. mulders and f. pijlman , nucl .
b * 667 * , 201 ( 2003 ) .
x. ji , j. p. ma and f. yuan , nucl .
b * 652 * , 383 ( 2003 ) [ arxiv : hep - ph/0210430 ] .
m. burkardt , phys .
d * 66 * , 114005 ( 2002 ) [ arxiv : hep - ph/0209179 ] ; nucl .
phys . a * 735 * , 185 ( 2004 ) [ arxiv : hep - ph/0302144 ] ; phys . rev .
d * 69 * , 074032 ( 2004 ) [ arxiv : hep - ph/0309072 ] ; arxiv : hep - ph/0505189 .
x. ji , j. p. ma and f. yuan , phys .
d * 71 * , 034005 ( 2005 ) [ arxiv : hep - ph/0404183 ] ; x. ji , j. p. ma and f. yuan , phys .
b * 597 * , 299 ( 2004 ) [ arxiv : hep - ph/0405085 ] ; jhep * 0507 * , 020 ( 2005 ) [ arxiv : hep - ph/0503015 ] .
j. c. collins and a. metz , phys .
* 93 * , 252001 ( 2004 ) [ arxiv : hep - ph/0408249 ] .
d. boer and w. vogelsang , phys .
d * 69 * , 094025 ( 2004 ) [ arxiv : hep - ph/0312320 ] .
a. bacchetta , c. j. bomhof , p. j. mulders and f. pijlman , arxiv : hep - ph/0505268 . m. glck , e. reya and a. vogt , z. phys .
c * 67 * , 433 ( 1995 ) .
a. v. efremov , k. goeke , s. menzel , a. metz and p. schweitzer , phys .
b * 612 * , 233 ( 2005 ) [ arxiv : hep - ph/0412353 ] .
m. anselmino , m. boglione , u. dalesio , a. kotzinian , f. murgia and a. prokudin , phys .
d * 71 * , 074006 ( 2005 ) [ arxiv : hep - ph/0501196 ] .
a. v. efremov , k. goeke and p. schweitzer , arxiv : hep - ph/0412420 ; and talk given by p. schweitzer at the `` sir05 workshop '' , jlab , may 15 - 18 , newport news , viginia .
d. boer , s. j. brodsky and d. s. hwang , phys .
d * 67 * , 054003 ( 2003 ) [ arxiv : hep - ph/0211110 ] ; l. p. gamberg , g. r. goldstein and k. a. oganessyan , phys .
d * 67 * , 071504 ( 2003 ) [ arxiv : hep - ph/0301018 ] ; f. yuan , phys .
b * 575 * , 45 ( 2003 ) [ arxiv : hep - ph/0308157 ] ; a. bacchetta , a. schfer and j. j. yang , phys .
b * 578 * , 109 ( 2004 ) [ arxiv : hep - ph/0309246 ] ; z. lu and b. q. ma , nucl .
a * 741 * , 200 ( 2004 ) [ arxiv : hep - ph/0406171 ] .
a. bacchetta , r. kundu , a. metz and p. j. mulders , phys .
b * 506 * , 155 ( 2001 ) [ arxiv : hep - ph/0102278 ] ; phys .
d * 65 * , 094021 ( 2002 ) [ arxiv : hep - ph/0201091 ] ; l. p. gamberg , g. r. goldstein and k. a. oganessyan , phys .
d * 68 * , 051501 ( 2003 ) [ arxiv : hep - ph/0307139 ] ; a. bacchetta , a. metz and j. j. yang , phys .
b * 574 * , 225 ( 2003 ) [ arxiv : hep - ph/0307282 ] ; d. amrath , a. bacchetta and a. metz , phys .
d * 71 * , 114018 ( 2005 ) [ arxiv : hep - ph/0504124 ] .
a. schfer and o. v. teryaev , phys .
d * 61 * , 077903 ( 2000 ) [ arxiv : hep - ph/9908412 ] .
x. artru , j. czyzewski and h. yabuki , z. phys .
c * 73 * , 527 ( 1997 ) [ arxiv : hep - ph/9508239 ] ; see also , n. makins , for example , talks given at the rbrc workshop `` single - spin asymmetries '' , june 1 - 3 , 2005 , and the rhic & ags users meeting , june 20 - 24 , 2005 , brookhaven national laboratory , upton , new york .
see : cern program library , http://wwwasd.web.cern.ch/wwwasd/cernlib/ a. metz , phys .
b * 549 * , 139 ( 2002 ) . c. j. bomhof , p. j. mulders and f. pijlman , phys .
b * 596 * , 277 ( 2004 ) [ arxiv : hep - ph/0406099 ] .
gsi - pax collab .
, p. lenisa and f. rathmann ( spokespersons ) _ et al .
_ , technical proposal , arxiv : hep - ex/0505054 ; arxiv : hep - ex/0412078 .
j. f. owens , rev .
phys . * 59 * , 465 ( 1987 ) .
h. contopanagos , e. laenen and g. sterman , nucl .
b * 484 * , 303 ( 1997 ) [ arxiv : hep - ph/9604313 ] ; n. kidonakis and g. sterman , nucl .
b * 505 * , 321 ( 1997 ) [ arxiv : hep - ph/9705234 ] ; n. kidonakis , g. oderda and g. sterman , nucl .
b * 525 * , 299 ( 1998 ) [ arxiv : hep - ph/9801268 ] ; nucl . phys .
b * 531 * , 365 ( 1998 ) [ arxiv : hep - ph/9803241 ] ; r. bonciani , s. catani , m. l. mangano and p. nason , phys .
b * 575 * , 268 ( 2003 ) [ arxiv : hep - ph/0307035 ] .
|
we study single - spin asymmetries in semi - inclusive deep inelastic scattering with transversely polarized target .
based on the qcd factorization approach , we consider sivers and collins contributions to the asymmetries .
we fit simple parameterizations for the sivers and collins functions to the recent hermes data , and compare to results from compass . using the fitted parameterizations for the sivers functions , we predict the single transverse spin asymmetries for various processes in @xmath0 collisions at rhic , including the drell - yan process and angular correlations in di - jet and jet - plus - photon production .
these asymmetries are found to be sizable at forward rapidities .
|
let @xmath12 be the unit sphere in @xmath1 , where @xmath13 denotes the euclidean norm . given a compact set @xmath14 ,
consider the class @xmath15 of unit positive borel measures supported on @xmath16 .
for @xmath17 the _ riesz @xmath5-potential _ and _ riesz @xmath5-energy _ of a measure @xmath18 are given , respectively , by @xmath19 where @xmath20 is the so - called _
riesz kernel_. the _ @xmath5-capacity _ of @xmath16 is then defined as @xmath21 for @xmath17 , where @xmath22 is the _ @xmath5-energy _ of the set @xmath16 .
a property is said to hold _ quasi - everywhere _
( q.e . ) , if the exceptional set has @xmath5-capacity zero .
when @xmath23 , there exists a unique minimizer @xmath24 , called the _ @xmath5-equilibrium measure on @xmath16 _ , such that @xmath25 . for more details
see ( * ? ? ?
* chapter ii ) . whenever @xmath26 ( we shall use @xmath27 ) , which occurs , for example , when @xmath28 and @xmath29 , we replace the riesz kernel @xmath30 by the _ logarithmic kernel _ @xmath31 ( in this case we define @xmath32 . ) we shall refer to a lower semi - continuous function @xmath33 $ ] such that @xmath34 on a set of positive lebesgue surface measure , as an _ external field_.
we note that the lower semi - continuity implies the existence of a finite @xmath35 such that @xmath36 for all @xmath37 .
the _ weighted energy associated with @xmath38 _ is then given by @xmath39 ( the terminology `` weighted energy '' is used here to indicate the presence of an external field , and should not be confused with `` weighted energy functionals '' , where the riesz @xmath5-kernel is multiplied by a weight function @xmath40 .
we leave the study of the external field problem for such generalized kernels for a future investigation . )
[ def : external.field.problem ] the riesz external field problem on the unit sphere @xmath0 for the external field @xmath2 is concerned with minimizing the weighted energy among all borel probability measures @xmath41 supported on @xmath0 . a measure @xmath42 with @xmath43
is called an _ @xmath5-extremal ( or positive equilibrium ) measure on @xmath0 associated with @xmath2_. if we consider only measures supported on some compact subset @xmath44 with positive @xmath5-capacity , then the minimizing measure is referred to as the _ @xmath5-extremal measure on @xmath16 associated with @xmath2 _ and denoted by @xmath45 . in the particular case when @xmath46 , the measure @xmath47 on @xmath0 is just the normalized unit surface area measure on the sphere for which we use the symbol @xmath48 .
we shall also consider the discrete analogue of the above external field problem which is defined as follows .
[ def : discrete.external.field.problem ] let @xmath49 or @xmath27 .
for a set of @xmath10 points @xmath50 the _ discrete weighted energy associated with @xmath2 _ is given by @xmath51.\ ] ] then the discrete external field problem on the sphere @xmath0 concerns the minimization @xmath52 where @xmath53 denotes the cardinality of the set @xmath54 .
a solution of the discretized minimization problem is called an _
@xmath10-point @xmath9-fekete set_. the existence of @xmath9-fekete sets is an easy consequence of the lower semi - continuity of the energy functional and the compactness of the unit sphere .
further , we remark that a standard argument establishes the following monotonicity property @xmath55 we remark that the discrete problem has application to image processing , namely the half - toning of images based on electrostatic repulsion of printed dots in the presence of an image - driven external field ; cf .
schmaltz et al .
@xcite and grf ( * ? ? ?
* section 6.5.2 ) .
the outline of the paper is as follows . in section
[ sec : characterization ] we provide frostman - type characterization theorems for the solution to the external field minimal energy problem on the sphere .
this is facilitated by a new restricted maximum principle on the sphere which holds for the range @xmath56 ( see theorem [ thm : restr.max.principle ] ) .
we also introduce the signed equilibrium measure and discuss its relation to the positive equilibrium measure . in section [ sec : separation ] we establish that for a large class of external fields @xmath2 , the sequences of @xmath10-point @xmath9-fekete sets are well - separated ; that is , have separation distance of order @xmath57 ( theorems [ thm : main3 ] and [ thm : main4 ] ) . in section [ sec : neg.external.field ] , for an external field due to a negative point charge , we provide a detailed analysis and give explicit representations of the signed equilibrium ( theorem [ thm : signeq ] ) and the @xmath5-extremal measure on @xmath0 ( theorem [ thm : s.equilibrium.measure ] ) .
this extends results in @xcite .
in section [ sec : examples ] we rigorously characterize the @xmath58-point @xmath9-fekete set for a general class of convex external fields and provide numerical results for the four point problem with riesz external fields ( figures [ fig3 ] and [ fig2 ] illustrate the analysis ) .
the proofs of our results are provided in section [ sec : proofs ] .
in @xcite the second and the third authors formulated the following frostman - type proposition , which deals with the existence and uniqueness of the measure @xmath47 , as well as a criterion that characterizes @xmath47 in terms of its potential .
the proof of this proposition follows closely the proof of ( * ? ? ?
* theorem i.1.3 ) .
it could also be derived as a particular case from the more general results in @xcite ( see especially theorems 1 and 2 , and proposition 1 of that paper ) .
[ prop:1 ] let @xmath59 . for the minimal energy problem on @xmath0 with external field
@xmath2 the following properties hold : * @xmath60 is finite .
* there exists a unique @xmath5-extremal measure @xmath61 associated with @xmath2 .
moreover , the support @xmath62 of this measure is contained in the compact set @xmath63 for some @xmath64 . *
the measure @xmath47 satisfies the variational inequalities @xmath65 where @xmath66 * inequalities and completely characterize the @xmath5-extremal measure @xmath67 in the sense that if @xmath68 is a measure with finite @xmath5-energy such that for some constant @xmath69 we have @xmath70 then @xmath71 and @xmath72 .
observe that , if the external field @xmath2 is continuous on @xmath0 , then the inequality in holds everywhere on @xmath0 .
proposition [ prop:1 ] remains true if @xmath0 is replaced with any compact subset @xmath73 with @xmath74 .
notationally , the dependence on @xmath75 will be indicated by a subscript @xmath75 ( e.g. , @xmath76 , @xmath77 , etc . ) . in the case
when @xmath78 , ( * ? ? ?
* theorem 1.3 ) analyzes further the characterization property from proposition [ prop:1](d ) by studying the supremum and the _ essential infimum _ of the weighted potential @xmath79 .
our first theorem extends this analysis to the larger range @xmath80 . to state the theorem
we introduce the notation @xmath81 to denote the _ essential infimum _ of @xmath82 with respect to a set @xmath83 ; that is , @xmath84 in other words , the infimum is taken quasi - everywhere .
[ thm : main ] let @xmath80 , @xmath2 be an external field on @xmath0 , and @xmath85 be defined as in .
for any measure @xmath86 we have @xmath87 & \leq f_{q , s } \label{eq : essinf } \intertext{and } \sup_{{\mathbf{x } } \in\operatorname{supp}(\lambda ) } \left [ u_s^\lambda({\mathbf{x } } ) + q({\mathbf{x } } ) \right ] & \geq f_{q , s}. \label{eq : sup}\end{aligned}\ ] ] if equality holds in both inequalities , then @xmath88 . for the restricted range @xmath89 , the proof of this theorem as given in @xcite utilizes the principle of domination for riesz potentials , which generally is stated for the parameter range @xmath90 and measures supported on any subsets of @xmath1 .
( a restricted version of the principle of domination for @xmath91 was established in ( * ? ? ?
* lemma 5.1 ) . ) via a different approach that utilizes the following _ restricted maximum principle on the sphere _ , we are able to prove the result for @xmath5 in the extended range @xmath56 ; see section [ sec : proofs ] .
[ thm : restr.max.principle ] let @xmath80 .
suppose @xmath41 is a positive measure with @xmath92 such that for some @xmath93 , the relation @xmath94 holds @xmath41-almost everywhere on @xmath0 .
then @xmath94 holds everywhere on @xmath0 .
an essential part of the analysis of external field problems is the determination the @xmath5-extremal ( equilibrium ) measure on @xmath0 associated with the external field @xmath2 and , in particular , its support . in principle , if the latter is known , the measure @xmath47 can be recovered by solving an integral equation for the weighted @xmath5-potential of @xmath47 arising from the variational inequalities and . a substantially easier problem is to find a ( signed ) measure that has constant weighted @xmath5-potential everywhere on @xmath0 .
the solution of this problem turns out to be useful in solving the harder problem .
this motivates the study of the signed equilibrium measure associated with an external field which is defined as follows .
[ def : signed.equilibrium ] given a compact subset @xmath95 ( @xmath96 ) and an external field @xmath2 , we call a signed measure @xmath97 supported on @xmath75 and of total charge @xmath98 _ a signed @xmath5-equilibrium on @xmath75 associated with @xmath2 _ if its weighted riesz @xmath5-potential is constant on @xmath75 ; that is , @xmath99 we note that if a signed equilibrium exists , then it is unique ( see ( * ? ? ? * lemma 23 ) ) .
a remarkable connection exists to the riesz analog of the _ mhaskar - saff @xmath100-functional _ from classical logarithmic potential theory in the plane ( see @xcite and ( * ? ? ?
* chapter iv , p. 194 ) ) .
@xmath101-functional _ of a compact subset @xmath44 of positive @xmath5-capacity is defined as @xmath102 where @xmath103 is the @xmath5-energy of @xmath16 and @xmath104 is the @xmath5-equilibrium measure ( without external field ) on @xmath16 .
let @xmath56 with @xmath49 .
if the signed equilibrium on a compact set @xmath73 associated with @xmath2 exists , then integration of with respect to @xmath105 shows that @xmath106 the essential property of the @xmath101-functional is the following ( cf .
* theorem 9 ) ) .
[ prop : f.s.functional ] let @xmath56 with @xmath49 and @xmath2 be an external field on a compact subset @xmath73 with @xmath74 . then the @xmath101-functional is minimized for the support of the @xmath5-extremal measure @xmath76 on @xmath75 associated with @xmath2 ; that is , for every compact subset @xmath107 with @xmath108 , @xmath109 given a compact subset @xmath73 , the _ extended support @xmath110 of @xmath76 _ is defined by @xmath111 the following theorem , which is the riesz analog of ( * ? ? ?
* theorem 2.6 ) and ( * ? ? ?
* lemma 3 ) , establishes a relation between the extended support @xmath112 of @xmath47 ( by this set contains the support of @xmath47 ) and the support of the positive part @xmath113 of the jordan decomposition @xmath114 of the signed equilibrium @xmath115 on @xmath0 associated with @xmath2 .
[ thm : signsupp.2 ] let @xmath56 and suppose that @xmath2 is an external field such that a signed @xmath5-equilibrium @xmath116 on a spherical cap @xmath117 exists .
then @xmath118 furthermore , if @xmath119 , then @xmath120 .
the theorem remains true if the @xmath5-extremal measure on a compact subset @xmath73 with @xmath74 and signed @xmath5-equilibria @xmath121 on compact subsets @xmath44 with @xmath122 are considered .
the characterization results for the shape of the support of the @xmath5-extremal measure on @xmath0 associated with a rotational symmetric external field given in ( * ? ?
? * theorem 10 ) immediately carry over to the external fields with extended range .
[ prop : connthm ] let @xmath123 with @xmath17 and the external field @xmath124 $ ] be rotationally invariant about the polar axis ; that is , @xmath125 , where @xmath126 is the altitude of @xmath127 , @xmath128 .
suppose that @xmath82 is a convex function on @xmath129 $ ] .
then the support of the @xmath5-extremal measure @xmath67 on @xmath0 is a spherical zone ; namely , there are numbers @xmath130 such that @xmath131 moreover , if additionally @xmath82 is increasing , then @xmath132 and the support of @xmath67 is a spherical cap centered at the south pole .
next we focus on the discretized version of the riesz external field problem given in definition [ def : external.field.problem ] . recall that the normalized counting measure associated with an @xmath10-point set @xmath133 is defined as @xmath134 where @xmath135 is the dirac - delta measure with unit mass at @xmath136 .
the continuous and discrete external field minimization problems are related in the following way .
let @xmath137 or @xmath27 .
then @xmath138 furthermore , if @xmath139 is any sequence of @xmath10-point @xmath9-fekete sets on @xmath0 ( see definition [ def : discrete.external.field.problem ] ) , then the sequence of the normalized counting measures @xmath140 associated with @xmath141 converges in the weak - star sense to the @xmath5-extremal measure @xmath47 .
the proof follows from a standard argument and utilizes the uniqueness result stated in proposition [ prop:1](b ) .
we are interested in determining sets that contain all the @xmath9-fekete sets . for this purpose
it is useful to investigate the _
weighted @xmath5-potential _ of the normalized counting measure @xmath142 which is defined as @xmath143 as an application of theorem [ thm : main ] we deduce the following result .
[ thm : main2 ] let @xmath144 .
let @xmath145 be a set of @xmath10 distinct points , and suppose that , for some constant @xmath146 , the associated weighted potential satisfies the inequality @xmath147 then ( cf . ) @xmath148 furthermore , @xmath149 we point out that this is an extension of ( * ? ? ?
* theorem 1.7 ) , which , as with theorem [ thm : main ] above , was originally established for @xmath78 . as in (
* corollary 1.9 ) , theorems [ thm : main ] and [ thm : main2 ] yield the following .
[ cor ] for @xmath80 , every @xmath9-fekete set is contained in the extended support @xmath112 .
we note that for most of the above theorems , @xmath28 marks the lower end of the stated range of the riesz parameter @xmath5 .
it turns out that the case @xmath28 is distinctive because new phenomena arise in the solution of the signed equilibrium problem , see section [ sec : neg.external.field ] .
moreover , for @xmath5 in the interval @xmath150 , the riesz-@xmath5 kernel becomes strictly superharmonic when considered in the stereographic projection space of @xmath0 ; consequently maximum principles and domination principles do not apply .
good separation of points is generally associated with the _ stability _ of an approximation or interpolation method ( e.g. , by splines or radial basis functions ( rbf ) ) ; cf .
, e.g. , @xcite . in this section
we shall apply results from section [ sec : characterization ] ( especially theorem [ thm : signsupp.2 ] and corollary [ cor ] ) to obtain explicit point separation estimates for sequences of @xmath10-point @xmath9-fekete sets ( cf . definition [ def : discrete.external.field.problem ] ) associated with a large class of external fields @xmath2 and establish that such sequences are `` well - separated '' in the following sense .
let @xmath151 denote the _ minimum distance _ among the points in @xmath152 .
then a sequence @xmath153 , @xmath145 for all @xmath10 , is called _ well - separated _ if @xmath154 is of order @xmath57 as @xmath155 .
( it suffices to show the existence of a constant @xmath69 such that @xmath156 since @xmath154 can not exceed the best - packing distance which is of order @xmath57 ; cf .
@xcite . ) in the potential - theoretical and field - free setting ( @xmath46 ) it has been known since dahlberg @xcite that fekete point sets ( harmonic case @xmath157 ) on a sufficiently smooth closed bounded @xmath158-dimensional surface in @xmath1 that separates @xmath1 into two parts will form a well - separated sequence ( but no explicit constant for the lower bound of @xmath154 has been given ) .
gtz @xcite studied the discrete external field problem on surfaces in @xmath159 where the energy functional is defined in terms of the green function for a domain @xmath160 .
his separation result generalizes dahlberg s result .
well - separation of minimal logarithmic energy configurations on @xmath161 in the field - free setting was first established by rakhmanov et al .
@xcite , and with an improved constant by dragnev @xcite . for minimal riesz @xmath5-energy configurations on @xmath0 in the field - free case , well - separation was established by kuijlaars et al .
@xcite for @xmath162 and by dragnev and saff @xcite for @xmath163 .
damelin and maymeskul @xcite give a separation result of order @xmath164 , @xmath165 , which is of sharp order in the boundary case @xmath28 .
it is expected but still unproven that minimal logarithmic and riesz @xmath5-energy ( @xmath166 ) configurations on @xmath0 , @xmath167 , are well - separated .
the references @xcite , @xcite and @xcite also provide an explicit constant in the lower estimate .
it should be noted that @xcite uses external fields to derive the desired separation estimates in the field - free setting . in the hyper - singular case @xmath11 , kuijlaars and saff @xcite
establish well - separation of minimal riesz @xmath5-energy configurations on @xmath0 .
we now present a generalization of ( * ? ? ?
* theorem 1.5 ) to the case when an external field is present and given by a potential .
[ thm : main3 ] let @xmath80 and @xmath168 for some signed measure @xmath169 with @xmath48-a.e .
finite riesz @xmath5-potential on @xmath0 .
assume the support of the negative part @xmath170 in the jordan decomposition @xmath171 satisfies that @xmath172 for some @xmath173 and that @xmath174 then any sequence @xmath175 of @xmath9-fekete sets on @xmath0 is well - separated ; more precisely , @xmath176 where @xmath177 it is understood that for @xmath29 and @xmath27 we replace @xmath178 by @xmath179 and @xmath5 by @xmath180 .
since the _ riesz @xmath5-energy of @xmath0 _ appearing in the separation constant in is given by the formula @xmath181 we have in the harmonic case @xmath182 that @xmath183^{-1/d}\ ] ] and in the limiting case @xmath28 ( and @xmath184 ) @xmath185^{-1/d},\ ] ] wheras for @xmath27 and @xmath29 @xmath186^{-1/2}.\ ] ] note that whenever the support of @xmath170 lies outside of @xmath0 , then both conditions and are satisfied by taking @xmath187 sufficiently close to @xmath179 . also observe that as @xmath4 approaches @xmath179 , the constant @xmath188 approaches @xmath180 . in case of @xmath189 and
@xmath190 the above theorem [ thm : main3 ] yields a known result for the well - separation of @xmath10-point minimal riesz @xmath191-energy configurations on @xmath0 ( @xcite but without explicit constants ) . also with @xmath189 , @xmath27 and @xmath29
we recover the same separation result as obtained in @xcite . here
we prove them separately ( with explicit constants in the former case ) , since they will be used to establish the separation bounds when an externalf field ( @xmath192 ) is given .
[ prop : well - separation.min.riesz.(d-2).energy ] for @xmath193 and @xmath194 we have @xmath195 for any @xmath196-point riesz @xmath191-energy we mean logarithmic energy . ] minimizing configuration @xmath197 on @xmath0 , where @xmath198 observe that @xmath199 when @xmath184 .
the first three values of @xmath200 are @xmath201 , @xmath202 , and @xmath203 .
curiously , @xmath204 is the ratio of the volume of the unit ball in @xmath159 divided by the surface area of the unit sphere in @xmath1 .
( this constant also appears as the coefficient of the leading term in the asymptotic expansion of the @xmath10-point minimal riesz @xmath158-energy as @xmath155 ( cf .
@xcite ) . ) finally , we present a well - separation result for sequences of @xmath9-fekete sets in the hyper - singular case @xmath205 . in this case
the ( strongly repellent ) short - range interactions between points on the sphere ensure well - separation of minimizing configurations for any continuous external field on @xmath0 .
in fact , it is enough that @xmath2 be integrable on some small subset of @xmath0 of positive surface area measure .
[ thm : main4 ] let @xmath11 .
suppose there is a subset @xmath206 such that @xmath207 and the fixed external field @xmath2 is integrable over @xmath208 with respect to @xmath48 .
then there is a constant @xmath69 independent of @xmath10 such that @xmath209 for any @xmath10-point @xmath9-fekete set on @xmath0 . in case of @xmath210
varies with @xmath10 , there may be no single fixed subset @xmath208 satisfying the hypotheses in theorem [ thm : main4 ] .
however , one can still deduce well - separation by requiring the following : there is a sequence @xmath211 of subsets of @xmath0 such that for some @xmath212 , @xmath213 for all @xmath10 , and for some @xmath214 independent of @xmath10 , @xmath215 where @xmath216 denotes the minimum of @xmath217 over @xmath0 .
these conditions are derived from the main inequality and the estimate . for large classes of external fields @xmath2 ( e.g. , continuous external fields ) , inequality and the estimate
can be made explicit which , in turn , yields an explicit constant in the separation estimate , as the following example illustrates .
let @xmath11 .
consider the external field @xmath218 , @xmath219 , @xmath220 , due to a point source above the north pole @xmath221 .
clearly , @xmath2 is continuous and thus integrable on @xmath222 .
thus theorem [ thm : main4 ] assures well - separation of @xmath10-point @xmath9-fekete sets on @xmath0 .
an explicit lower bound can be easily derived from and .
we find @xmath223 and , consequently , @xmath224 where @xmath225 \bigg\}^{-1/s}.\ ] ] note that @xmath226 as @xmath155 .
the constants @xmath227 and @xmath228 are given in and , respectively , and the representation of @xmath229 appears in .
in the following we consider external fields @xmath2 that are generated by negative sources .
the required lower semi - continuity of @xmath230 $ ] implies that no negative singularities can be on the sphere but it may support a negative `` continuous '' charge distribution ( with no discrete part relative to @xmath0 ) .
we give a detailed analysis for the riesz external field @xmath231 where @xmath232 , that is due to a negative point source at @xmath233 below the south pole and which also provides the basis for more general axis - supported fields defined by superposition of point source fields .
our analysis thus extends and complements results in @xcite where positive axis - supported external fields were considered .
intuitively , a negative point source under the south pole will `` pull '' charge towards the south pole and if sufficiently strong will cause a negatively charged spherical cap around the north pole to appear on a grounded sphere .
grounding of the sphere imposes constant weighted potential everywhere on @xmath0 .
this naturally leads to the signed equilibrium problem on the whole sphere or on its parts , say , the spherical cap @xmath234 centered at the south pole . on a positively charged isolated sphere
a sufficiently strong negative field will produce a spherical cap around the north pole that is free of charge .
we are specifically interested in the charge distribution on the remaining part @xmath235 , that is the @xmath5-extremal measure on @xmath0 associated with the external field @xmath236 and its support @xmath235 .
we will use the methods and results of @xcite .
an essential concept is the _ @xmath5-balayage _ of a measure . recall that given a measure @xmath237 and a compact set @xmath75 ( of the sphere @xmath0 ) , the balayage measure @xmath238 preserves the riesz @xmath5-potential of @xmath237 onto the set @xmath75 and diminishes it elsewhere ( on the sphere @xmath0 )
let @xmath239 denote the signed @xmath5-equilibrium on @xmath240 associated with the external field @xmath236
. then it can be expressed as @xmath241 \nu_t - q \epsilon_t,\ ] ] where @xmath242 are the @xmath5-balayage measures onto @xmath240 of the positive unit point charge at @xmath233 and the uniform measure @xmath48 on @xmath0 . the function @xmath243 , defined by @xmath244 in terms of the @xmath5-energy of @xmath0 , given in and norms @xmath245 and @xmath246 , plays an important role in what follows .
indeed , @xmath247 and using that @xmath248 and @xmath249 on @xmath240 by , at every @xmath250 @xmath251 by definition [ def : signed.equilibrium ] , @xmath252 and relates @xmath243 to the @xmath101-functional by means of @xmath253 , whereas proposition [ prop : f.s.functional ] implies that the latter is minimized by the support of the @xmath5-extremal measure on @xmath0 associated with the external field which turns out to be a spherical cap @xmath235 .
we will see that the unique minimum of @xmath243 in the interval @xmath129 $ ] will provide this critical parameter @xmath254 ( see theorem [ thm : s.equilibrium.measure ] ) . moreover , the remark following theorem [ thm : signeq ] provides the necessary and sufficient conditions ( involving @xmath243 and therefore @xmath255 ) under which the signed @xmath5-equilibrium measure @xmath239 on @xmath240 turns into the @xmath5-extremal measure on @xmath0 associated with the external field . throughout ,
@xmath256 and @xmath257 denote the gauss hypergeometric function and its regularized form a negative integer . ] with series expansions @xmath258 where @xmath259 and @xmath260 for @xmath261 is the pochhammer symbol .
we also recall that the incomplete beta function and the beta function are defined as @xmath262 whereas the regularized incomplete beta function is given by @xmath263 first , we give the representation of the signed equilibrium on the whole sphere @xmath0 , which is well - known from elementary physics ( cf .
@xcite ) in the classical coulomb case , and provide a necessary and sufficient condition when it also is the @xmath5-extremal measure on @xmath0 .
[ prop : signeq ] let @xmath137 and @xmath264 .
the signed @xmath5-equilibrium @xmath265 on @xmath0 associated with the riesz external field , where in fact @xmath266 , is given by @xmath267 furthermore , @xmath268 where @xmath269 has the following representation : @xmath270 moreover , if @xmath271 , then @xmath272 if and only if @xmath273 in such a case @xmath274 .
indeed , it can be readily verified that @xmath275 given by is the signed equilibrium measure on the whole sphere on observing that @xmath276 with @xmath277 by using a suitably defined kelvin transformation ; for details see ( * ? ? ?
* proof of theorem 2 ) which can be easily extended to hold for the external field .
the unique signed equilibrium @xmath275 on @xmath0 associated with @xmath236 coincides with the unique positive @xmath5-equilibrium measure on @xmath0 if and only if @xmath275 is a measure .
this is a consequence of the uniqueness of either measure and variational inequalities ( proposition [ prop:1](d ) ) .
as the strictly decreasing density function @xmath278 assumes its minimum value at the north pole @xmath221 , @xmath279 we obtain the necessary and sufficient criterion in . in light of it
is natural to ask if there is a critical distance @xmath280 such that the support of @xmath281 is all of @xmath0 for @xmath282 but is a proper subset of @xmath0 for @xmath283 . for external fields with positive charge @xmath284
such a critical distance always exists ; cf .
* remark 4 and example 5 ) and @xcite .
however , for negative charge @xmath284 such a critical distance does not always exist .
we discuss this phenomenon in more detail in @xcite .
how are the riesz-@xmath5 external fields , generated by a positive point charge @xmath285 above the north pole ( @xmath286 ) and a negative point charge @xmath287 below the south pole ( @xmath288 ) , related ?
let @xmath289 and @xmath278 be the density functions of the respective signed equilibrium on @xmath0 .
then it can be readily verified that for same center distances @xmath290 and total charge @xmath291 the densities complement each other ; that is , @xmath292 on the other hand , when it is assumed that both @xmath289 and @xmath278 attain the same value at the north pole , then again by ( * ? ? ?
* theorem 2 ) and proposition [ prop : signeq ] , @xmath293 here , in the case @xmath290 ( i.e. @xmath294 ) , equation is equivalent with @xmath295 and in the case @xmath296 ( i.e. @xmath297 ) , equation becomes @xmath298 observe that the last relation between @xmath299 and @xmath300 does not depend on the charge @xmath284 .
figure [ fig : signed.equilibria.sphere ] illustrates theses cases .
comparison of density functions of signed equilibria on @xmath161 associated with riesz external fields generated by a positive point charge above the north pole ( blue curves in colored version , i.e. smaller value at @xmath301 ) and a negative point charge below the south pole ( red curves in colored version , i.e. larger value at @xmath301 ) having the same density function value at @xmath179 ( north pole ) . left column : fixed center distance @xmath302 ( @xmath303 golden ratio ) , @xmath304 and @xmath287 from .
right column : total charge @xmath296 with @xmath305 , @xmath306 and @xmath299 from . from top to bottom : @xmath307 , @xmath179 , @xmath308 .
( no solution for @xmath309 in right column.),title="fig : " ] + comparison of density functions of signed equilibria on @xmath161 associated with riesz external fields generated by a positive point charge above the north pole ( blue curves in colored version , i.e. smaller value at @xmath301 ) and a negative point charge below the south pole ( red curves in colored version , i.e. larger value at @xmath301 ) having the same density function value at @xmath179 ( north pole ) . left column : fixed center distance @xmath302 ( @xmath303 golden ratio ) , @xmath304 and @xmath287 from .
right column : total charge @xmath296 with @xmath305 , @xmath306 and @xmath299 from . from top to bottom : @xmath307 , @xmath179 , @xmath308 .
( no solution for @xmath309 in right column.),title="fig : " ] comparison of density functions of signed equilibria on @xmath161 associated with riesz external fields generated by a positive point charge above the north pole ( blue curves in colored version , i.e. smaller value at @xmath301 ) and a negative point charge below the south pole ( red curves in colored version , i.e. larger value at @xmath301 ) having the same density function value at @xmath179 ( north pole ) . left column : fixed center distance @xmath302 ( @xmath303 golden ratio ) , @xmath304 and @xmath287 from .
right column : total charge @xmath296 with @xmath305 , @xmath306 and @xmath299 from . from top to bottom : @xmath307 , @xmath179 , @xmath308 .
( no solution for @xmath309 in right column.),title="fig : " ] + comparison of density functions of signed equilibria on @xmath161 associated with riesz external fields generated by a positive point charge above the north pole ( blue curves in colored version , i.e. smaller value at @xmath301 ) and a negative point charge below the south pole ( red curves in colored version , i.e. larger value at @xmath301 ) having the same density function value at @xmath179 ( north pole ) . left column : fixed center distance @xmath302 ( @xmath303 golden ratio ) , @xmath304 and @xmath287 from .
right column : total charge @xmath296 with @xmath305 , @xmath306 and @xmath299 from . from top to bottom : @xmath307 , @xmath179 , @xmath308 .
( no solution for @xmath309 in right column.),title="fig : " ] comparison of density functions of signed equilibria on @xmath161 associated with riesz external fields generated by a positive point charge above the north pole ( blue curves in colored version , i.e. smaller value at @xmath301 ) and a negative point charge below the south pole ( red curves in colored version , i.e. larger value at @xmath301 ) having the same density function value at @xmath179 ( north pole ) . left column : fixed center distance @xmath302 ( @xmath303 golden ratio ) , @xmath304 and @xmath287 from .
right column : total charge @xmath296 with @xmath305 , @xmath306 and @xmath299 from . from top to bottom : @xmath307 , @xmath179 , @xmath308 .
( no solution for @xmath309 in right column.),title="fig : " ] next , we investigate the signed equilibrium on spherical caps @xmath240 .
ultimately , our goal is to use this signed equilibrium to obtain the @xmath5-extremal measure on @xmath0 associated with the external field in when its support is not all of @xmath0 .
we remark that the proofs of the results and remarks in this section can be obtained by inspecting the proofs of the related results in @xcite ( details will be presented in a later paper @xcite ) . [ thm : signeq ] let @xmath310 .
the signed @xmath5-equilibrium @xmath239 on the spherical cap @xmath311 , @xmath312 , associated with @xmath236 in is given by .
it is absolutely continuous in the sense that for @xmath313 , @xmath314 where ( with @xmath315 and @xmath316 ) @xmath317 the density @xmath318 is expressed in terms of regularized gauss hypergeometric functions .
furthermore , if @xmath319 , the weighted @xmath5-potential is given by @xmath320 where @xmath321 and @xmath322 is the regularized incomplete beta function .
there is a simple relation between the positive and negative point charge problem with regard to their signed equilibria when these charges are on opposite sides of an axis .
namely , if @xmath323 and @xmath324 are the riesz external fields generated by a positive point charge at @xmath325 above the north pole and a negative point charge at @xmath233 below the south pole , then the weighted @xmath5-potentials of the respective signed equilibria @xmath326 and @xmath275 on the spherical cap @xmath240 satisfy @xmath327 everywhere on @xmath0 . furthermore , if @xmath328 , then the following principle holds : @xmath329 for all @xmath330 such that @xmath331 .
the next remark , leading up theorem [ thm : s.equilibrium.measure ] , emphasizes the special role of @xmath332 .
it can be shown that the signed equilibrium @xmath239 on @xmath240 associated with @xmath236 is a positive measure with support @xmath240 if and only if @xmath333 whereas the weighted @xmath5-potential of the signed equilibrium @xmath239 on @xmath240 associated with @xmath236 exceeds the value @xmath243 assumed on @xmath240 strictly _ outside _ of @xmath240 ( but on @xmath0 ) if and only if @xmath334 as the right - hand side above is negative , one also has the weaker restriction @xmath335 .
for @xmath239 to coincide with the @xmath5-extremal measure on @xmath0 associated with @xmath236 with support @xmath240 both and have to hold ( cf .
proposition [ prop:1](d ) ) .
the difficult part of the next statement is to verify that the arising equation has a unique solution , which can be done as in the proof of ( * ? ? ?
* theorem 13 ) .
[ thm : s.equilibrium.measure ] let @xmath336 . for the external field
the function @xmath337 given in has precisely one global minimum @xmath338 $ ] .
this minimum is either the unique solution @xmath339 of the equation @xmath340 or @xmath341 when such a solution does not exist .
in addition , @xmath243 is greater than the right - hand side above if @xmath342 and is less than if @xmath343 . moreover , @xmath344 .
the extremal measure @xmath281 on @xmath0 is given by @xmath345 ( see ) , and @xmath346 . in the limiting case
@xmath28 with @xmath49 it can be shown that the @xmath5-balayage measures @xmath347 exist and both have a component that is uniformly distributed on the boundary of @xmath240 .
each of these measures is the weak - star limit as @xmath348 of the respective measure in .
an inspection of the proofs in @xcite yields that the signed @xmath5-equilibrium @xmath349 on the spherical cap @xmath240 associated with @xmath350 ( @xmath271 and @xmath351 with @xmath220 ) is given by @xmath352 \overline{\nu}_t - q \overline{\epsilon}_t , \qquad \overline{\phi}_{d-2}(t ) { \,{:=}\,}w_{d-2}(\mathbb{s}^d ) \left ( 1 + q \left\|\overline{\epsilon}_t\right\| \right ) / \left\|\overline{\nu}_t\right\|.\ ] ] more explicitly , it can be decomposed into a continuous and a discrete part by means of@xmath353 where the density with respect to @xmath48 restricted to @xmath240 has the form @xmath354 and the boundary charge uniformly distributed over the boundary of @xmath240 is @xmath355.\ ] ] the vanishing of this boundary charge characterizes the critical distance @xmath254 for which @xmath356 becomes the @xmath191-extremal measure on @xmath0 associated with @xmath357 ( details will be presented in a later paper @xcite ) . a similar result holds for the logarithmic case @xmath27 on @xmath161 .
let @xmath325 be a fixed point in @xmath359 with @xmath360 .
it will be regarded as the source of an external field @xmath361 , where @xmath82 is a strictly convex and decreasing function .
then as we show below the optimal configuration ( minimizing ) on @xmath161 for every @xmath17 is an equilateral triangle perpendicular to the main axis passing through @xmath325 and the center of the sphere ; cf .
figure [ fig3 ] .
the intercept @xmath362 of the plane supporting this unique ( up to rotation about the main axis ) triangle with the main axis varies with @xmath82 . from these facts it is easy to see that @xmath363^{s/2 } } + 12 f ( 1 - 2 r t_0 + r^2 ) , \ ] ] where the negative intercept @xmath362 is the unique minimum of @xmath364 satisfying the relation @xmath365^{s/2 + 1 } } = 4 r \left [ - f^\prime ( 1 - 2 r t_0 + r^2 ) \right].\ ] ] it is interesting to note that for the case @xmath366 we deduce the well - known fact that the tetrahedron has minimal riesz-@xmath5 energy for four points on @xmath161 interacting via the riesz-@xmath5 potential ;
see the first frame of figure [ fig3 ] .
typical @xmath58-point @xmath9-fekete sets for weaker ( @xmath367 , left ) and stronger ( @xmath368 , right ) riesz external fields @xmath369 in the coulomb case @xmath370 for selected values of center distances @xmath371 , @xmath372 , and @xmath373 ( @xmath303 is the golden ratio ) ; compare with fig .
[ fig2].,title="fig : " ] typical @xmath58-point @xmath9-fekete sets for weaker ( @xmath367 , left ) and stronger ( @xmath368 , right ) riesz external fields @xmath369 in the coulomb case @xmath370 for selected values of center distances @xmath371 , @xmath372 , and @xmath373 ( @xmath303 is the golden ratio ) ; compare with fig .
[ fig2].,title="fig : " ] typical @xmath58-point @xmath9-fekete sets for weaker ( @xmath367 , left ) and stronger ( @xmath368 , right ) riesz external fields @xmath369 in the coulomb case @xmath370 for selected values of center distances @xmath371 , @xmath372 , and @xmath373 ( @xmath303 is the golden ratio ) ; compare with fig .
[ fig2].,title="fig : " ] we now establish that the aforementioned configuration is indeed optimal . for this purpose let @xmath374 be an optimal configuration . without loss of generality , we assume that the segment connecting @xmath375 and @xmath376 lies in the @xmath377-plane and is parallel to the @xmath378-axis , say , with parameterization @xmath379 and @xmath380 , @xmath381 . furthermore , we can write @xmath382 since @xmath383 , the convexity of the function @xmath384 implies that @xmath385 where equality holds only if @xmath386 or @xmath387 .
similarly , since @xmath388 the convexity of @xmath82 implies that @xmath389 where equality holds only if @xmath390 or @xmath387 .
moreover , from the fact that @xmath82 is strictly decreasing we get @xmath391 and equality holds only if @xmath392 or @xmath393 .
consequently , the optimal configuration must satisfy @xmath394 and @xmath395 . since the choice of @xmath375 and @xmath376 was arbitrary ,
the same is true for any pair , which implies that the configuration is an equilateral triangle with equidistant points from @xmath325 .
finally , it is easy to see that there is a unique solution for the intercept @xmath362 of the equilateral triangle and the axes . let four unit charges be restricted to move on the unit sphere @xmath161 and the fifth point ( placed on the polar axis above the north pole ) act as the source of an external field @xmath396 , where @xmath397 , @xmath398 , and @xmath221 the north pole . a @xmath399-fekete set minimizes the discrete weighted energy associated with @xmath400 , @xmath401,\ ] ] among all four point configurations @xmath402 on @xmath161 .
this optimization problem is highly non - linear and currently eludes explicit solution .
heuristic considerations based on the symmetries of the problem and backed by numerical experiments suggest three basic types of optimal configurations ( see figure [ fig2 ] ) : * ( a ) * a triangular pyramid @xmath403 with one point at the south pole and three points forming an equilateral triangle parallel to the equator ( one degree of freedom ) with discrete weighted energy @xmath404 \right),\ ] ] where @xmath405 denotes the intercept of the triangle s plane with the polar axis ; * ( b ) * the set @xmath406 consisting of two pairs of opposite points in planes parallel to the equator rotated by @xmath407 ( two degrees of freedom ) with discrete weighted energy @xmath408 where @xmath405 and @xmath409 denote the intercepts of these two planes with the polar axis ; and * ( c ) * the four points forming a square @xmath410 parallel to the equator ( one degree of freedom ) with discrete weighted energy @xmath411 where @xmath405 is the intercept of the square s plane with the polar axis .
using elementary calculus one can show that @xmath412 is a strictly convex function on @xmath413 with @xmath414 as @xmath415 and thus has a unique minimum at a @xmath416 satisfying @xmath417 similarly , the function @xmath418 is strictly convex on @xmath413 with @xmath419 as @xmath415 and thus has a unique minimum at a @xmath416 satisfying @xmath420 the set @xmath410 is a special case of @xmath406 when the two planes merge into one .
figure [ fig1 ] shows a comparison of the discrete weighted energy for these configurations for the coulomb case @xmath370 and for @xmath421 and @xmath422 as the distance of the external source varies .
( as we have no explicit formulas , numerics provide the optimal values of the free parameters , that is , the positions of the planes supporting the points of @xmath403 , @xmath406 and @xmath410 that yield the smallest respective weighted energy . )
deviation of the minimal discrete weighted energy of @xmath403 ( @xmath423 ) , @xmath406 ( @xmath424 ) and @xmath410 ( @xmath425 ) from the putative global minimum versus distance from center @xmath426 ( sampled at @xmath427 ) for @xmath421 ( left ) and @xmath422 ( right ) and coulomb case @xmath370.,title="fig : " ] deviation of the minimal discrete weighted energy of @xmath403 ( @xmath423 ) , @xmath406 ( @xmath424 ) and @xmath410 ( @xmath425 ) from the putative global minimum versus distance from center @xmath426 ( sampled at @xmath427 ) for @xmath421 ( left ) and @xmath422 ( right ) and coulomb case @xmath370.,title="fig : " ] in the case when @xmath421 ( weak field ) , the ( numerically ) smallest discrete weighted energy appears to occur for the @xmath403-configuration for all @xmath398 , whereas for @xmath422 ( strong field ) the optimal configurations change type from @xmath403 to @xmath406 to @xmath410 as the distance @xmath426 decreases and passes certain critical distances ; cf .
figure [ fig2 ] for typical examples .
similar to the case of the three point problem , the choice @xmath428 for the four point problem corresponds to the equilibrium configuration for riesz-@xmath5 energy with no external field for _ five _ points on @xmath161 . for @xmath370
this is illustrated in the first frame of figure [ fig2 ] .
recently , schwartz @xcite gave a computer - assisted proof that the triangular bi - pyramid @xmath429 is indeed the minimizing configuration for @xmath370 and @xmath430 .
for other choices of @xmath284 in @xmath431 different configurations , such as the square based pyramid , may arise ; see the second frame of figure [ fig2 ] .
typical * putative * @xmath432-point @xmath433-fekete sets for weaker ( @xmath434 , left ) and stronger ( @xmath368 , right ) riesz external fields @xmath435 , @xmath436 , in the coulomb case @xmath370 for three choices of center distances @xmath371 , @xmath372 , and @xmath373 ( @xmath303 is the golden ratio ) ; compare with fig .
[ fig3].,title="fig : " ] typical * putative * @xmath432-point @xmath433-fekete sets for weaker ( @xmath434 , left ) and stronger ( @xmath368 , right ) riesz external fields @xmath435 , @xmath436 , in the coulomb case @xmath370 for three choices of center distances @xmath371 , @xmath372 , and @xmath373 ( @xmath303 is the golden ratio ) ; compare with fig .
[ fig3].,title="fig : " ] typical * putative * @xmath432-point @xmath433-fekete sets for weaker ( @xmath434 , left ) and stronger ( @xmath368 , right ) riesz external fields @xmath435 , @xmath436 , in the coulomb case @xmath370 for three choices of center distances @xmath371 , @xmath372 , and @xmath373 ( @xmath303 is the golden ratio ) ; compare with fig .
[ fig3].,title="fig : " ]
the proof of theorem [ thm : main ] relies on the restricted version of the maximum principle for measures supported on @xmath0 ( see theorem [ thm : restr.max.principle ] ) .
the latter can be shown using the principle of domination and the maximum principle for riesz potentials . for convenience
we state them here .
[ prop : thm1.29 ] let @xmath437 .
suppose @xmath41 is a positive measure in @xmath438 whose potential @xmath439 is finite @xmath41-almost everywhere , and that @xmath440 is a @xmath441-superharmonic function .
then if the inequality @xmath442 holds @xmath41-almost everywhere , it holds everywhere in @xmath438 .
before we provide the proof of theorem [ thm : restr.max.principle ] and then show theorem [ thm : main ] , we recall some basic facts about the kelvin transformation ( spherical inversion ) of points and measures . inversion in a sphere is a basic technique in electrostatics ( method of electrical images , cf .
@xcite ) and , more generally , in potential theory ( cf .
@xcite and @xcite ) . the kelvin transformation ( of a function )
is linear , preserves harmonicity , and preserves positivity .
let us denote by @xmath443 the kelvin transformation ( stereographic projection ) with center @xmath444 and radius @xmath445 ; that is , for any point @xmath446 the image @xmath447 lies on a ray emanating from @xmath325 , and passing through @xmath136 such that @xmath448 the transformation of the distance is given by the formula @xmath449 the image of @xmath0 under the kelvin transformation is a hyperspace orthogonal to the radius - vector @xmath325 , which we can identify with @xmath159 in a natural way .
for @xmath450 , let @xmath451 denote the image of @xmath54 under the kelvin transformation @xmath443 .
clearly , @xmath451 is then contained in @xmath159 .
next , we recall the definition of the kelvin transform of measures . given a measure @xmath237 with no point mass at @xmath325 , its _
@xmath5-kelvin transformation _
@xmath452 is a measure defined by @xmath453 clearly , and imply the duality @xmath454 .
we also note that @xmath455 in the particular case when @xmath456 and @xmath457 we have @xmath458 suppose that @xmath91 or if @xmath459 , then @xmath460 .
any point @xmath325 in the level set @xmath461 , which has positive @xmath41-measure , can serve as a center of inversion for the kelvin transformation with radius @xmath445 given in .
select one such point as center .
define the measure @xmath462 \sigma_d$ ] .
then @xmath463 for all @xmath464 .
neither @xmath41 nor @xmath237 has a point mass at @xmath325 , hence gives @xmath465 since the right - hand side is @xmath466-superharmonic in @xmath159 , by proposition [ prop : thm1.29 ] the inequality extends to all @xmath467 .
the inverse kelvin transformation yields @xmath468 for all @xmath469 .
so , @xmath94 for all @xmath469 .
using a different center of inversion , one obtains @xmath470 .
suppose now that @xmath29 and @xmath27 .
using we write @xmath471 since @xmath472 has finite logarithmic energy , the principle of domination for logarithmic potentials ( * ? ? ?
* theorem ii.3.2 ) implies that this inequality holds everywhere in the complex plane .
the inverse kelvin transformation then gives the desired inequality on the sphere except at the center of inversion .
this restriction can be removed by moving the inversion center .
suppose to the contrary that there is a measure @xmath473 such that fails ; that is , there is a constant @xmath474 such that @xmath475 applying we obtain that @xmath476 from proposition [ prop:1 ] we have that @xmath47 has finite @xmath5-energy , therefore its support @xmath477 will have positive @xmath5-capacity .
hence , the @xmath5-extremal measure associated with @xmath477 is well - defined .
thus , we may integrate both sides of inequality with respect to @xmath478 and , using fubini s theorem , we derive @xmath479 recall that the gauss variational inequalities for the @xmath5-extremal measure @xmath478 state that ( see @xcite ) @xmath480 by the maximum principle the first inequality can be extended to all of @xmath1 if @xmath481 and still remains true on @xmath0 when @xmath482 by the sphere maximum principle ( see theorem [ thm : restr.max.principle ] ) .
thus , we get the contradiction @xmath483 this establishes relation for @xmath123 .
we derive similarly , utilizing instead .
assume there is a measure @xmath473 and a constant @xmath484 , such that @xmath485 integration with respect to @xmath486 yields that @xmath486 has finite @xmath5-energy ( recall that @xmath85 is finite and @xmath487 ) .
this implies that @xmath488 , and that the measure @xmath489 is well - defined . from
we get @xmath490 now , we integrate both sides of this inequality with respect to the @xmath5-extremal measure @xmath489 and using fubini s theorem we arrive at @xmath491 applying again the gauss variational inequalities now to @xmath492 and using the sphere maximum principle we obtain a similar contradiction .
let @xmath56 with @xmath49 . for brevity set @xmath493 and @xmath494 , where @xmath495 is the spherical cap @xmath496 .
then @xmath497 using and we get @xmath498 now we can rewrite as @xmath499 which also holds @xmath500-a.e . on @xmath501 . setting @xmath502 , we obtain that ( see also ) @xmath503 selecting a center of inversion @xmath504 , the potentials of the kelvin transformations @xmath505 and @xmath506 satisfy @xmath507 note that by the principle of domination ( see proposition [ prop : thm1.29 ] ) the first inequality holds on all of @xmath159 .
now , we can apply a de la valle poussin - type theorem ( see ( * ? ? ?
* section 3 ) , ( * ? ? ? * theorem 2.5 ) ) to conclude that @xmath508 the inverse kelvin transformation yields @xmath509 this implies that @xmath510 and that if @xmath511 , then @xmath512 .
the theorem for the riesz case @xmath56 with @xmath49 is proved . when @xmath456 and @xmath457 we utilize and modify the argument above using the regular principle of domination in the complex plane and the original de la valle poussin theorem ( * ? ? ?
* theorem iv.4.5 ) .
next we establish the proof of theorem [ thm : main2 ] , which allows for the important conclusion that the @xmath9-fekete sets @xmath141 ( that is , the supports of the minimizers of the discrete weighted @xmath5-energy associated with @xmath2 ) are contained in the extended support @xmath112 of the continuous minimizer @xmath47 . from theorem [ thm : main ]
we conclude that @xmath513 . using
, inequality yields that @xmath514 let @xmath486 be a multiple of the @xmath5-equilibrium measure on @xmath0 , so that @xmath515 for all @xmath516 .
then the last inequality becomes @xmath517 if @xmath518 , then we can extend this inequality to any @xmath519 by the lower semi - continuity of @xmath520 and the continuity of @xmath521 at such @xmath522 .
as @xmath523 and @xmath524 for @xmath525 , holds for all @xmath526 . if @xmath477 is a proper subset of @xmath0 , then we can find a point @xmath527 .
since @xmath47 has finite @xmath5-energy , inequality holds @xmath47-almost everywhere . using kelvin transform centered at @xmath325 with radius @xmath445
, we derive that @xmath528 the principle of domination ( see proposition [ prop : thm1.29 ] ) enables us to extend this inequality to all @xmath529 .
inverse kelvin transformation then implies for all @xmath526 except at the center @xmath325 and this restriction can be lifted by moving the center of inversion .
definition [ def : discrete.external.field.problem ] implies that any point @xmath530 is a global minimum for the weighted potential ( cf . )
@xmath531 , @xmath516 , where @xmath532 . hence , inequality holds with @xmath533 .
thus holds in particular for @xmath534 which reduces to @xmath535 .
therefore , @xmath536 . here
we prove our separation results .
the proof of theorem [ thm : main3 ] has three parts .
the first considers the case @xmath336 .
the second establishes explicit bounds in the limiting case @xmath28 in the field - free setting and thus shows proposition [ prop : well - separation.min.riesz.(d-2).energy ] .
the third considers the limiting case @xmath190 in the presence of an external field .
let @xmath336 .
we shall modify the approach from @xcite .
let @xmath537 be an @xmath10-point @xmath9-fekete set on @xmath0 .
each point in @xmath141 defines an external field by means of @xmath538 observe that @xmath539 is an @xmath540-point @xmath541-fekete set on @xmath0 .
since all the @xmath542-point @xmath541-fekete sets @xmath543 are contained in the extended support @xmath544 of the @xmath5-extremal measure @xmath545 on @xmath0 associated with @xmath546 by corollary [ cor ] , we will be done if we show that there is a spherical cap with radius @xmath547 such that when centered at @xmath548 its intersection with @xmath544 is empty .
note that the constant @xmath188 should only depend on @xmath2 and @xmath5 .
we shall use the fact that whenever the compact subset @xmath73 contains the support @xmath549 of the @xmath5-extremal measure @xmath545 on @xmath0 , then @xmath550 ( cf .
theorem [ thm : signsupp.2 ] and remark following it ) . from proposition [ prop:1](b )
it follows that @xmath549 is contained in a set where @xmath551 for some constant @xmath552 .
this implies that there is a spherical cap @xmath553 , @xmath554 , such that @xmath555 .
hence it suffices to consider the continuous riesz external field problem on @xmath556 ( instead of @xmath0 ) and study signed @xmath5-equilibria @xmath557 on spherical caps @xmath558 satisfying @xmath559 . the signed @xmath5-equilibrium @xmath557 on a spherical cap @xmath495 exists for any @xmath560 .
indeed , @xmath561 where @xmath562 is the dirac - delta measure with unit charge placed at @xmath548 and @xmath563 is a normalizing constant such that @xmath564 . using similar analysis as in ( * ? ? ?
* proof of theorem 1.5 ) we shall derive that this signed measure will be negative on a band containing the rim of the spherical cap @xmath495 for any @xmath565 , and in particular for @xmath566 .
hence , for any such @xmath567 the support of the positive part @xmath568 is contained in a smaller spherical cap @xmath569 with @xmath570 .
since @xmath544 is contained in any such @xmath571 , we conclude that @xmath572 , i.e. @xmath573 to finish the proof we have to show the negativity of the signed equilibrium measures @xmath574 near the rim for all such @xmath575 .
the middle term in , @xmath576 , can be written as @xmath577 with the notations @xmath578 and @xmath579 we rewrite as follows : @xmath580 note that inequality between two signed measures may be understood in terms of their densities or that the difference of the two signed measures is a positive measure .
since balayage may be done in steps , we have @xmath581 the superposition representation of the balayage measure yields @xmath582 the formula for the point mass balayage on the unit sphere ( see ( * ? ? ?
* theorem 2 ) ) yields @xmath583 where @xmath584 is such that @xmath585 . as balayage preserves positivity and is linear , application of balayage to the inequality @xmath586 yields @xmath587
combining and , and using that @xmath588 , we arrive at @xmath589 as balayage reduces the norm of a measure ( i.e. , @xmath590 and @xmath591 ) , we obtain @xmath592 the dominating signed measure at the right - hand side above has total charge @xmath593 ( which does not depend on @xmath594 and @xmath567 ) and can be rewritten as @xmath595}{\| \nu_\rho \| } \ , \nu_\rho - \frac{1}{c_{\sigma } \left ( n-2 \right ) } \ , \epsilon_\rho \right),\ ] ] where the parenthetical expression is the signed @xmath5-equilibrium on @xmath495 associated with the riesz-@xmath5 external field generated by a point charge of size @xmath596 $ ] at @xmath548 .
( there is no other dependence on @xmath594 than that @xmath548 determines the axis of symmetry of @xmath495 . )
it follows from ( * ? ? ?
* theorem 13 with @xmath371 ) ( also cf .
analysis in ( * ? ? ? * theorem 1.5 ) ) that this signed measure has a negative part for each @xmath597 , where the critical @xmath598 solves the equation . in @xcite
the `` altitude '' of the boundary is used . ]
@xmath599 and @xmath598 is also the unique minimizer of @xmath600 .
hence @xmath601 and it follows that @xmath602^{1/d } \leq \rho_0,\ ] ] where ( substituting for @xmath284 ) @xmath603 the last inequality holds for all @xmath604 such that @xmath605 and the constant @xmath188 takes the form given in which reduces to the constant given in ( * ? ? ?
* theorem 1.5 ) in the field - free setting .
specialization for @xmath157 gives .
this completes the proof of the theorem for @xmath336 .
let @xmath184 and @xmath28 .
we proceed similar as in the first part of the proof of theorem [ thm : main3 ] .
let @xmath606 be a riesz @xmath191-energy minimizing @xmath10-point configuration on @xmath0 .
each point in @xmath197 in turn can be identified with the north pole @xmath221 and thus defines a riesz external field @xmath607 where @xmath608 , so that @xmath609 is an @xmath540-point @xmath610-fekete set on @xmath0 .
every @xmath542-point @xmath610-fekete set on @xmath0 is contained in the ( rotational symmetric ) extended support @xmath611 of the @xmath191-extremal measure @xmath612 on @xmath0 by corollary [ cor ] .
it suffices to show that @xmath613 for some constant @xmath614 independent of @xmath10 .
observe , that the set @xmath611 is contained in the extended support of the @xmath191-extremal measure on a compact subset @xmath44 whenever @xmath16 contains the support @xmath615 of @xmath612 .
hence , by theorem [ thm : signsupp.2 ] and following remark , the set @xmath611 is a subset of the support of the positive part of the signed @xmath191-equilibrium measure on @xmath16 associated with @xmath2 whenever @xmath616 .
in fact , proposition [ prop:1](b ) yields that @xmath617 for some positive @xmath618 depending on @xmath10 . ( note that @xmath615 is a spherical cap by proposition [ prop : connthm ] . )
we deduce that @xmath619 of the positive part @xmath620 of the signed equilibrium measure @xmath621 on the spherical cap @xmath622 for each @xmath623 , where @xmath624 coincides with @xmath47 .
the radius @xmath625 and its estimate in terms of @xmath10 gives the desired lower bound for the separation of points in @xmath197 .
it can be shown that the @xmath191-balayage measures onto the spherical cap @xmath558 of the uniform measure @xmath48 and a unit point charge at @xmath221 , @xmath626 exist and both have a component that is uniformly distributed on the boundary of @xmath495 . from (
* theorem 15 ) , and letting @xmath627 , we obtain that @xmath628 with @xmath629 where @xmath630 using the formulas in @xcite it is easy to show that @xmath631 has a unique minimum in @xmath632 at the critical radius @xmath625 ( cf .
* theorem 15 ) ) .
this @xmath625 is the solution of the equation @xmath633 therefore , @xmath634 and the relations and follow after substitution of @xmath635 .
let @xmath29 and @xmath27 . in the logarithmic case the external field in
is replaced with @xmath636 where @xmath637 .
again , @xmath638 is an @xmath540-point @xmath639-fekete set on @xmath161 .
similarly as before we are led to the investigation of the signed logarithmic equilibrium problem on spherical caps @xmath495 .
theorem 17 in @xcite gives ( as @xmath627 ) that @xmath640 { \,d}\delta_{1-\rho^2/2}(u ) { \,d}\sigma_{d-1 } ( \overline{{\mathbf{x } } } ) , \quad ( \sqrt{1 - u^2 } \ , \overline{{\mathbf{x } } } , u ) \in \mathbb{s}^d.\ ] ] clearly , the signed logarithmic equilibrium @xmath641 has a negative boundary charge ( uniformly distributed over the boundary of @xmath495 ) for every @xmath623 , where at the critical distance @xmath642 this boundary charge vanishes ; that is , @xmath643 the relation for the separation follows .
let @xmath28 and @xmath184 .
we proceed as in the first part of the proof of theorem [ thm : main3 ] .
let @xmath644 be an @xmath10-point @xmath610-fekete set on @xmath0 .
each point in @xmath645 defines an external field by means of @xmath646 again corollary [ cor ] and proposition [ prop:1 ] guarantee that there is a spherical cap @xmath556 such that @xmath647 for every @xmath542-point @xmath648-fekete set on @xmath0 .
hence , we may study the continuous riesz @xmath191-external field problem on @xmath556 and use signed @xmath191-equilibria on spherical caps @xmath495 in our analysis .
the signed @xmath191-equilibrium @xmath649 on a spherical cap @xmath495 exists for any @xmath560 , since ( with the notations @xmath650 and @xmath651 ) @xmath652 where @xmath563 is a normalizing constant such that @xmath653 . in the limiting case @xmath28 , balayage introduces a boundary charge uniformly distributed over the boundary of @xmath495 . indeed , from (
* lemmas 33 and 36 ) , and letting @xmath627 , we have in ( * ? ? ?
* lemmas 33 and 36 ) . ]
@xmath654 furthermore , @xmath655 note that @xmath656 ( since it is subtracted in ) can only have a non - positive contribution to the boundary charge on @xmath495 .
this leaves @xmath657 .
iterating balayage , we have @xmath658 where @xmath659 . from
we get @xmath660 hence , using that the @xmath191-balayage also does not increase the norm ( i.e. , @xmath661 and @xmath662 ) , we arrive at @xmath663 the right - hand side has total charge @xmath664 and can be rewritten as @xmath665 , where @xmath666 is the signed @xmath191-equilibrium on @xmath495 for the riesz-@xmath191 external field due to a point charge of size @xmath667 $ ] placed at @xmath548 .
now we may proceed as in the last part of the proof of proposition [ prop : well - separation.min.riesz.(d-2).energy ] to conclude that the critical distance @xmath625 satisfies @xmath668 that is @xmath669^{1/d } , \qquad \frac{q}{q+1 } = \frac{1}{\overline{c}_\sigma ( n - 2 ) + 1 } \geq \frac{1}{\overline{c}_\sigma \ , n}.\ ] ] the result for @xmath28 and @xmath184 follows . in the logarithmic case @xmath27 and @xmath29
we follow the same argument utilizing ( * ? ? ?
* lemmas 39 and 41 ) .
note that the separation estimate in the logarithmic case is the limit as @xmath670 of the separation results for @xmath671 . in the proof of theorem [ thm : main4 ] we will use the following facts about _ spherical caps _ @xmath672
centered at @xmath516 with radius @xmath673 provided in @xcite ; namely @xmath674 and the estimate ( @xmath194 ) @xmath675 here , @xmath676 denotes the surface area of @xmath0 and @xmath677 reference @xcite also gives @xmath678 where @xmath679 as @xmath680 .
a finer analysis of the integral at the right - hand side gives the following explicit estimates .
[ lem : explicit.estimates ] the remainder term in satisfies @xmath681 , where @xmath682 \dfrac{s/2 - d}{\left ( s - d \right ) \left ( s - d - 2 \right ) } & \text{for $ 2d < s \leq 2d + 2 $ , } \\[1.05em ] \dfrac{s/2 - d}{d \left(d+2\right ) } & \text{for $ s > 2d + 2$. } \end{cases}\ ] ] the funk - hecke formula ( cf .
@xcite ) gives @xmath683 the standard substitution @xmath684 and ( * ? ? ?
* eq . 15.6.1 ) enables us to obtain an expression in terms of a gauss hypergeometric function by means of @xmath685 where the last relation follows from the last linear transformation in ( * ? ? ?
* eq . 15.8.1 ) . using that @xmath686
, we obtain the estimate @xmath687 where @xmath688 note that the hypergeometric function at argument @xmath179 evaluates as @xmath689 by ( * ? ? ?
* eq . 15.4.20 ) .
setting @xmath690 , the mean - value theorem and the differentiation formula ( * ? ? ?
* eq . 15.5.1 ) gives @xmath691 for some @xmath692 .
( the last step follows again from ( * ? ? ?
* eq . 15.6.1 ) . )
the result follows by considering the integral ( which reduced to a beta function integral ) for the three given cases @xmath693 , @xmath694 and @xmath695 and the observation that @xmath696 .
we follow the argument in ( * ? ? ?
* corollary 4 ) .
let @xmath697 be a @xmath9-fekete set on @xmath0 .
set @xmath698 for @xmath699 , where @xmath206 is such that @xmath700 and @xmath701 , and where @xmath702 .
setting @xmath703 , we have from that @xmath704 consider , for a given index @xmath705 , the function @xmath706 as @xmath707 minimizes the discrete weighted @xmath5-energy associated with @xmath2 , we have @xmath708 using the inclusions @xmath709 and , we have @xmath710 where @xmath711 as @xmath712 ; in fact , by lemma [ lem : explicit.estimates ] , @xmath713 dividing through by @xmath714 in , subtracting off the constant @xmath715 ( if it is negative ) from both sides of the new inequality and substituting the estimates above , we arrive at @xmath716 note that @xmath717 is well - defined and finite by the lower semi - continuity of @xmath2 and minimality of @xmath707 .
since @xmath718 ( @xmath699 , @xmath719 ) is a lower bound for the left - hand side , we have @xmath720 integrability ( with respect to @xmath48 ) of @xmath2 over @xmath721 and @xmath722 , where @xmath723 is the minimum of @xmath2 over @xmath0 , yield @xmath724 hence , the result follows from , and the definition of @xmath725 .
j. h. conway and n. j. a. sloane . , volume 290 of _ grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences]_. springer - verlag , new york , third edition , 1999 . with additional contributions by e. bannai , r. e. borcherds , j. leech , s. p. norton , a. m. odlyzko , r. a. parker , l. queen and b. b. venkov .
p. d. dragnev . on the separation of logarithmic points on the sphere . in _
approximation theory , x ( st . louis , mo , 2001 ) _ , innov .
math . , pages 137144 .
vanderbilt univ . press ,
nashville , tn , 2002 .
e. a. rakhmanov , e. b. saff , and y. m. zhou .
electrons on the sphere . in _ computational methods and function
theory 1994 ( penang ) _ , volume 5 of _ ser .
_ , pages 293309 .
world sci .
publ . , river edge , nj , 1995 .
|
we consider the minimal energy problem on the unit sphere @xmath0 in the euclidean space @xmath1 in the presence of an external field @xmath2 , where the energy arises from the riesz potential @xmath3 ( where @xmath4 is the euclidean distance and @xmath5 is the riesz parameter ) or the logarithmic potential @xmath6 .
characterization theorems of frostman - type for the associated extremal measure , previously obtained by the last two authors , are extended to the range @xmath7 the proof uses a maximum principle for measures supported on @xmath0 .
when @xmath2 is the riesz @xmath5-potential of a signed measure and @xmath8 , our results lead to explicit point - separation estimates for @xmath9-fekete points , which are @xmath10-point configurations minimizing the riesz @xmath5-energy on @xmath0 with external field @xmath2 . in the hyper - singular case @xmath11 , the short - range pair - interaction enforces well - separation even in the presence of more general external fields . as a further application , we determine the extremal and signed equilibria when the external field is due to a negative point charge outside a positively charged isolated sphere .
moreover , we provide a rigorous analysis of the three point external field problem and numerical results for the four point problem .
|
since the seminal work by skyrme @xcite , where the first three - dimensional topological defect solution arising from a nonlinear field theory was presented in the context of particle physics , a very large number of studies reporting the impact of some kind of topological defect has appeared in the physics literature and , in recent years , this area of research remains one the most active fields in many areas of the physics , including condensed matter physics @xcite , field theory topological - ft , and cosmology @xcite .
for instance , in an astrophysical scenario , topological defects arises largely from grand unified theories ( guts ) of elementary particles . in this case
, it is expected that the known local gauge symmetry group @xmath0 resulted from an underlying symmetry group @xmath1after a series of spontaneous symmetry breakings .
thus , the early universe has gone through a number of phase transitions , with one or more several types of topological defects possibly being left behind .
another interesting background , where we can find topological configurations , is high energy physics @xcite . in this context
, topological structures are found for example in the massive thirring model @xcite , in nonlinear chiral theories duff - isham , in the gross - neveu model @xcite , in a chiral @xmath2 @xmath3 @xmath2 model @xcite , in an artificial spin - ice lattice @xcite , and where the peccei - quinn symmetry is broken after inflation @xcite .
nowadays we know that the carbon is the key ingredient in the called organic chemistry . from the viewpoint of condensed matter physics , the carbon plays a crucial role in the formation of new classes of topological materials casprp09 with a large variety of physical properties @xcite .
in particular , the material named in the literature as graphene is the most famous due to their great potential for technological applications .
those materials are a two - dimensional ( 2d ) allotrope of carbon , which consists of a monotonic layer of carbon atoms on a honeycomb lattice .
its abundant electronic properties was studied for the first time using the tight - binding approximation @xcite , where the structure of the electronic energy bands and brillouin zones for graphite has been calculated .
although it is a nonrelativistic system , because the fermions move with a speed @xmath4 , which is 300 times smaller than the speed of light @xmath5 , the most remarkable feature of the graphene is that electronic properties can be described by the relativistic dirac equation , as initially pointed by semenoff @xcite .
thus , we can use the well - established methods of quantum field theory in order to study the physics and topological properties of the graphene@xcite . a special point of interest in graphene
are the topological excitations , which are a direct consequence of its strong chemical bonding @xcite responsible for assuring a stability in the geometrical properties of the underling lattice .
topological defects in graphene can lead to the appearance of an effective gauge field , which can be induced by wrinkles or ripples in its structures @xcite , when the graphene sheet is under tension @xcite , and by displacements along bond directions chamon .
topological excitations of the 2d lattice have its counterpart in the electronic properties of the material .
one example , is the jackiw - pi theory that predicts a dirac zero mode , associated with vortex solutions of the gauge field coupled to the electrons , leading to charge fractionalization @xcite .
the jackiw - pi theory was further generalized to non - chiral models @xcite , where the implications of different gauge groups in the gap formation , considering scalar , vector , and fermion excitations were studied .
in addition the finite - energy vortex solutions of the gauge models having the flux of the electromagnetic fields quantized leads to the bohm - aharonov effect even in the absence of external electromagnetic field . furthermore , we can find an interesting supersymmetric model for graphene in refs .
@xcite .
another example , of interest , is the possibility of solitonic excitations in the 2d lattice from the nonlinear solutions of the field equations in nanoribboins ( see e.g. @xcite ) .
the static deformation of the scalar field forming a domain wall in a graphene layer has been analyzed in @xcite , where it was discovered an electronic domain wall spectrum associated with localized electronic states transverse to the wall , and progating only in one dimension .
they studied gapped graphene and found three confined states , one the zero - mode and one in the conducting and valence band . in strained graphene nanoribbon
it was also studied a soliton trap which has electronic states confined to the wall associated with a dirac zero mode solution in the trap @xcite .
the topological confinement of electrons in bilayer graphene was proposed in @xcite , where also the associated spectrum was computed numerically ( for a recent review on the electronic properties of bilayer graphene see @xcite ) .
such theoretical proposal was experimentally verified by observing the valley transport properties at the bilayer graphene domain wall @xcite , and quite recently @xcite it was possible to control these one - dimensional conducting channels in a gapped set up .
the imaging of the confined states was done in @xcite by using scanning tunnelling microscopic , and it was found that the associated conducting channels are mainly located close to the two edges of the domain wall . we should remind that jackiw and rebbi @xcite in 1976 proposed that fermions can couple to bosonic soliton configurations in non - linear theories in ( 1 + 1 ) dimensions . in that work , it was concluded that there is a zero - energy bound state , where the solitons are degenerate doublets with fermion number @xmath6 . following this seminal paper ,
several works have been done with theories involving fermionic fields coupled to topological defects , such studies have included fermions and vortex solutions in abelian and nonabelian gauge theories @xcite , monopoles paolonpb1977 , supersymmetric field theories @xcite , condensed matter and relativistic field theories @xcite , nanorings @xcite , dense quark matter @xcite , aharonov - bohm and aharonov - casher problems @xcite , and topological superconductors @xcite .
motivated by the work of ref .
@xcite , chu and vachaspati chuprd2008 obtained the full spectrum of fermion bound states on a scalar z@xmath7 kink in ( 1 + 1 ) dimensions .
indeed , a decade before in davisprd1988 , it was speculated that the presence of possible fermionic bound states would inhibits the decay of cosmic strings .
the method developed in ref .
@xcite to solve the dirac equation on a kink background has been shown to be extremely powerful .
it enables to obtain analytically the bound state solutions .
later on , dutra and correa @xcite have extended such method to study fermion bound - states and zero modes in the background of kinks within models presenting a structure with a false and a true vacuum .
the spectrum of fermion bound - states of a class of asymmetrical scalar field potentials was analytically obtained in that work .
therefore , the findings of chuprd2008,rafaelplb2010 , in one - dimensional theory , suggest that a domain wall , or a soliton trap for electrons in 2d materials , should also presents an analytical representation of the rich spectrum and the corresponding states adding this feature to the previous finding of @xcite and @xcite . in the present work
we generalize the analytical methods developed in chuprd2008 for field theories of fermions coupled to scalar kinks in ( 1 + 1 ) to ( 2 + 1 ) dimensions in order to obtain the electronic spectrum in the presence of such topological defect .
we found analytically the excited states trapped by this domain wall , adding those to the ones previously obtained by @xcite and @xcite .
furthermore , we will explore the properties of the spectrum under the influence of constant external fields coupled to the fermions and also the associated impact due to the change of the kinetic term of the scalar field , which is given by lorentz symmetry violation terms in the lagrangian .
our motivation is to check the robustness of the electron spectrum when the properties of scalar field associated with modes of the 2d lattice are modified , in particular with respect to the propagation speed .
in fact , in a pioneer work , kostelecky and samuel @xcite proposed lorentz symmetry violation ( lsv ) in string theories , that was explored in several other physical scenarios ( see e.g. @xcite and references therein ) .
in particular , some works have explored topological defects in the presence of lsv .
for instance , we can find investigations using monopole and vortices @xcite , in a gauged @xmath8 sigma model @xcite , on topological defects generated by two real scalar fields @xcite , in the propagation of electromagnetic waves @xcite , and in traveling solitons systems lv48 , lv51 .
this paper is organized as follows : in section ii we present the theoretical framework of the sme which is going to be analyzed and we find the classical field configurations associated to it . in section iii we compute the full spectrum of two - dimensional fermion states . in section
iv we present our conclusions and final remarks .
in this section , we will show the general form of the relativistic lagrangian for a free spin-@xmath9 dirac fermion @xmath10 of mass @xmath11 in the so - called standard - model extension . at this point
it is important to remark that this theory is based on the idea of spontaneous lorentz breaking in an underlying theory and has been used for various investigations placing constraints on possible violations of lorentz symmetry , several of which depend crucially on the nonrelativistic physics of free massive fermions .
our development is a generalization of the work @xcite , where it was studied the kink - fermion coupling in @xmath12 dimension without lorentz breaking .
we extend the @xmath12 theory studied in @xcite to @xmath13 dimensions including lorentz breaking terms in the model in order to study the electronic properties of 2d materials in the presence of a kink .
the extended lagrangian density , adopted in the present work , has the following form : @xmath14 where @xmath15 , @xmath16 is a real scalar field , @xmath10 is a two - component spinor , @xmath17 is the potential , given in terms of a real scalar field , defining the bosonic sector of the specific model under analysis , @xmath18 is the corresponding yukawa coupling , @xmath19 is a dimensionless coefficient for lorentz violation that preserves _ _
cpt__. it can be taken as real , symmetric , and traceless .
the polarization tensor @xmath19 in the context of 2d materials , represents the change in the scalar mode propagation with the direction , and its speed can be different from the electron fermi velocity . furthermore , the quantities @xmath20 and @xmath21 are parameters that control the extent of lorentz and _ cpt _ violation in the theory . at this point , it is important to remark that , in the context of the standard - model and quantum electrodynamics ( qed ) extensions , these parameters are determined by expectation values of lorentz tensors arising from spontaneous lorentz breaking in a more fundamental theory . finally , in the above lagrangian density , the @xmath22 are the dirac matrices , which in this work will be written as @xmath23 \ , \ \
\gamma ^{1}=i\sigma ^{1}=i\left [ \begin{array}{cc } 0 & 1 \\ 1 & 0\end{array}\right ] , \label{gm1 } \\ \
\gamma ^{2 } & = & i\sigma ^{2}=\left [ \begin{array}{cc } 0 & 1 \\ -1 & 0\end{array}\right ] .\end{aligned}\ ] ] in order to justify the choice of such matrix representation , we point out that some years ago , these description has enabled to obtain the spectrum of fermion bound - states on the background of kinks of a class of asymmetrical scalar field potentials @xcite . in that work , such matrix structure allowed to find analytically the fermion configurations which has important consequences for cosmological and condensed matter systems .
another important matrix that will be useful is the @xmath24 . in this context , its representation in given by @xmath25 .
\label{gm2}\ ] ] note that for the bidimensional irreducible representation of the spinor space , the pseudovector and vector lorentz breaking terms in the lagrangian are the same , as the rotation of 180@xmath26 in the plane is equivalent to the parity transformation , and mathematically the algebra is closed with the identity and the three pauli matrices . assuming isotropy for the given frame , namely @xmath27 , the net effect of the corresponding lagrangian terms , produce only a trivial shift the dirac spectrum and from now on we set @xmath28 . here
, we would like to emphasize that , as expected , the gamma - matrices eq .
( [ gm1 ] ) and eq .
( [ gm2 ] ) satisfy the usual clifford algebra@xmath29 in this work , as an illustrative example , we choose the @xmath30 theory @xcite which is given by a symmetric double - well potential@xmath31 where @xmath32 and @xmath33 are real parameters . as we can see , this double - well potential has two minima , at @xmath34 .
moreover , there is reflectional symmetry since @xmath35 and the vacuum manifold has two - fold degeneracy . at this point
, it is important to remember that , in this model , the vacumm has an expectation value @xmath36 @xmath37 .
therefore , in such a vacuum , the elementary fermion will have a mass @xmath38 , where we are taking @xmath39 and @xmath36 @xmath40 . on the other hand ,
the mass associated to the scalar sector is @xmath41 . as we are studying a theory in @xmath13 dimensions , the tensor @xmath19 in eq .
( [ lagrangian ] ) is represented by a @xmath42 matrix written in the form @xmath43 it is important to remark that the above matrix has arbitrary elements .
however , if this matrix is real , symmetric , and traceless , the _ cpt _ symmetry is kept . recently
, a great number of works using a similar process to break the lorentz symmetry , with a tensor like @xmath19 , have been used in the literature , from microscope @xcite to cosmological scales @xcite . from the lagrangian density ( [ lagrangian ] ) , the euler - lagrange equation for @xmath16 is @xmath44 where @xmath45 . note that , in the above equations we are assuming that the back reaction due to the yukawa coupling between the dirac field and the scalar one can be neglected @xcite , in other words , the scalar field behaves like a classical background field @xcite . as we are interested in static solutions , where the field @xmath16 has dependence only in @xmath46 , namely @xmath47 .
thus , the above equation becomes @xmath48 where @xmath49 , @xmath50 with @xmath51 .
we observe that the dimensionless factor @xmath52 renormalizes the propagation velocity of the scalar mode with respect to the fermion maximum speed , namely @xmath53 this trivially breaks the covariance of the theory under lorentz boosts , but applying it to 2d materials , seems rather natural , as the propagation speed of perturbations in the lattice is not required to be the same as the one for electrons , which in the massless case , corresponds to the fermi velocity , that for graphene is @xmath54 .
an analytic solution of the above nonlinear equation , which interpolates between the different boundary conditions , is the so - called kink . in this case , follows that @xmath55 , \label{kinkphisolution}\ ] ] where @xmath56 is a constant of integration , it can be viewed as specifying the position of the kink . for simplicity , and without loss of generality , we can choose a localized kink centred about @xmath57 , which takes @xmath16 from @xmath58 at @xmath59 to @xmath33 at @xmath60 .
the size of the kink is controlled by : @xmath61 note that by varying the speed of the scalar mode by the lv term in the lagrangian the size of the wall can be tuned .
now , from eq .
( [ lagrangian ] ) , we have the following dirac equation in the kink background @xmath62where @xmath16 is the kink solution given by the eq .
( [ kinkphisolution ] ) written in terms of the @xmath46 variable .
from now on , we will solve explicitly the eq .
( [ dirac ] ) . at this point , it is important to highlight that we will only be interested in determining the bound states .
then , as we are working with two spatial dimensions , the fermions are described by two - component spinors , which we will write as @xmath63 .
\label{sd2}\ ] ] therefore , using the above representation into eq .
( [ dirac ] ) , and after some straightforward manipulations , we can find the following coupled pair of first order differential equations @xmath64 from now , we will turn our attention concerning the problem of how proceed to decouple the equations ( [ betap ] ) and ( [ betaequations ] ) . in this case , our goal will be to reduce the differential equations in schrdinger - like equations , which can be used with advantage in order to understand the essential features of the quantum mechanics of system under analysis .
then , after the substitution we obtain the corresponding second order differential equations @xmath65where we are using the corresponding definitions @xmath66 .\ ] ] and@xmath67 moreover , we can note that eq .
( [ schrodinger ] ) and eq .
( schrodinger1 ) are one - dimensional schrdinger - like equations with potential energy given by * * * * @xmath68 $ ] , and where both solutions has the same eigenvalue @xmath69 . in the next section , we will calculate both a zero mode and a fermionic bound state on the background of kink given by eq .
( [ kinkphisolution ] ) , with momentum along @xmath70 , transverse to the direction of action of the trapping potential .
in this section , we find the fermion bound states by solving the schrdinger equation on the background of the kink from the @xmath30 theory given by eq .
( [ kinkphisolution ] ) . in this case , we follow closely the approach developed by chu and vachaspati @xcite . for the clarity of the presentation and to define our notation , we repeat the essential steps of that work for the 2 + 1 case , including the lorentz symmetry breaking terms of the model . to begin , we rewrite the potentials @xmath71 using the kink solution eq .
( [ kinkphisolution ] ) . in this case , we get@xmath72where @xmath73 and @xmath74 is given by eq .
( [ alpha0 ] ) .
the @xmath75 factor gives @xmath46 in units of distance .
next , one needs to calibrate the model parameters to a real situation , as we explain in the following . in a recent bilayer graphene experiment
, it was obtained the images of conducting channels driven by topological states@xcite .
the domain wall was created by deformations of one of the layers .
the observed conducting channels correspond to states localized at the edges of the domain wall .
this experimental configuration leads to the formation of a topological zero mode states which is localized at the border of the domain wall .
naively we can expect the destruction of the peak at the center of the zero mode , as the potential created by the topological deformation for this experimental set up , should be deformed at the middle of the domain wall to include a small barrier at the center , which can imply a double humped state .
of course this effect is not included in our analytical model of the kink , as the effective potential is symmetrical and has the deep at the center of the kink . . ]
nevertheless the first excited shows the general characteristic peaks at the edges of the domain wall , and due to that we set the parameters of the model to reproduce the relative distance between the two maximum of the first excited state to the observed separation of the conducting channels of 8 nm .
we obtain a kink parameter of @xmath76 and @xmath77 mev . as we are going to see the deep at the minimum of the excited state probability density
is controlled by the ratio @xmath78 which should be at least about @xmath79 . to illustrate our choice of parameters , in fig .
[ figdensdistr ] we anticipate our results for the charge carrier density of the first excited trapped state , where it is clearly seen the separation of 8 nm between the peaks , and they are located at edge of the domain wall .
we will explore different cases of @xmath80^{-1}\ , , \ ] ] to check the effect of breaking the lorentz invariance , reminding that the fermi velocity for the electron is @xmath81 in graphene .
we should emphasize that the kink in the present topological model , is much simpler than the alluded experimental set up .
there are other interesting topological models of superposition of kinks @xcite , which could generate a potential closer to the experimental set up , however it is not yet known the analytical solutions of the trapped fermionic modes in these non - trivial kink backgrounds .
it is worth noting that for values of @xmath82 the potential @xmath83 has the asymptotic maximum of @xmath84 , and minimum value of @xmath85 at @xmath86 .
as it is well known from quantum mechanics in one - dimension , a time - independent attractive potential that tends to zero asymptotically will have at least one bound state .
in addition , the schrdinger - like equation derived for the upper component , has a particular combination of the kink and its derivative , which gives a zero mode @xcite . in general grounds ,
the zero mode solution of the dirac equation for a scalar domain wall presents a zero mode ( see e.g. @xcite ) , and was theoretically discovered in the graphene layer near the wall in @xcite . later on , in ref .
sasnjp10 the zero mode solution was found for the dirac electron for strained graphene nanoribbons in a soliton trap .
therefore , it is guaranteed that @xmath83 has at least one bound state for every @xmath18 , which is the zero mode . moreover , since @xmath83 gets deeper with increasing @xmath18 , there are more and more bound states that appear with larger values of @xmath18 @xcite . the domain wall created by a kink in 2d materials can present a rich spectrum near the wall driven by the effective potentials @xmath87 , illustrated in fig .
[ figvp ] .
in addition , the lorentz violating term drives the potentials by changing the relative speed of the scalar particle to the fermi velocity eq .
( [ velratio ] ) allowing to dial the spectrum with @xmath88 .
we observe in the upper panel of fig .
[ figvp ] that the increase of @xmath88 turns @xmath89 shallow but wider , which indeed increases the number of excited states , as can be seen in @xmath90 in the lower panel of the figure , which becomes deeper and wider allowing more states .
we are going to detail analytically such properties of the spectrum .
( upper frame ) and @xmath91 ( lower frame ) obtained with @xmath92 mev , @xmath93 and @xmath94^{-1}$ ] , for different ratios @xmath95 of @xmath96 ( dashed line ) , @xmath97 ( thin solid line ) and @xmath98 ( thick solid line).,title="fig : " ] ( upper frame ) and @xmath91 ( lower frame ) obtained with @xmath92 mev , @xmath93 and @xmath94^{-1}$ ] , for different ratios @xmath95 of @xmath96 ( dashed line ) , @xmath97 ( thin solid line ) and @xmath98 ( thick solid line).,title="fig : " ] in order to solve the dirac equation , we need a non - trivial bound state of the @xmath99 schrdinger - like equation which has the same energy eigenvalue as for @xmath100 .
only then @xmath101 will solve the first order equations , eq .
( [ betaequations ] ) , except if @xmath102 and for that we can take @xmath103 . for @xmath104 , @xmath105 is in the shape of a potential barrier and clearly has no bound states .
this shows that for @xmath104 , the only possible bound state is with @xmath102 and @xmath106 ; the and the zero mode solution is @xmath107where @xmath108 .
this zero mode solution near the wall matches the one found in @xcite .
this state corresponds to a massless particle moving in the transverse direction , @xmath70 , with momentum @xmath109 , and total energy @xmath110 , as well .
now , since that there are more bound states for @xmath111 , we will begin our search for these bound states . on the other hand ,
let us now carry out the quantitative calculation of the stationary states @xmath112 , for do this we write @xmath113where @xmath114 . then , substituting the eq .
( [ q1 ] ) into eqs .
( [ schrodinger ] ) and ( [ schrodinger1 ] ) , we obtain @xmath115 \mathrm{sech}^{2}(\alpha _ { 0}z/(\hbar v_{f}))f_{+}(z)=0,\right .
\notag \\ & & \notag \\ & & \left
. \partial _ { z}^{2}f_{-}+2\alpha _ { 0}\vartheta \tanh ( \alpha _ { 0}z/(\hbar v_{f}))\partial _ { z}f_{-}\right .
\label{q4.1 } \\ & & \left .
+ \left [ \alpha _ { 0}^{2}\vartheta ( 1+\vartheta ) -g(g-\alpha _ { 0})\right ] \mathrm{sech}^{2}(\alpha _ { 0}z/(\hbar v_{f}))f_{-}(z)=0.\right .
\notag\end{aligned}\ ] ] according to @xcite , by the variable transformation @xmath116 , \]]the equations ( [ q2.1 ] ) and ( [ q4.1 ] ) , becomes @xmath117 f_{\pm } ( u)=0,\right .
\label{q5 } \\ & & \notag \\ & & \left .
u(u-1)\partial _ { u}^{2}f_{-}(u)+(\vartheta + 1)(2u-1)\partial _
{ u}f_{-}(u)+\right . \notag \\ & & \left .
\left [ \vartheta ( \vartheta + 1)-g(g-\alpha _ { 0})/\alpha _ { 0}^{2}\right ] f_{-}(u)=0,\right . \label{q7}\end{aligned}\ ] ] as we can see , the above equations are the so - called hypergeometric equations , which has the following solutions@xmath118 \right .
\\ & & \left .
+ \mathcal{n}_{+}^{(2)}\mathcal{f}\left [ a_{+}+1-c_{+},b_{+}+1-c_{+};2-c_{+};u\right ] , \right . \notag \label{q9}\end{aligned}\]]@xmath119 \right . \\ & & \left .
+ \mathcal{n}_{-}^{(2)}\mathcal{f}\left [ a_{-}+1-c_{-},b_{+}+1-c_{-};2-c_{-};u\right ] , \right .
\notag \label{q12}\end{aligned}\ ] ] where the arguments of the hypergeometric function are defined as @xmath120 therefore , we can write the solutions for @xmath112 in the following form @xmath121\right .
\label{q151 } \\ & & \left .
+ \mathcal{n}_{\pm } ^{(2)}e^{\alpha _ { 0}\vartheta z}\mathcal{f}[a_{\pm } -c_{\pm } + 1,b_{\pm } -c_{\pm } + 1;2-c_{\pm } ; u].\right
. \notag\end{aligned}\ ] ] as shown in @xcite , for reason of normalizability , it is necessary to impose that @xmath122
. thus , the equation ( [ q151 ] ) becomes@xmath123 . \label{q16}\ ] ] in addition , we also have the condition @xmath124 therefore , after straightforwardly mathematical manipulations , we obtain the energy spectrum @xmath125 } , \notag \\ & & e_{n_{+}}=\pm \sqrt{(\hbar \omega ) ^{2}+n_{+}\alpha _ { 0}(2g-\alpha _ { 0}n_{+})}\ , .
\label{band1}\end{aligned}\]]at this point , it is important to remark that the energy eigenvalues coincide @xmath126 , which implies in the consistence condition @xmath127 .
moreover , the normalizability requires @xmath128 , as a consequence we have the constraint @xmath129the consistence condition for the quantum numbers leads to @xmath130 , where @xmath131 .
just to remind the zero mode is indexed with @xmath132 .
our solution reduces to the ones found in @xcite , when @xmath133 , namely the lv term of the scalar field in the lagrangian vanishes .
note that as have shown the zero mode still survives , independently of this lv interaction term .
it builds a zero gap band , as already found in previous works @xcite and @xcite .
as we have already discussed , the motivation for the choice of the parameters comes from the experimental work @xcite , where they found by imaging the confined states that the associated conducting channels are mainly located close to the two edges of the domain wall . in that experiment
the distance between the two peaks is about 8 nm . in our model
the zero mode is peaked at the center of the domain wall , while the first excited state is concentrated at the edges of the domain wall . in order to give some concrete scales of the spectrum
, we find the gap @xmath134 between the valence and conducting bands belonging to the first excited state in the trap potential obtained with @xmath135 and @xmath136 in eq .
( band1 ) : @xmath137with @xmath76 and @xmath77 mev , one gets @xmath138 mev , which is quite large when compared to the value of 80 mev from @xcite .
this deference comes from the description of potential in the experimental situation , where one expects to have a wider potential with a barrier in the middle and therefore the level position can be shifted down . the spectrum for these parameters is shown in fig .
the dispersion relation for the zero mode and the first 38 excited states are presented .
the chosen parameters leads to @xmath139 from eq .
( [ numberexc ] ) , giving a large number of conducting channels , concentrated at the edges of the kink .
we observe that the quantity @xmath140 can be interpreted as an effective mass for each of the excited trapped states , which can be written in terms of the gap as : @xmath141 and @xmath142 for the first 39 states , including the zero mode .
in this situation the number of states is 101 from @xmath139 as follows from eq .
( [ numberexc]).,scaledwidth=45.0% ] on the kink background the wave function of the n - th fermionic state , @xmath143 , trapped by the wall , for positive and negative energies , including the zero mode solutions are built from eqs .
( [ sd2 ] ) , ( betazm ) and ( [ q16 ] ) .
the zero mode eigenstate of the dirac hamiltonian trapped by the soliton is : @xmath144 , \right.\]]which is the solution already found and thoroughly discussed in sasnjp10 , apart the lv effect .
we have to observe that the zero - mode state in the layer corresponds to the solution of the one dimensional dirac equation with a `` massless '' fermion , as the dispersion relation is @xmath145 .
in addition the spinor state is an eigenstate of @xmath146 .
the trapping potential gives a width to the charge distribution along the @xmath46 direction , as carried the @xmath147 term .
the lv term in the scalar field lagrangian changes only the velocity of the scalar particle @xmath148 , and in the zero mode solution it only modifies the @xmath149 and the net effect is to broaden or shrink the wave function , as one sees in fig .
[ fig : zm ] for @xmath88 equals to 0.1 , 0.05 and 0.02 ; as the velocity ratio becomes larger wider is the state density , as the effective potential also becomes wider as shown in fig .
[ figvp ] .
that comes because the value of @xmath149 according to eq .
( [ alpha0vphivf ] ) , is inversely proportional to @xmath88 , and @xmath149 larger / smaller implies in confining potentials narrower / wider .
we have shown results for small values of the velocity ratio and of course could be enhanced by changing more drastically the ratio .
this effect could be useful to dial the width of the charge density in the fermion zero mode state in the solitonic trap by modifying the properties of the collective mode propagation , either doping the 2d material , or introducing defects on the lattice , or by a varying strain , etc .
this one - dimensional conducting state lives at the middle of the domain wall , while as we are going to show the higher excited states are located mainly close to the edges of the kink . and @xmath150^{-1}$ ] , for different ratios @xmath151 of @xmath96 ( dashed line ) , @xmath97 ( thin solid line ) and @xmath98 ( thick solid line ) . ] along the wall , with @xmath152 mev , @xmath153 mev and @xmath77 mev.,scaledwidth=75.0% ] the wave function of the excited states in the trapping potential of the kink with @xmath154 , are given by : @xmath155 , \right.\end{aligned}\]]where@xmath156@xmath157 \right ] , \notag \\ & & \\
f_{-}(z ) & = & \mathcal{f}\left [ a_{-},b_{-};c_{-};\frac{1}{2}\,\left [ 1-\tanh \left ( \alpha _ { 0}z/(\hbar v_{f})\right ) \right ] \right ] . \notag \\ & & \end{aligned}\ ] ] the effect of the lorentz symmetry breaking term in the scalar field lagrangian , which corresponds to the scaling of the propagation velocity by @xmath158 or @xmath52 , allows to increase / decrease the number of trapped states by increasing / decreasing the speed of the scalar mode in the 2d material , as the condition for the number of states ( numberexc ) shows for a fixed value of @xmath1 .
furthermore , the decrease in @xmath149 increase the density of states . in fig .
[ figspect ] , we illustrate the charge carrier density for the lowest states obtained with @xmath159 mev , @xmath160 mev and @xmath161 500 mev . in this case
the density is dominated by @xmath162 due to the energy factor , as seen in eq.([wfexc ] ) .
the effect of the lv term expressing the change in the relative speeds of the electron and the scalar quantum , modifies @xmath149 as shown by eq .
( [ alpha0vphivf ] ) , impact both the trapped excited spectrum and in the associated charge carrier densities .
for example , by decreasing @xmath149 from the increase of the scalar mode velocity with respect to @xmath163 , the domain wall becomes softer the charge carrier density swallows .
in this work we have investigated the full spectrum of two - dimensional fermion states in a scalar soliton trap with a lorentz breaking background .
it is important to remark that in the context of the novel 2d materials , the lorentz symmetry should not be strictly valid for a field theory that aims to represent the charge carriers close to the dirac points and the collective modes of the lattice represented by bosonic fields .
the dirac electron and bosonic fields reflect the non - relativistic dynamics of the electrons and lattice . in that sense , it is natural that the bosonic fields have propagation velocities different from @xmath163 , and the effect of this kind of lorentz violating effect should be expected when modelling the electronic and lattice excitations in 2d materials , like graphene .
therefore , it is necessary that the lorentz symmetry breaking should be taken into account in the chosen field theory model . having that motivation , we have performed an analysis of this particular lorentz violation effect in a 2d material with a relativistic model of the electrons and fields .
we explore theoretically , the asymmetry between the fermi and scalar mode velocities , in the wave function of the trapped electrons in a kink background . to accomplish
that , we extended the analytical methods developed in the context of 1 + 1 field theories to study the role of confined fermions to the stability of cosmic strings ( see @xcite ) , to explore the effect of the lorentz symmetry breaking in the charge carrier density of 2d materials in the presence of a domain wall with a kink profile .
the width and the depth of the trapping potential from the kink in the presence of the lorentz violating effect associated with the asymmetry of the fermi and scalar mode velocities can be manipulated , and is reflected in the band structure and properties of the levels localized along the kink axis and propagating in the transverse direction .
the present model provides an analytical form for the rich spectrum enlarging a previous findings of the edge states obtained with domain wall @xcite and in particular in strained graphene nanoribbon in a chiral gauge theory @xcite . in future
it will be interesting to generalize our analysis , when transitions between the two valleys are allowed in the field theoretical model , which leads a four dimensional representation of the spinor . in summary ,
the bound states solutions of dirac equation in the kink background are not destroyed by the lorentz violation due to the different velocities of the dirac electron and boson modes .
one should also expect that theories with gauge fields that aims to describe electronic properties of 2d materials are not necessarily covariant , and lorentz symmetry breaking effects should be taken into account .
we found that by tuning the lorentz breaking parameter , it is possible to change the density of trapped states and make the associated charge carrier density wider or narrow , which may be of interest in actual applications of 2d materials to electronic devices .
reversely , one could also search the consequences of lorentz violation in controlled table top experiments with 2d materials , which could be relevant to understand subtle phenomena in the universe .
racc would like to thank so paulo research foundation ( fapesp ) , grant 2016/03276 - 5 , for financial support .
wdp , asd and tf thank to capes , cnpq and fapesp for financial support .
racc gratefully acknowledge jos abdalla helayel - neto for discussions at an early stage .
a. j. chaves , g. d. lima , w. de paula , c. e. cordeiro , a. delfino , t. frederico and o. oliveira , phys .
b * 83 * , 153405 ( 2011 ) a. j. chaves , t. frederico , o. oliveira , w. de paula and m. c. santos , j. phys
. condens .
matter * 26 * , 185301 ( 2014 ) c. popovici , o. oliveira , w. de paula and t. frederico , phys .
b * 85 * , 235424 ( 2012 ) e. m. c. abreu , m. a. de andrade , l. p. g. de assis , j. a. helayel - neto , a. l. m. a. nogueira , and r. c. paschoal , j. high energy phys .
* 05 * , 001 ( 2011 ) .
|
the full spectrum of two - dimensional fermion states in a scalar soliton trap with a lorentz breaking background is investigated in the context of the novel 2d materials , where the lorentz symmetry should not be strictly valid .
the field theoretical model with lorentz breaking terms represents dirac electrons in one valley and in a scalar field background .
the lorentz violation comes from the difference between the dirac electron and scalar mode velocities , which should be expected when modelling the electronic and lattice excitations in 2d materials .
we extend the analytical methods developed in the context of 1 + 1 field theories to explore the effect of the lorentz symmetry breaking in the charge carrier density of 2d materials in the presence of a domain wall with a kink profile .
the width and the depth of the trapping potential from the kink is controlled by the lorentz violating term , which is reflected analytically in the band structure and properties of the trapped states .
our findings enlarge previous studies of the edge states obtained with domain wall and in strained graphene nanoribbon in a chiral gauge theory .
|
the characteristic energies of gamma - rays are measured in _ _ energy units of mev__ev ) , corresponds to a wavelength of about 10@xmath4 m or a frequency of 10@xmath5hz . ] , 56 orders of magnitude above the typical energies of atomic transitions which shape spectra in the optical domain .
cosmic gamma - rays thus are messengers of high - energy processes in cosmic sites .
the typical binding energy of electrons in atoms are several ev , while binding energies of nucleons in atomic nuclei are of order several mev ; hence , mev gamma - rays are often related to _ nuclear transitions_. characteristic temperatures for gamma - ray emission , according to wien s displacement law , would be 10@xmath6k ( = gk ) for thermal gamma - rays ; at such temperatures , objects would be unstable unless confined , e.g. the interior of a star may be at temperatures of millions of k as the large gravitational mass holds the object together , while nova and supernova explosions feature gk temperatures in their interiors .
mev energies are above the rest mass energy of electrons , hence electrons and positrons at these are _
relativistic_. = 6 is the subject of the current paper .
( adapted from @xcite ) . ]
.strongest spectral lines in the nuclear - physics part of the gamma - ray domain between @xmath7 0.1 and 10 mev , and their characteristics [ cols="^ , < , < " , ] [ tab : lines ] astrophysical sources of cosmic gamma - rays , therefore , are : \(1 ) nuclear - burning sites such as stellar explosions , as they release radioactive nuclides which have been produced through nuclear fusion reactions in their hot interiors @xcite .
these could be supernovae , such as from core - collapse of a massive star at its terminal phase of stellar evolution ( _ supernova types ii and ib , ic _ ) , thermonuclear supernovae related to disruptions of a white dwarf star ( _ supernovae type ia _ ) , or novae , which are thermonuclear events on the surface of a white dwarf star , which are not disruptive to the host star itself ( reviews and more detail on supernova variants can be found in @xcite and @xcite ) .
thermonuclear events on more - compact objects such as neutron stars and black holes will be less likely to lead to significant expansion of the nuclear - burning site , and thus remain non - transparent to gamma - rays produced herein ; on neutron star surfaces , such events are termed _ type - i x - ray bursts _ @xcite . also , some products of nuclear burning in the cores of normal stars may , under special circumstances , be mixed out to the stellar surface , and hence be released as part of the stellar wind , e.g. in the wolf - rayet phase of stars more massive than about 25 m@xmath8 .
several lines have been observed , from supernovae , and from radioactive nuclides accumulated in interstellar space ( see table [ tab : lines ] ) .
\(2 ) high - energy collisions in objects or interstellar space , which lead to excitations of nuclear levels , followed by de - excitation with accompanied characteristic gamma - ray line emission @xcite .
such high - energy interactions also produce continuum gamma - rays through the processes of inverse - compton scattering , bremsstrahlung , and other radiation processes related to acceleration of charges in strong fields such as curvature radiation and synchrotron emission . for the purposes of this paper ,
we focus on spectral signatures rather than continuum emission .
such high - energy collisions occur in interstellar space through cosmic - ray bombardment of interstellar gas , but also near cosmic relativistic - particle accelerators such as solar flares , supernova remnants , pulsar wind nebulae , accreting binaries and supermassive black holes , and gamma - ray bursts .
often , in these sites , continuum processes dominate in brightness , and therefore only in solar flares have the characteristic spectral - line signatures from nuclear transitions been detected up to now .
interactions of low - energy cosmic rays ( lecr ) appear to be closest to detection thresholds of current telescopes , while characteristic lines from nuclear transitions in accreting compact stars and their plasma jets are beyond reach at present .
\(3 ) annihilation of particles with their anti - particles , such as electron - positron annihilation , which results in a characteristic line at 511 kev energy from two - photon annihilation , and a characteristic spectral feature extending from 511 kev down into the 100 kev region , which originates from 3-photon annihilation from the triplet state of the positronium atom which is formed on the most - likely annihilation pathways @xcite .
this characteristic emission has been mapped to occur in an extended region throughout the inner parts of our galaxy , has been reported also from a few transient events in compact sources , and has been observed in great detail from solar flares @xcite .
at gamma - ray energies , interactions with the detector material of the astronomical telescope have a character different from lower energies : up to x - ray energies , photons interact within a small volume , while at gamma - ray energies the penetration depth into any material is macroscopic , of size mm to cm .
therefore , the mirror or lens optics which characterizes astronomical telescopes at lower energies ( or longer wavelengths ) becomes unfeasible , and the aperture of the telescope is identical to the detector s surface area .
the detector itself , correspondingly , converts the photon energy into a signal across a macroscopic depth range rather than a thin surface region .
the physical processes which are responsible for such interaction are \(1 ) compton scattering , i.e. the inelastic scattering of the incident photon on an electron , which produces a secondary photon with reduced energy and an energized secondary electron , and \(2 ) pair production , i.e. the conversion of the electromagnetic energy of the photon into a pair of electron and anti - electron ( positron ) in the electric field of a nucleus of the detector material .
pair production has a threshold energy of 1.022 mev , the rest mass of the pair of electron and positron . the excess energy and momentum of the incident photon
are distributed to the pair of particles , and thus the tracks of these ionizing particles can be traced with appropriate detectors .
bremsstrahlung may convert some of the particle energy of the secondary into photon energy , however , and may complicate tracking due to the larger range of photons in material compared to the charged leptons .
compton scattering results in a secondary photon , with an energy of about the same order of magnitude as the primary photon , hence with a similar range within the detector before a second interaction and partial energy deposit occurs .
therefore , successive compton scatterings and their total range constitute the upper end of the required detector size .
the interaction depth will vary from photon to photon , reproducing the probability distribution given by the effects of compton scattering ( between about 0.1 and about 2 mev ) and pair production ( at higher energies ) , respectively .
but event by event , interaction depths will vary by macroscopic amounts .
this leads to a degradation of imaging resolution , from the scatter of interaction depth in the detector pixel .
similarly , spectroscopic resolution depends on the homogeneity of energy deposits and its conversion into the signal amplitude across the pixel depth ( and area ) . at the high - resolution limit , the intrinsic energy distribution of the electrons in the detector material , with which the photon interacts , sets a lower limit on energy resolution : atomic electrons which compton - scatter the incident photon will have different energies according to their atomic orbits , and impose statistical fluctuations at the tens - ev level to the energy of the secondary electron from such compton scatter , which together with the secondary photon carries the energy of the incident photon .
the secondary photons and particles which are characteristic for high - energy interaction processes of photons with materials have energies below and up to the primary photon s energy , say mev .
in contrast to detectors at gev energies or above , the number of secondaries is rather small , and the number of total interactions from the primary interaction down to deposition of the total energy of the primary photon also is relatively small . in gev detectors , secondaries produce copious secondary ionization , and , since energies are large , one may add thick passive intervening mass layers into the path of secondaries to simply multiply their number , thus enhancing the ionization trace of secondaries . at mev energies ,
each single interactions is precious and should be measured with as much precision as feasible , to maximize the information about the primary photon , specifically about its energy and its arrival direction . for these reasons ,
the technology of _ compton telescopes _ has been assessed to be most promising and effective as an astronomical instrument . here
, a large active detector volume ensures that both primary and secondary interactions can be measured in sufficient detail through measurements of energy depositions and momentum vectors .
still , the interpretation of these signals is not straightforward , as the compton scattering , bremsstrahlung , and pair processes can not be inverted directly into reconstruction of the primary photon s energy and arrival direction @xcite . the statistical fluctuations inherent to measurements of energy deposits and track directions add to the resolution limitations of detector elements in spatial and energy dimensions , and are responsible for instrumental limitations .
additionally , the lack of focusing and therefore identity of photon detection and collection areas imply that the instrumental background present in the detector itself is more important , not being suppressed by a collection / detection enhancement factor which is common to other telescopes
. moreover , instrumental background dominates the total number of measured events , and those are difficult to disentangle from the few desired cosmic - photon detection events .
this is due to the physical origins of background , which are mostly the same processes that also create cosmic photons of interest : high - energy particle collisions from cosmic - ray bombardment of the instrument and its supporting structures result in continuum photons as well as nuclear line emission , as discussed above .
the _ prompt _ background events can be rejected when the primary charged particle or cosmic ray can be detected with a suitable detector system , and used in anti - coincidence to eliminate detection events which occur within a short time window after the high - energy particle traversal .
but _ delayed _ background events from short- or long - lived radioactivities or nuclear excitations can not be rejected .
these must be identified from their behavior in data space and their characteristics , as they might differ systematically from cosmic photon events .
such filtering is not perfect , due to statistical broadening of signals , and also may eliminate some desired cosmic photon events .
therefore , the processing of gamma - ray detector signals and rejection / suppression of background is complex , and often a major threshold for astronomers to make use of mev gamma - ray telescope data in broader studies .
mev gamma - ray spectroscopy is mostly done by a small group of astrophysicists which also are specialists in such instruments .
cosmic gamma - ray spectroscopy was initiated by space exploration programs and nuclear radiation detectors in the 1960ies .
the first cosmic lines reported were the positron annihilation line near 500 kev from the sun and from the galaxy , later the detection of interstellar decay of @xmath2al , and the first detection of supernova radioactivity gamma - rays in sn1987a .
broader studies of the gamma - ray sky were then undertaken through nasa s compton gamma - ray observatory mission ( 1991 - 2000 ) , and through esa s international gamma - ray astrophysics laboratory ( integral ) mission .
gamma - rays from the sun and its flares had been studied with the solar maximum mission ( smm ) ( 1980 - 1989 ) and its scintillator - based gamma - ray spectrometer instrument ( forrest et al .
1980 ) , and later with the reuven ramaty high - energy solar spectroscopic imager ( rhessi ; 2002present ) @xcite , which featured ge detectors and corresponding high ( kev - type ) spectral resolution at gamma - ray energies .
several balloon - borne instruments contributed also significant spectroscopy results , such as the ge detector equipped experiments of gris @xcite and hexagone @xcite in the 90ies , and more recently the tgrs experiment on the wind spacecraft @xcite and the nuclear compton telescope ( nct ) balloon experiment @xcite .
cgro had two gamma - ray spectroscopy instruments on board , the comptel compton telescope @xcite , and the osse spectrometer @xcite ; both featured a modest spectral resolution of order 10% ( fwhm ) .
this is inadequate for any gamma - ray spectroscopy in the sense of identifying new lines or constraining kinematics of source regions through line shape measurements , as we now know from high - resolution instruments based on semiconductor detectors , such as integral s spi spectrometer which provide resolutions of @xmath7 0.1% .
the grse instrument which was originally foreseen to address nuclear - line spectroscopy on this spacecraft , was removed during mission preparation following cost overruns . a nuclear astrophysics explorer ( nae) @xcite mission was discussed for a while , and was integrated later through a gamma - ray spectrometer into esa s integral mission concept @xcite .
but the first - generation instruments on cgro were essential in exploration of the gamma - ray sky , and comptel s imaging capabilities at mev energies are still the best we have .
the integral space mission of esa @xcite is a current example of cosmic gamma - ray instrumentation .
the satellite platform hosts two main instruments employing the coded - mask technique of imaging , the spi spectrometer and the ibis imager , in addition to monitor instruments , the jem - x coded - mask x - ray telescope addressing lower energies with a wider field of view , the omc optical monitor camera for simultaneous optical exposure of the target sky regions , and the irem charged - particle monitor detector system recording the cosmic - ray irradiation .
the field of view of the main instruments are of similar size of order 10@xmath9(ibis 9x9@xmath9 , spi 16x16@xmath9 , fully - coded parts , corner - to - corner ) .
the large field of view is defined by the hexagonal arrangements of the 19-element ge detector camera ( 70 mm deep , 55 mm wide detectors , covering a densely - packed 268 mm wide camera plane ) and the 127-element tungsten mask with 6 cm wide elements , which is placed 170 cm above the camera itself .
_ dithering _ the satellite pointing direction around a celestial target direction in @xmath7 2@xmath9steps helps to add more modulation to the sky signal , so it can be separated from the ( @xmath7 constant ) instrumental background .
the background should not vary within small attitude changes , and a very eccentric 3-day orbital period reaching far out and away from radiation belts ( maximum distance from earth 150,000 km)was chosen to provide slow background variations .
typically , the satellite pointing is changed every @xmath7 1800 s , and observations way vary from tens of ks to ms , depending on science target .
the spi spectrometer is integral s instrument for gamma - ray spectroscopy .
it features a set of 19 coaxial ge detectors operating in the energy range 158000 kev , with a detection area of 250 @xmath10 total and a nominal sensitivity of @xmath7 3 10@xmath11 ph @xmath12s@xmath13 ( 3@xmath14 , 1 ms observing time .
spi s energy resolution is typically e/@xmath15e=500 or 3 kev at 13001800 kev , due to high - purity n - type ge semiconductor detectors operated at @xmath7 80 k and @xmath7 4 kv , with annealings maintaining this spectral performance .
this allows spectroscopy of nuclear lines through unambiguous identification of line energies .
other current facilities in the field of cosmic gamma - ray spectroscopy include the rhessi solar observatory @xcite , the gamma - ray burst monitor scintillation detectors ( modest energy resolution ( 10% fwhm ) ) aboard the _ fermi _ gamma - ray satellite launched in june 2008 @xcite , and the nustar mission @xcite launched in june 2012 and equipped with a focusing hard x - ray mirror and detectors operating between 6 and 79 kev , thus including lines from radioactive decay of @xmath1ti at 68 and 78 kev , the lowest - energy lines known from any cosmic radioactive decay .
detectors for gamma - ray spectrometry in the 0.1 mev to gev range could be based on scintillators , solid - state detectors , drift or time projection chambers , and trackers made up of spark chambers or solid state detector stacks .
the physical challenge is to produce a cascade of inelastic interactions of the primary photon and its secondaries within the volume of the detector , and to produce an electrical signal which is proportional to the total energy deposit of the cascade .
high levels of background radiation lead to detectors which respond quickly , and have short dead times . here , scintillation detectors have an advantage .
the issue with scintillation detectors is to ensure a homogeneous and linear light collection over the volume of the scintillation detector .
imperfections result in different signal amplitudes per energy deposit , depending on the location of interactions within the detector volume .
the spectral resolution required for identification of gamma - ray lines and relating them to specific nuclear transitions practically can only be achieved through solid state detectors .
solid state detectors operate through collection of the charge liberated from photon interactions as electrons are activated into the conduction band . in semiconductor detectors ,
a small band gap of few ev only allows very sensitive high - resolution detectors .
germanium detectors have been established as standard in terrestrial nuclear - physics experiments , and also space borne cosmic gamma - ray experiments ; ge detectors have been reviewed recently @xcite . in recent years , cdzntl detectors have become popular , because they can be operated at room temperature , rather than the cryogenic temperatures required for ge detectors , at nearly similar performance .
charge collection implies propagation of electrons and holes within the detector volume , which is achieved through an electric field between the two electrodes of the detector .
depending on detector and electrode geometry , charge collection also may be inhomogeneous across the detector volume , thus limiting spectral resolution .
moreover , bombardment with cosmic rays in the typical space environment where such detectors for cosmic photons are operated will destroy the crystal properties locally within the detector volume , thus producing traps for charge transport and therefore impeding charge collection with cumulative cosmic - ray exposure .
figure [ fig : spi_annealings ] demonstrates this effect for the spi detectors on integral , showing the resolution for an instrumental line at 1764 kev as a function of time ( in units of satellite orbits called revolutions which last 3 days each ) over ten years .
the degradations are repaired through annealings every few months : here , the entire detector platform is heated from cryogenic operational temperatures around 80 k to about 370k for 10 - 14 days , thus curing the crystal defects which impeded charge collection .
the spectral response of the ge detectors on spi is monitored and stored in precise detail , from many instrumental lines fitted by a parametrized representation of the response to photons of specific energy .
the response function is composed of a gaussian centered at the incoming photon energy , a step function from this energy towards lower energies , and a one - sided exponential from the incoming photon energy downward to represent the degrading charge collection efficiency .
figure [ fig : spi_degradation ] shows how the degradation thus is extracted in consistent and precise detail . from a set of instrumental - background lines , the spectral response is decomposed into a nearly - gaussian part plus a one - sided exponential that describes degradation of charge collection
this response is then used in spectroscopic analysis later on .
spectroscopy at mev energies suffers from a dominance of instrumental background in measured events , the cosmic photons of interest contribute only a small ratio of events at the percentage level or below .
typical source intensities in lines are 10@xmath16 ph @xmath12s@xmath13 for the strongest line at 511 kev to 10@xmath17 ph @xmath12s@xmath13 for the weakest lines detected so far ( from @xmath3fe decay ) , and many predicted lines from objects of interest are about 1 - 2 orders of magnitude fainter still .
thus , a critical component of cosmic gamma - ray spectroscopy is the determination of the instrumental background as it may vary across the months of typical exposure times needed to collect the cosmic photons from sources of interest .
the physical processes which cause the instrumental background are rather clear : cosmic ray and secondary energetic - particle interactions with the spacecraft and instrument materials causing inelastic high - energy photon production and nuclear excitations .
nevertheless , the prediction of instrumental backgrounds from software implementations of such physical processes ( e.g. geant ) fail to achieve the required precision of better than a % -range precision , with their typical uncertainties of 30% .
these uncertainties derive from unknown energetic - particle spectra and flux variations , and from limitations in modeling the spacecraft and instrument in its geometry and materials composition at sufficient precision ; small admixtures of specific isotopes are often responsible for the strongest background lines .
moreover , those energetic particle interactions also create radioactive isotopes within the spacecraft which are not part of the materials library of these software packages , as they have been established from terrestrial high - energy physics experiments .
therefore , the method of choice for background determination is empirical fitting of models which describe the background in its spectral and temporal variations . the main challenge is to find a background representation which is orthogonal , or rather least degenerate , with variations of the cosmic photon signal .
figure [ fig : spi_eventrate ] shows the count rate variation over ten years of the spi detectors on integral .
the general increase of background over the first years corresponds to a decline of solar activity , as energetic particles from the sun ( solar wind ) change the morphology of the earth s magnetosphere . in near - earth orbits of space satellites , therefore in phases of high solar activity the shielding of cosmic rays is more effective .
this can be seen in fig .
[ fig : spi_eventrate ] from background declining again as the solar activity picks up in cycle # 24 after the year 2010 ( see fig .
[ fig : sunspots ] ) . superimposed
are short - term rate variations , which arise from solar flare events , or magnetic storm particle events which expose the satellite to an increased particle flux for a short time .
additionally , the approach of radiation belts at the perigee of the satellite orbit leads to strongly - increased energetic particle irradiations ; but here the instrument is turned off to avoid damages from saturations , so that only a declining activation signature from this perigee exposure can be seen after the instrument is turned on again leaving the radiation belt zones .
the integral orbit has been chosen to have a high eccentricity with an apogee of about 150,000 km , so that the few minutes of post - perigee activation decline are rather negligible and background generally is stable across days of this 3-day orbit .
for modeling of instrumental backgrounds , it had turned out to be most effective to make use of other , independent , detectors of energetic - particle - flux related events near the cosmic photon detectors . in the case of spi ,
the rate of saturated events in ge detectors has proven valuable , as has the rate of charged - particle detections in the anticoincidence shield detector ( based on bgo scintillator material ) and a veto detector plate below the coded mask ( which is based on a plastic scintillator ) @xcite .
moreover , ge detector counts in continuum energy bands adjacent to the line of interest proved valuable . but
all of these _ tracers _ of instrumental background have been found to change their functional relation to background on time scales of days , thus require scaling parameters to be fitted during the cosmic source exposures .
therefore , a typical spectroscopy study with spi consists of fitting a multi - parameter model to observed spectral counts per the ge detector set and thousands of 30-minute pointings of overlapping sky regions .
the model includes instrumental background through a few tracer components adjusted at different time intervals , and a spatial - distribution model of the cosmic photon sources of interest , which is folded through the instrument s imaging response to predict the cosmic event contribution .
once the intensity scaling of the cosmic source is determined in fine energy bins across the line of interest , the spectral response is then accounted for to determine the line position and shape parameters for the cosmic source .
an illustration of this procedure can be found below , from the detailed spectroscopy of the celestial @xmath2alline .
understanding the progenitor systems and the dynamics of the carbon - burning triggered explosion of supernovae of type ia is motivated by two main reasons .
snia are the dominating sources of cosmic iron and similar elements , and therefore key to the metal enrichment of the universe .
snia also are key to current cosmology studies , being used as standardizable sources with known intrinsic brightness for tracing cosmological expansion @xcite .
but up to now , no consistent physical explanation of snia explosions has been established , in spite of many proposed scenarios and progenitors .
currently - discussed models are ( i ) explosions triggered in white dwarf central regions as the chandrasekhar mass limit is reached , ( ii ) explosions triggered by an off - center event such as a he - shell flash on white dwarfs of a broader mass range , and ( iii ) explosions resulting from merging of two white dwarfs as it may terminate the evolution of a binary system .
more complex scenarios of binary evolution with recurrent pulsations culminating in a supernova are also discussed @xcite . in all cases
, the runaway nuclear carbon burning will provide the energy which disrupts the white dwarf ; this is the commonly agreed part of snia models ( see@xcite for recent reviews ) .
the issues discussed relate to the nuclear c - burning flame propagation in the different density regimes , between the core of the white dwarf and the exploding outer layers . as a result
, the snia nuclear ashes may range from almost - pure ( near 1 m@xmath8 ) @xmath0ni as it would result from complete burning in nuclear statistical equilibrium ( nse ) , through deflagration with nse and incomplete - nse burning , down to @xmath0ni amounts below 0.1 m@xmath8and mostly intermediate - mass element ashes resulting from partial burning , such as expected for double - degenerate merger or detonation models @xcite .
correspondingly , the spatial distribution of @xmath0ni ashes may range from central - only to distribution throughout the entire exploding supernova .
the initial brightness of snia is due to the sudden deposition of the explosion energy of @xmath7 10@xmath18erg .
but soon , after a few days , supernova brightness dominated by the energy deposition of gamma - rays and positrons from radioactive decay of @xmath0ni .
clearly , the brightness critically depends on how @xmath0ni deposits its radioactive energy into the gas of the rapidly - expanding envelope , leading to thermal emission of the envelope gas at its ( inward - moving ) photosphere .
a ( bolometric , thermal - emission ) luminosity maximum is obtained at @xmath7 20 days after explosion in optical emission .
radiation propagation models are used to relate this low - energy photon emission to the original @xmath0ni and its radioactive energy deposition ( @xmath73 mev per decay of a @xmath0ni nucleus ) @xcite .
measurements of the total @xmath0ni content through gamma - rays are more direct , and rather independent ( at times > 100 days ) of the density structure of the expanding supernova .
the expanding supernova becomes transparent to gamma - rays on a time scale of @xmath7 100 days , so that @xmath0ni decay energy deposition fades .
the rise of radioactive gamma - ray luminosity , as well as gamma - ray line shapes and line - to - continuum ratio convey information about the large - scale spatial distribution of @xmath0ni and the envelope structure .
typically , the gamma - ray luminosity maximum is expected at 7090 days after explosion , much later than the luminosity maximum in thermal ( optical to infrared emission ) supernova light .
the apparent optical luminosity fading with a @xmath7 100-day time constant thus is fortuitous , and results from radioactive decay of @xmath0ni through @xmath0co to @xmath0fe convolved with energy deposition properties of the rapidly - expanding supernova envelope .
sufficiently - nearby events , closer than about 5 mpc , should occur once every 23 years ( as estimated by @xcite from different sn rate proxies in the local universe ) , although numbers vary by factors @xmath7 2 depending on the assumptions about the supernova rate in this more - heterogeneous population of nearby galaxies . the cgro mission in 19912000 had two opportunities : sn1991 t occurred early in the mission , at a distance of 13 mpc the detected hint for the characteristic gamma - ray lines appeared surprising .
but it was quickly recognized that sn1991 t was an anomalously - bright event also from other observed phenomena , and hence not typical for snia .
the second event during the cgro mission was sn1998bu at @xmath7 17 mpc .
it was much fainter , and no gamma - ray emission could be detected . during the integral mission ,
two snia were close enough for useful integral observations , but no signal could be found .
sn2003gs probably was much too distant , at 16 mpc .
sn2011fe occurred on august 24 , 2010 , in the nearby galaxy m101/ngc5457 @xcite , at a distance of 6.4 mpc . with lines likely to be significantly broadened ,
the high spectral resolution does not help , considering the instrumental - background level .
integral observed sn2011fe for 4 ms total , only to provide upper limits @xcite ( see fig .
[ fig_sn2011fe ] ) .
thus , after the comptel and integral gamma - ray telescopes operating for 20 years , snia gamma - rays still have not contributed to help understand snia .
a significant advance in sensitivity is needed to not depend only on luck with sufficiently nearby events , to below 10@xmath17 ph @xmath12s@xmath13 as shown in several proposed mission concepts ( see esa s cosmic - vision program selection , e.g. @xcite ) .
massive stars with initial masses above 810 m@xmath8evolve through a series of central nuclear fusion stages beyond hydrogen burning up to silicon burning .
as this energy reservoir of the release of nuclear binding energy is exhausted eventually , gravitational collapse will terminate this evolution after 10100 my @xcite .
the collapse occurs on a time scale of seconds , once either central nuclear burning starves from exhaustion of the fuel for energy - liberating fusion reactions or electron capture sets in and takes away the pressure support from the ( degenerated ) electron gas .
infalling nuclei are decomposed in the shock wave above the newly - forming neutron star into nucleons and @xmath19 particles , which consumes part of the gravitational energy , but neutrinos emitted by the neutron star that forms in the interior of the collapsing star probably lead to sufficient energy deposits in the infalling matter to trigger an explosion .
how exactly this occurs is a matter of current studies and debate .
it appears to be the result of instabilities in infalling and expanding gas flows within the delicate balance of gravitational , nuclear reaction , neutrino interaction , and hydrodynamical flow energies .
clumps and jets may be part of such explosions , as observed in several core - collapse supernova remnants .
nuclear reactions are expected to occur in the dense shock region approaching nuclear statistical equilibrium ( nse ) , hence producing fe - group elements and also substantial amounts of radioactive @xmath0ni . at the same time , in regions nearer to the neutron star , neutrino interactions with nuclei occur in the _ @xmath20-process _ and mainly liberate nucleons from nuclei , thus stimulating proton and neutron capture reactions of remaining nuclei . further out ,
decomposition of in falling matter will also provide abundant free nucleons and @xmath19-particles , and thus add an _
@xmath19-rich _ flavor to nuclear burning .
once the explosion sets in , material will expand and cool , and nuclear burning will _ freeze out_. yet , the explosion will drive a shock through the outer envelope beyond the core , before these parts could take notice of the inner collapse , and some _ explosive nucleosynthesis _ will occur in those shock - heated regions .
this explosive nucleosynthesis is characterized by short nuclear burning times which lead to large deviations from equilibrium and hydrostatic nuclear burning patterns .
altogether , one expects that abundant oxygen , silicon , and iron - group nuclei will be ejected , mixed with small but important contributions of heavy element products from the neutrino - driven wind zone .
radioactive ejecta which could be observed with gamma - ray spectrometers include @xmath0ni and @xmath1ti , at estimated amounts of @xmath7 0.1 and @xmath7 few 10@xmath21 m@xmath8 , respectively .
ejection of @xmath0ni is certified by the supernova light resulting from its decay energy , but amounts vary by almost three orders of magnitude ) .
again , proximity is key to gamma - ray observations , and more constraining due to the much smaller amounts of @xmath0ni compared to snia .
@xmath1ti ejection is much less clear , as it is produced predominantly in the supernova s interior region very close to the mass cut separating ejecta from material which will end up in the central neutron star .
one - dimensional models of type ii or type ib supernovae obtain typical @xmath1ti yields of 5 10@xmath11 m@xmath8 , with large ( factors 24 ) variations with progenitor mass @xcite . analyzing the impacts of deviations from spherical symmetry and the entropy - dependency of @xmath19-rich freeze - out nucleosynthesis , an overall enhancement of @xmath1ti relative to @xmath0ni of a factor of @xmath7 5
was considered plausible @xcite ; but it remains to be shown that this holds for a realistic and detailed supernova model .
three - dimensional simulations of a nucleosynthesis network under a variety of plausible supernova - interior conditions @xcite obtain mass fraction variations over 68 orders of magnitude for @xmath1ti production , and thus reinforce that the specific supernova explosion trajectory in the @xmath22 phase space will determine the ejected amount of @xmath1ti .
measurements of @xmath1ti gamma - rays from all accessible young supernova remnants can clarify nucleosynthesis and thus the physical conditions in these inner supernova regions .
@xmath1ti decay results in gamma - ray emission in three lines : the first stage of its decay chain ( half - life of 59 y ( @xmath230.3 y ) ) occurs predominantly ( 99.3% ) through capture of an electron into @xmath1sc , which de - excites by emission of the excitation - level energy through two x - rays at 68 and 78 kev . @xmath1sc itself
@xmath24-decays quickly ( with half life 3.7 h ) to excited @xmath1ca , which in almost all cases de - excites through emission of the characteristic 1157 kev @xmath25-ray line .
the two low - energy lines are within the range of high - energy x - ray telescopes such as integral s coded - mask telescopes , and the nustar x - ray mirror telescope , while the gamma - ray line requires gamma - ray instruments such as comptel aboard cgro and spi on integral . ti decay . _
( left : ) _ the @xmath1ti decay line at 1157 kev was discovered with comptel .
the detection was based mainly on imaging , detecting cas a as a point source in the energy band of this line . _ ( right : ) _ integral / spi did not clearly detect the line ; the continuous line represents a narrow line ( instrumental width ) with a flux of 1.6 10@xmath11ph @xmath12s@xmath13 , within uncertainty ( @xmath231.2 ) consistent with the consolidated flux of 2.5 10@xmath11ph @xmath12s@xmath13 . if we assume that the data do not show any line at the expected energy of 1157 kev , the cas a line must be broadened by at least 500 km s@xmath13 for statistical consistency ( 2@xmath14 ) .,title="fig:",scaledwidth=53.0%]ti decay . _
( left : ) _ the @xmath1ti decay line at 1157 kev was discovered with comptel .
the detection was based mainly on imaging , detecting cas a as a point source in the energy band of this line . _ ( right : ) _ integral / spi did not clearly detect the line ; the continuous line represents a narrow line ( instrumental width ) with a flux of 1.6 10@xmath11ph @xmath12s@xmath13 , within uncertainty ( @xmath231.2 ) consistent with the consolidated flux of 2.5 10@xmath11ph @xmath12s@xmath13 . if we assume that the data do not show any line at the expected energy of 1157 kev , the cas a line must be broadened by at least 500 km s@xmath13 for statistical consistency ( 2@xmath14 ) .,title="fig:",scaledwidth=45.0% ] the cas a supernova has become a key test case for core - collapse models .
the remnant is at a distance of 3.4 kpc ( + 0.3/-0.1 kpc ) @xcite , and its 5 diameter is suitable for observations also with modern telescopes throughout the electromagnetic spectrum , considering their often small fields of view .
its explosion is dated to year 1671 with a precision of @xmath7 few years ( see @xcite ) . at present ,
the outer shock of the explosion blast wave is clearly seen in radio and x - ray bands , while the reverse shock that is expected can not be clearly identified , but supposedly now has propagated through @xmath72/3 of the remnant already , re - heating the remnant gas to x - ray emission temperatures .
although the age of cas a is well beyond the 89 y decay lifetime of @xmath1ti , @xmath1ti radioactivity observations have become an important complement in the study of the cas a supernova explosion .
comptel observations obtained discovery of @xmath1ti from cas a ( fig . [ fig : casa - spec ] ) , and stimulated several followup measurements .
both integral main telescopes are sensitive to @xmath1ti gamma - rays : the ibis telescope obtained a flux value of ( [email protected] ) 10@xmath11ph @xmath12s@xmath13 @xcite from the 68 and 78 kev lines . from the spi spectrometer , an initial report of a detection consistent with expectations @xcite
was corrected later towards an upper flux limit , owed to large variations of background during observations .
this has been translated into an amount of @xmath1ti at the time of explosion of 1.6 x 10@xmath21 m@xmath8 , with an uncertainty of ( @xmath26/@xmath27 ) m@xmath8 .
thus the @xmath1ti ejection from the cas a supernova was @xmath7a factor of three above predictions from models .
the non - detection of the 1157 kev decay line by integral s spi instrument ( fig .
[ fig : casa - spec ] ) suggests that @xmath1ti - rich ejecta move at velocities above 500 km s@xmath13 . as doppler broadening increases with photon energy
, this makes the high - energy decay line much broader than spi s instrumental resolution , hence the signal - to - background ratio for this line degrades and may escape detection , less so the 68 and 78 kev lines which have been measured consistently from cas a. the nustar imaging telescope has recently mapped the @xmath1ti emission morphology for cas a ( harrison , _ priv.comm . , _ 2013 ) .
this provides insights into supernova asymmetries , and revealed a rather extended , though incomplete , symmetry rather than major clumping of inner ejecta through @xmath1ti , which have been seen from other ejecta traces in the outer parts of the cas a supernova remnant @xcite .
supernova sn1987a in the lmc galaxy was _ the _ event which had major influence on our current understanding of core - collapse supernovae .
the detection of @xmath0co decay lines with the gamma - ray spectrometer scintillation detector instrument on the solar maximum mission @xcite was a surprise , as the supernova envelope had been expected to become transparent to inner nucleosynthesis products much later .
this suggested major deviations from sphericity of the supernova explosion .
integral observations in 2010/2011 provided a detection of sn1987a in just an energy band characteristic for the 68 and 78 kev lines from @xmath1ti decay @xcite .
as with cas a , the reported flux is considerably above expectations of @xmath1ti synthesis from models , again pointing to non - spherical explosion models . considering the rate of core collapse supernovae in our galaxy of @xmath7 2/century @xcite , one would expect to detect @xmath1ti emission from several objects along the plane of the galaxy , if @xmath1ti ejection were a typical property of core - collapse supernovae .
both the comptel and integral suverys fail to detect the expected numbers of @xmath1ti sources . in spite of the low number of events involved ,
this is a significant observation .
thus , also from the rate of @xmath1ti emitting core - collapse supernovae , it seems that object to object variations must be large , and @xmath1ti ejection non - typical , probably much larger than current models predict in case they do happen . in
non - spherical explosions with clumps and jets , @xmath1ti ejection could vary by more than an order of magnitude , as explorations of the parameter space in density and temperature and their impact on nuclear reaction networks have shown @xcite .
radioactive nuclei ejected from cosmic sources of nucleosynthesis into interstellar space will accumulate and produce a diffuse glow , when their decay time is long compared to the time between ejection events .
this is the case for @xmath2al and @xmath3fe from massive stars and their supernovae , given the radioactive decay times of 0.7 and 2.6 my , respectively , and massive - star evolution times of my with a core - collapse supernovae rate of @xmath7 2 per century .
unlike in the case of the supernovae discussed above , the measured gamma - rays therefore can not be attributed to a single object , and rather are due to the nucleosynthesis of a population or group of stars ( see ) . ) , showing continuum emission and the additional line feature from positron annihilation , as well as the lines at 511 kev , and from @xmath2al , and @xmath3fe radioactive decay ( adapted from @xcite ) .
the spi data points are shown as crosses , ibis results are shown with square symbols , and other lines ( solid , dashed ) show different components as expected to contribute to galactic continuum emission ( for details discussing the continuum components of bremstrahlung , inverse - compton , and superimposed stellar sources see @xcite and @xcite ) . ,
scaledwidth=68.0% ] al _
( left ) _ , and @xmath3fe _ ( right ) _
.,title="fig:",scaledwidth=48.0% ] al _ ( left ) _ , and @xmath3fe _
( right ) _
.,title="fig:",scaledwidth=48.0% ] the diffuse and extended gamma - ray emission of the galaxy is dominated by a bright , power - law - type continuum , which is due to interactions of cosmic ray electrons with interstellar radiation and matter ( figure [ fig : galridge_spec ] , @xcite .
superimposed are also seen a few line features , from positron annihilation ( together with its lower - end continuum ) at 511 kev , and from @xmath2alat 1808.63 kev and @xmath3fe at 1773 and 1332 kev .
these line emissions appear extended and diffuse in the galaxy ; this has been established for the bright @xmath2alemission with sufficient signal for imaging , and inferred indirectly for @xmath3fe ; the diffuse glow of positron annihilation emission , which should go hand in hand with nucleosynthesis of isotopes on the proton - rich side of the stability valley of isotopes , such as @xmath2al or @xmath1ti or @xmath0ni , has not yet been detected unambiguously from the galaxy as a whole , and also not from the prominent @xmath2al - bright massive - star regions , but rather is found concentrated in a bright and extended emission region associated with the galaxy s bulge region .
it is these lines which will be discussed in more detail in the following .
the @xmath2alline from the interstellar medium in the galaxy was the first nuclear line detected from outside the solar system @xcite .
it had been attributed to current nucleosynthesis in our galaxy , although the source remained unclear for a while between novae , wolf - rayet stars , supernovae , and even solar system origins .
the comptel gamma - ray survey then mapped @xmath2alemission across the sky , and found rather clumpy emission extended along the entire plane of the galaxy .
this was interpreted as massive stars being the dominant sources , evolving in groups and leading to local emission peaks when stars reach wolf - rayet phases and first core - collapse events , while novae were considered less important because their more - frequent ejection events should produce a smoother and more centralized glow of the galactic plane @xcite .
spi confirmed the interstellar glow throughout the galaxy of @xmath2al gamma - rays early in the integral mission , and has been refining those measurements since .
figure [ fig_al - spectra_years ] shows the @xmath2alsignal measured with integral / spi from the entire galaxy , adopting spatial distribution on the sky according to the skymap based on the comptel measurements @xcite . shown
are results from different datasets over the years , the latest result using also multiple - detector events and data up to orbit number 1142 ( february 2012 ) .
the @xmath2alline parameters of intensity , line position , and width , are all consistent with earlier results based on single - detector events only .
the statistical precision increases from the larger number of events , although some additional uncertainty from background modeling arises .
nevertheless , the @xmath2alemission is seen at 32 @xmath14 significance .
this signal from @xmath2al has been studied in detail now from the galaxy at large , see fig .
[ fig : al - fe_spectraspi ] , and in specific , localized regions ( see below ) .
in the inner galactic ridge , a flux at 1808.63 kev of ( 2.63 @xmath230.2 ) 10@xmath21ph @xmath12s@xmath13rad@xmath13 has been derived .
alline measurements with spi during the integral mission.,scaledwidth=100.0% ] the total mass of @xmath2al in the galaxy had been inferred from earlier comptel measurements to be 23 m@xmath8@xcite .
in such a mass determination , one must adopt a spatial source distribution to resolve the distance uncertainty , when converting a measured gamma - ray flux into a quantity of isotopes present in the galaxy .
generally , large - scale models of galactic structure or sources have been used , such as exponential - profile disk , azimuthally - symmetric nested rings , or spiral - arm models inferred from other observations . yet
, the irregularity of the large - scale emission as seen by comptel from galactic @xmath2al already suggested that the massive - star population in the galaxy may be more clumpy than such large - scale models describe .
the localized @xmath2al emission seen from the cygnus and sco - cen regions supports this view .
accounting for such localized enhancements , with more - advanced integral measurements the galactic mass of @xmath2al has been re - determined as ranging from 1.5 to 3.6 m@xmath8 , considering all uncertainties . using this galaxy - wide amount of @xmath2al together with @xmath2al yields for massive stars across the entire mass range , and an initial - mass distribution
, one can derive the total population of stars which corresponds to measured @xmath2al gamma - rays @xcite .
as massive stars above @xmath7 8 - 10 m@xmath8all are believed to end their evolution as core - collapse supernovae , this corresponds to a determination of the _ galactic _ rate of core - collapse supernovae .
we obtain a value of 1.54 ( @xmath230.89 ) supernovae per century , or a supernova every 65 years ( somewhat lower than the value published in @xcite from @xmath2al , due to accounting for foreground emission now attributed to the nearby scorpius - centaurus sources , rather than distant galactic @xmath2al . ) .
the importance of this measurement of the galactic supernova rate derives from the underlying method . using penetrating gamma - rays from radioactivities ejected by supernova related sources throughout the current galaxy ,
this approach does not suffer from corrections for occulted sources , such as e.g. a supernova rate determination based on o and b star counts does .
other measurements inherently assume negligible possible differences between our galaxy and other galaxies , as they measure core - collapse supernova related objects in milky - way - like galaxies which are seen face - on and hence free from occultation - bias of distant parts of our own galaxy .
it is interesting that the @xmath2al - based supernova rate determination agrees with measurements undertaken with those alternative approaches , and falls into the lower - value part among all those measurements ( as discussed in @xcite ) . with fine ge - detector spectroscopy , the shape of the @xmath2alis resolved , and
may thus be studied in terms of astrophysical signatures .
figure [ fig_al - width_2012 ] shows the probability distribution of the line width for an additional gaussian line broadening beyond the instrumental resolution alone .
an astrophysical line broadening was discovered , with a value of 1.4 kev ( @xmath230.3 kev ) , which corresponds to 175 km s@xmath13 ( @xmath2345 km s@xmath13 ) in velocity space . for comparison ,
the expectations from averaging over different velocities from large - scale rotation of the galaxy are about 1.35 kev , and compatible with this measurement .
alline measurements with spi along the plane or the galaxy , for different lines of sight .
the systematic doppler shift from large - scale rotation in the galaxy can be seen clearly ( _ left _ ) . a nearby instrumental line at 1764 kev ( _ right _ )
does not show differences in line position , from the same data.,scaledwidth=90.0% ] the large - scale galactic rotation shows up as a line centroid shift in the measurements of @xmath2al : bulk motion differs for different lines of sight along the plane of the galaxy , with systematic variations leading to a blue shift of the @xmath2alline towards the 4@xmath28 quadrant of the galaxy , and a red shift towards the 1@xmath29 quadrant , correspondingly .
the doppler shift signature was indicated in early results , and is now consolidated ( see fig.[fig_al - spectra_plane ] ) .
this allows to study motion of hot interstellar gas as it is shaped around massive - star regions , and compare its dynamics to dynamics of stars and gas as we know it otherwise .
for example , co line emission has been used to trace the molecular gas in the galaxy @xcite . here , the galactic ridge was recognized clearly , with peculiar and high cloud dynamics in the vicinity of the galaxy s central supermassive black hole . as star formation occurs from dense clumps in molecular clouds , @xmath2alejected from recently - formed stars could show how closely its kinematic motion can be related to the current molecular cloud population .
al line doppler - shift signature along the inner galaxy , as plausible from large - scale galactic rotation , and compared to co data @xcite ( color ) .
the data points are derived from independent longitude segments and show the velocity of @xmath2al - enriched interstellar gas ; the blue - hatched area shows the range of uncertainty in these values from imaging resolution .
the solid blue line shows expectations from @xmath2alsources moving with molecular clouds as seen in co data , the red - dashed line shows how locations of @xmath2alsources in the galaxy s bar would affect the kinematic signature .
the dashed line matching the @xmath2aldata points shows expectations from a model based on @xmath2albeing ejected from the leading edges of inner galactic spiral arms , with a net bulk velocity of @xmath7 200 km s@xmath13 due to asymmetries of the superbubbles created by massive star groups ; see kretschmer et al
. , 2013 , submitted , for more detail.,scaledwidth=78.0% ] the longitude - velocity diagram as derived from integral @xmath2almeasurements is shown in fig .
[ fig:26al_long - vel ] .
the derived velocities extend from @xmath7 -200 to @xmath7 + 200 km s@xmath13 at the extremes ( tangential directions to galactic - ridge directions with largest relative motion with respect to the sun , at longitudes @xmath2330@xmath9 .
it is apparent that the @xmath2alvelocities extend over higher velocities on large scale than known for other sources and in particular for cold , molecular gas ( also shown in fig . [ fig:26al_long - vel ] in color ) .
this appears surprising .
if we consider @xmath2alsources across the galaxy traced by free electrons produced by ionizing starlight , which has been mapped through pulsar dispersion measurements , we would expect a longitude - velocity trend as shown by the thick ( blue ) line in fig .
[ fig:26al_long - vel ] .
but this representation of free electrons according to @xcite is deficient in the inner galaxy , from an observational bias of available pulsar data .
thus it may not be surprising that this apparently does not represent @xmath2alobservations of motions , providing a hint that some inner - galaxy sources may be missing in this free - electron model .
adding an additional @xmath2alsource population also within a galactocentric radius of 2 kpc ( where the free - electron map is @xmath7 empty ) , and applying a density structure according to the galaxy s inner bar @xcite plus rotational behavior as extrapolated towards the inner galaxy from available measurements @xcite , the @xmath2alkinematic data are more closely reproduced ( kretschmer et al . ,
submitted ) .
the generally - larger velocities seen in @xmath2alsources suggest that preference towards the direction of galactic rotation may be given to massive - star ejecta as they leave their sources .
it remains to be shown if more detailed modeling of star formation along the galaxy s bar and inner spiral arms can provide an explanation of the observed @xmath2alvelocities .
surprisingly , @xmath2alenriched gas appears to show systematically larger velocities than expected .
this suggests that the ejection of nucleosynthesis material from massive - star sources occurs into surrounding interstellar medium which has been shaped by massive - star winds and the gas accumulation in spiral arms in peculiar ways ( kretschmer et al .
2013 , in press ) .
long searched for , gamma - rays from decay of @xmath3fe were discovered from the plane of the galaxy in 2005 by rhessi @xcite and integral , see fig .
[ fig : al - fe_spectraspi ] .
this signal was found to be much fainter than the one from @xmath2al , although the same massive - star populations are the most - plausible sources for both those relatively long - lived isotopes which decay long after they have been injected into interstellar space from their sources .
the @xmath2alyields of massive stars used in above considerations are based on models of stellar evolution and supernovae ( see @xcite for a summary and references ) .
[ fig : al - fe_nucleosynthesis ] shows the complex interior structure of massive stars , as they evolve from initial core hydrogen burning towards the final core - collapse supernova .
nuclear - burning and convective regions are indicated ( hatched ) . within the star , at later phases ,
intermittent shell burning occurs simultaneously and in addition to core nuclear burning of the different fuels , he , c , o and si burning .
the locations of @xmath2al ( and also @xmath3fe ) synthesis are indicated ( arrows ) : @xmath2al is produced in the hydrogen - burning phases from initial @xmath30 mg , and later in the carbon and neon burning core and shell as proton - release reactions allow further processing of remaining @xmath30 mg , plus during the supernova explosion as explosive ne / c burning . only @xmath2al from core hydrogen burning
is mixed into the envelope and thus ejected during the intense stellar - wind phase .
@xmath3fe is expected to be synthesized from successive neutron capture reactions on @xmath31fe , as neutron release reactions such as @xmath32c(@xmath19,n ) and @xmath5ne(@xmath19,n ) are activated ; but here , a subtle balance of neutron capture and convective transport away from neutron - rich regions is required to ensure that @xmath3fe is not destroyed by further neutron captures .
reaction rates for these neutron captures on unstable and n - rich fe isotopes are all being re - assessed by current and planned laboratory experiments ( see @xcite for a discussion ) .
all @xmath2al and @xmath3fe produced within the star after the main sequence phase is released only with the supernova explosion , together with additional contributions from explosive burning in the supernova itself .
note that stellar evolution uncertainties also arise from nuclear reaction rate uncertainties of the 3@xmath19 and @xmath33c(@xmath34o reactions , and in particular affect the yields of @xmath3fe @xcite . ) .
the production sites of @xmath2al and @xmath3fe are indicated , as well as the mass loss from stellar wind in the wolf - rayet phase .
, scaledwidth=68.0% ] if we consider the likely formation of a _ group _ of massive stars from a dense molecular - cloud core , their concerted release of @xmath2al and @xmath3fe radioactive isotopes occurs over a period of @xmath7 3 to 20 my after star formation .
the measurement of @xmath2al and @xmath3fe radioactivity gamma - rays , therefore , is a tool to verify our _ stellar - mass averaged _ predicted yields from models of massive - star evolution and nucleosynthesis . in particular , since the same massive stars are plausible producers of both those isotopes , yet from different regions and epochs inside the stars , the measurement of the ratio of @xmath2al to @xmath3fe emission provides a valuable tool , as systematic uncertainties in the stellar populations themselves ( richness , distance ) cancel in such ratio .
the @xmath3fe gamma - ray line emission occurs in a cascade of two gamma - ray lines , at 1172.9 and 1332.5 kev , with approximately - equal intensities ( 99.85% of decays produce the 1172.9 kev line , versus 99.98% for the 1332.5 kev line ; a 59 kev line is emitted in 2% of decays from a transition in @xmath3co ) . in none of these lines , separately analyzed in spi
single - detector and multiple - detector events , a clearly - convincing celestial gamma - ray line signal could be seen . combining line intensities in both these lines , the spectrum shown in fig .
[ fig : al - fe_spectraspi ] has been derived , which shows @xmath3fe emission from the sky with a combined significance of @xmath7 5@xmath14 . in any case , the total gamma - ray brightness in @xmath3fe decay is substantially below the one seen from @xmath2al .
the measured gamma - ray brightness ratio is @xmath7 15% , with an uncertainty of @xmath7 5% . for @xmath3fe
, no imaging decomposition could be derived yet due to this low brightness . in steady state , this brightness ratio thus constrains massive - star interiors globally for an average over all massive - star groups throughout the galaxy . for stellar groups at specific ages , the steady - state assumption does not apply , however , and in particular the re - assessed decay time of @xmath3fe of 3.8 my @xcite ( the formerly - used value was 2.2 my ) must be taken into account for determinations of the isotope ratio ( see discussion of _ specific regions _ below ) . from nucleosynthesis models of massive stars and supernovae
, it appears that more @xmath3fe would be expected , although currently - predicted yields are still in agreement with the observations , within quoted uncertainties .
note that @xmath3fe nucleosynthesis may also occur in a rare subtype of snia explosions at high yields of m@xmath8@xcite , although a single bright source being responsible for observed @xmath3fe gamma - rays appears unlikely ( the rhessi and spi detections were derived assuming extended emissions along the galactic plane ) .
@xmath3fe yields are particularly high ( as compared to @xmath2al yields ) in massive - star models for the upper mass ranges above @xmath7 60 m@xmath8@xcite .
so , even considering the uncertainties of the gamma - ray measurement , the @xmath3fe/@xmath2al ratio is lower than expected .
possible causes could be suppression of @xmath3fe by more - efficient neutron capture reactions , leading to neutron - richer fe isotopes .
alternatively , @xmath24-decay of @xmath35fe could occur at a higher rate than inferred from current nuclear theory .
both reaction types are being investigated in laboratory experiments at the present time .
note that the @xmath24-decay lifetime of the @xmath3fe isotope had been re - determined in 2008 @xcite , and found to be substantially longer , with ( exponential ) lifetime increased from 2.15 to 3.8 my ( t@xmath36 from 1.5 to 2.6 my ) .
another cause of lower galaxy - wide @xmath3fe glow could be that the explosion as a supernova does not occur throughout the entire range of stellar masses , but deviations from the standard initial - mass distribution on the high - mass end of massive stars @xmath3760
m@xmath8 , or islands of explodability in that mass range occur .
the observational bound of @xmath7 15% for the @xmath3fe/@xmath2al gamma - ray intensity ratio provides a significant constraint to be met by massive - star models as a whole , including aspects of stellar structure , nucleosynthesis , and evolution towards supernovae .
particularly interesting would be the @xmath3fe/@xmath2al ratio for source populations of specific ages , as the ratio varies significantly due to wind - released @xmath2al before any core - collapse supernova would eject @xmath3fe and more @xmath2al .
@xmath3fe is exclusively released in supernovae , although predominantly produced in the late shell - burning phase before the collapse of the core .
only the cygnus region appears within integral s sensitivity range for this , however .
algamma - ray line ( e@xmath38 1808.63 kev ) as measured by spi from different regions along the plane of the galaxy for @xmath2al .
, scaledwidth=98.0% ] @xmath2al is bright enough to also be seen from localized regions along the galactic plane hosting many massive stars ( fig .
[ fig : al_region_spectraspi ] ) . for a specific source region , distance uncertainties
can often be resolved from other observations , and , moreover , the stellar population is determined through star catalogues .
therefore , a comparison of predicted versus observed amounts of @xmath2alcan be made ( see fig .
[ fig7:al_cygnus ] ) , and provide a more specific test for massive - star models than can be obtained from galaxy - wide analysis as discussed above . with deeper exposure , this became possible in the late integral mission for the cygnus , carina , and scorpius - centaurus regions ; for other candidate locations of massive - star groups integral s sensitivity is insufficient , due to their fainter @xmath2alemission . the brightest individual @xmath2alemission region appears to be the cygnus region . between 6 and 9 ob associations
potentially contribute to the signal , whose ages range from 2.5 to 7.5 my . probably the cyg ob2 association dominates by far @xcite : about 120 stars in the high - mass range ( 20120 m@xmath8 ) have been identified to relate to cyg ob2 ; the other associations typically are ten times smaller .
the age and distance of cyg ob2 is 2.5 my and 1.57 kpc , respectively . because of its young age , stellar evolution even for the most - massive stars should still not be completed , and contributions from core - collapse supernovae to @xmath2alproduction should be small or absent . instead ,
wolf - rayet - wind ejected @xmath2alfrom hydrostatic nucleosynthesis may be assumed to dominate , currently originating from cyg ob2 stars .
the @xmath2alemission towards the cygnus direction has been measured with integral in the longitude interval [ 70@xmath9,96@xmath9 ] as a total of @xmath406 10@xmath11ph @xmath12s@xmath13 .
accounting for a large - scale galactic background , the contribution from the _ cygnus complex _ is @xmath403.9 10@xmath11ph @xmath12s@xmath13 .
accumulation of the expected @xmath2alproduction from the stellar census has always shown that @xmath2algamma - rays seemed 23 times brighter than predicted . in fig .
[ fig7:al_cygnus ] , the horizontally - shaded area presents the range given by the @xmath2algamma - ray data , the dashed lines bracket the uncertainty range of predictions from recent massive - star models through population synthesis .
expectations from such populations synthesis are on the low side of observed @xmath2algamma - rays , and predict the main @xmath2alejection still to come within the next @xmath7 3 my . while much of this discrepancy could be assigned to the occultation of stars by molecular clouds , some under - prediction ( @xmath4025% for solar metallicity ) remained , though still within uncertainties of both observation data and models . the metallicity dependence of wolf - rayet - phase @xmath2alyields results for cygnus region metallicity in a revised @xmath2alprediction by a factor @xmath7 4 below the measurement ( see fig .
[ fig7:al_cygnus ] at time=0 ) . for a young and active region of massive - star action
, one may plausibly assume that the interstellar medium would be peculiar and probably more dynamic than in a large - scale average . with the fine spectroscopic resolution of the integral measurements ,
therefore initial hints for a broadened @xmath2algamma - ray line were tantalizing . with better data
, it turns out that the @xmath2alline seen from the cygnus region is compatible with the laboratory energy ( i.e. no bulk motion exceeding tens of km s@xmath13 ) and with instrumental line width ( i.e. no excessive doppler broadening beyond @xmath40200 km s@xmath13 . note that @xmath2alejection from wolf rayet winds would be with @xmath401500 km s@xmath13 , decelerating as circum - stellar gas is swept up .
the nearby region corresponding to stars from the scorpius - centaurus association has been discriminated against the galactic emission in @xmath2al@xcite .
this stellar association and its subgroups are located at a distance of about 100150 pc .
several subgroups of different ages ( 5 , 16 , and 17 my , with typical age uncertainties of 12 my ; , @xcite ) have been identified .
similar to the orion region , this promises that even modest spatial telescope resolution such as from spi ( 2.7@xmath9 ) could reveal @xmath2alemission displaced from its sources , and thus teach us about ejecta flows . with spi , @xmath2alemission from the direction overlapping the current location of the upper - sco group of stars
could be discriminated against the large - scale galactic @xmath2alemission , due to its favorable location about 20@xmath9above the plane of the galaxy .
the @xmath2alejected from the scorpius - centaurus stars should in fact be distributed over a large region on the sky .
but the distribution might help to distinguish contributions from different - age subgroups , thus helping to study the scenario of triggered star formation .
the @xmath2alline from the scorpius - centaurus region source may be slightly blue - shifted : a centroid energy of 1809.46 kev ( @xmath230.48 kev ) has been found , and implies a blue shift of @xmath400.8 kev corresponding to bulk streaming towards the sun at about ( 137@xmath2375 ) km s@xmath13 .
interestingly , measurements of hot gas in the solar cavity also suggested gas inside the local cavity streaming from this general direction towards the sun @xcite .
the orion region is the most - nearby region of massive stars , at a distance of @xmath40450 pc .
its location towards the outer galaxy and at galactic latitudes around 20@xmath9is favorable , as potential confusion from other galactic sources is negligible .
the dominating group of massive stars is the orion ob1 association with three major subgroups of different ages , one oldest subgroup _ a _ at 812 my , and two possibly coeval subgroups _ b _
( 58 my ) and _
c _ ( 26 my ) ; subgroup _ d _ is the smallest and youngest at 1 my or below .
subgroup c holds the most massive stars , about 45 in the mass range 4120 m@xmath8 .
these groups are located on the near side of the orion a and b molecular clouds , which extend from 320 to 500 pc distance away from the sun , and span a region of @xmath40120 pc perpendicular to our viewing direction .
@xmath2alemission from orion was seen with comptel ( a gamma - ray flux of 7 10@xmath11ph @xmath12s@xmath13 at 5@xmath14 significance , @xcite ) , and appears quite extended and not concentrated near the orion ob1 association .
a huge interstellar cavity extends from the orion molecular clouds towards the sun , the eridanus cavity , which extends over almost 300 pc and is seen in x - ray emission @xcite .
@xmath2alejected from current - generation stars would find a pre - shaped cavity directing the flow of ejecta into it , rather than towards the dense remains of the parental molecular clouds on the far side of the ob association . with integral s spectrometer the @xmath2alline centroid ( bulk motion towards the sun ? ) and width ( @xmath2alfractions moving within the hot cavity versus @xmath2aldeposited at the cavity walls ?
) may provide interesting high - resolution spectroscopy constraints , from observations taken in 2012/2013 .
nuclei such as the above - discussed @xmath2al , @xmath1ti , and @xmath0ni isotopes decay through @xmath41-decays , thus releasing positrons into interstellar space , with typical energies of @xmath7 mev . colliding with their antiparticles , the electrons , in the interstellar medium , annihilation photons are released , with a total energy equivalent to the rest mass of the two leptons of 1.022 mev plus their kinetic energy . from momentum and spin conservations ,
the photon spectrum is dominated by photons at 511 kev , plus a continuum with a maximum energy of 511 kev from annihilation through the ortho - positronium atom made of the two leptons as an intermediate state .
there are other candidate sources of interstellar positrons , such as magnetized and rotating neutron stars and accreting binary systems , which are expected to release electron - positron plasma with much higher energies ( @xmath7gev ) , and dark - matter particle interactions may also plausibly contribute to produce positrons ( for a review see @xcite ) .
the positron annihilation gamma - ray line was the first cosmic gamma - ray line ever detected , in 1972 , with a low - resolution nai detector instrument @xcite , and later identified as annihilation line through the ge detector measurement providing sufficient spectroscopic precision for line identificantion@xcite .
apparently time - variable annihilation emission was pursued for a while , hoping for a key diagnostic of pair plasma in compact sources @xcite .
but most of the apparent variability was later found to appear because different instruments with different field - of - view sizes recorded different fractions of the diffuse galactic annihilation emission @xcite .
still , a few transient 511 kev flashes from x - ray transients have been reported , and remain a target to identify a unique high - energy source process , either pair plasma ejection , or short - lived @xmath41-decay radioactivity such as expected for novae ( see next section below ) .
integral has imaged the positron annihilation gamma - rays across the sky in great detail , and confirmed the diffuse nature of annihilation across our galaxy .
the scientific surprise of this emission had already been apparent in earlier results from the compton observatory @xcite , and was consolidated by spi measurements : the annihilation emission predominantly arises in the inner galaxy in an extended region of size @xmath7 10@xmath9 . by comparison ,
the disk of the galaxy is much fainter , wit a bulge - to - disk intensity ratio of 1.4 from a total luminosity of @xmath7 2 10@xmath16ph @xmath12s@xmath13 .
the line shape of the 511 kev line , and the line - to - continuum ratio of the emission from positron annihilation , both carry information about the conditions in the annihilation sites : doppler broadening from the kinetic energy of the annihilating leptons is expected to increase the line width , unless annihilation occurs at low ( thermal ) energies only , or the momentum is transferred to a heavy atomic nucleus or dust grain surface where annihilation may occur . from the rather narrow 511 kev line , neutral gas appears to be an important ingredient of annihilation sites , and positrons form a positronium atom capturing an electron from interstellar hydrogen , before annihilating . on average ,
annihilation environments are only moderately ionized ( percent level ) , and moderately hot at @xmath7 8000 k ; this suggests that annihilation does not occur in the hot interstellar medium that is characteristic for the surroundings of most candidate sources , but rather in the outer , partially - ionized boundary layers of molecular clouds . in view of the variety of candidate sources of cosmic positrons @xcite , it remains challenging to attribute positron source intensities to each of them .
although their positron injection energies may be different , and also their characteristic locations in the galaxy may be different , positron propagation from their sources to their annihilation sites is quite uncertain .
this leaves interpretational freedom for slowing down , and rearrangements of the spatial distributions .
spi images ( figure [ fig : annihilation_image_spi ] ) include hints for deviations from the ideal symmetry expected for dark - matter origins of the bulge emission .
asymmetric disk emission was interpreted as possibly related to x - ray binaries @xcite .
but in fact , each of the candidate sources may have its own deviations from perfect galactic symmetry , and conclusions are model dependent and rely on the quality of the imaging information .
therefore , efforts in those studies focus on precise mapping of the annihilation emission and its correlation with source distribution models , and on searches from known / specific candidate positron - producing objects .
nova explosions are understood as the nuclear ignition of the accreted surface layer on a white dwarf .
nucleosynthesis is characterized by explosive hydrogen burning , producing preferentially nuclei on the proton - rich part of the valley of stable nuclei .
this leads to the expectation of positron annihilation gamma - rays ( see above ) following the radioactive decay of those freshly - produced nuclei , with their often short lifetimes @xcite .
however , this _ annihilation flash _ would occur reveal days before the nova achieves its optical peak brightness , and thus would precede the detection and identification of the nova as a celestial source .
therefore , post - analyzing sky survey data for serendipitous exposures of such novae are needed to discover such a characteristic signature .
recurrent novae probably only process small amounts of previously - accreted material , while classical novae accrete total amounts of 10@xmath21 - 10@xmath11 m@xmath8at very slow rate of order 10@xmath42 m@xmath8 y@xmath13 before ignition of the nova runaway .
about 1/3 of all novae may be due to the more - massive and more - evolved white dwarf progenitors , which are enriched in heavier seed nuclei such as na , mg , and ne . here
it appears plausible that h burning also passes through the ne - na cycle and leaves behind significant amounts of @xmath5na , a radioactive isotope with decay time of 3.8 years , and thus a candidate emitter of gamma - rays at 1274.53 and 511 kev .
a major uncertainty in nova models is the amount of admixed white - dwarf material , another the total amount of ejected material .
the variety of gamma - ray signals which could arise from novae of the co and o - ne types have been summarized in @xcite .
na line , derived testing a spatial source distribution as plausible for nova events @xcite .
the fitted line at 1275 kev can hardly be seen , and corresponds to an ( insignificant ; 1@xmath14 ) flux of 1.3 10@xmath11ph @xmath12s@xmath13 ( p. jean , private communication).,scaledwidth=78.0% ] none of the gamma - ray missions so far have resulted in a detection of any of the gamma - ray lines expected from novae , neither positron annihilation , nor @xmath5na nor @xmath43be decay signals have been seen .
smm @xcite , wind - tgrs @xcite , cgro - comptel @xcite , and integral @xcite all have been used to search for these , and some hints in comptel data seemed promising .
but gamma - ray telescopes all suffer from intense activation and instrumental background in both the 511 kev and more so in the 1275 kev part of the measured spectrum from aluminum activation by cosmic ray bombardments in space . as an example
, we show the search for @xmath5na emission from novae with integral / spi @xcite . here , the sky distribution was modeled from the galactic nova distribution assumed to follow the 2.4 @xmath44 m dust emission in the galaxy @xcite , and was fitted to data from two years of observations along the plane of the galaxy , with 1.8 ms of total exposure .
only an insignificant hint of the expected line at 1275 kev could be seen ( figure [ fig:22na_spi ] ) . for the spatial model by @xcite , which includes both a spheroidal and a significant disk population of novae , a 2@xmath14 limit of 2.5 10@xmath21ph
@xmath12s@xmath13 was obtained . translating this into a limit for the nova population attributed to this spatial distribution , adopting a nova rate of 20 ... 40 per year and an o - ne - mg enriched nova fraction of 13 ... 33% , an average ejected - mass limit of 2.55.7 10@xmath45m@xmath8was obtained .
cosmic - ray interactions with interstellar gas are expected to produce spectral signatures at gamma - ray energies through excitation of nuclear levels , and through pion decay and higher - energy nucleonic excitations @xcite .
we concentrate on signatures from atomic nuclei here , leaving pion and gev lines aside ( see @xcite for a review of continuum processes ) . in the nuclear line region
, lines are expected from @xmath33c and @xmath46o at 4.43 and 6.1 mev , respectively .
a nuclear line de - excitation spectrum including those c and o lines may be expected also from particle acceleration sites and sources , as energetic particles interact with ambient gas within or near the accelerating region . for a supernova remnant
, a detailed prediction is found in , showing rather broad lines from c and o at 4.4 and 6.1 mev .
estimated fluxes fall close to the sensitivity of comptel ; due to the large line width , these are well below integral s capabilities .
the capture of neutrons on hydrogen is expected to produce a line at 2.223 mev , as thermalized neutrons form deuterium with hydrogen nuclei and liberate the neutron binging energy .
such gamma - ray line emission is prominent in instrumental lines , as atmospheric neutrons capture on hydrogen - rich propellants present in satellites @xcite , but also could arise from the surface of compact stars ( white dwarfs , neutron stars ) , in particular in accreting binary systems .
line shape information could be valuable , as orbiting material in accretion disks could be identified , or gravitational redshifts be exploited wrt .
neutron star accretion geometry .
a search for this line and corresponding point sources was negative , and only provided upper limits@xcite .
this is in line with expectations for the candidate sources and their plausible accretion rates .
searches for above - discussed nuclear lines have been performed , from data addressing regions where cosmic ray interactions plausibly produce such excitations .
targets of interest are nearby clusters of massive stars , and along the inner ridge of the galaxy where many of those star clusters should exist along any line of sight @xcite .
the measurements with high - resolution spectrometers are less sensitive to the typical and largely - broadened lines ; therefore , integral and rhessi have not obtained useful measurements . but
comptel s scintillation detector resolution of @xmath7 10% ( fwhm ) had been suitable for setting constraints .
unfortunately , the discussion and assessment of instrumental backgrounds had been difficult ( and controversial ) within the comptel team , and conflicting results were obtained .
an early report of nuclear lines from the orion region @xcite created much excitement @xcite , also because the reported intensities would have implied very efficient cosmic - ray acceleration in this region @xcite .
but after excluding possibly - contaminated parts of the data , the flux constraints had to be relaxed @xcite , and no clear line detection remained . it appeared that the line intensities expected from low - energy cosmic ray excitations both from orion and from the inner galaxy are just about near the instrumental limits of the comptel observations .
thus , no new lessons can be taken yet , and must await a more - sensitive instrument / mission in the mev energy band .
energetic particle acceleration is a key characteristic of solar flares , according to the established model @xcite : reconnection events of magnetic field lines in the upper solar corona set up some kind of fermi accelerator , at significant altitude above the solar chromosphere .
it remains to be understood how the acceleration is set up and evolves @xcite .
it seems clear that the energy for particle acceleration derives from the magnetic field , and that changes in its configuration set up the critical conditions to create relativistic particles . as a result , high - energy protons , nuclei , and electrons
hit the deeper layers of the solar chromosphere , and produce a variety of observational signatures once interacting with denser gas towards the solar photosphere , at typical gas densities 10@xmath47cm@xmath16 or higher . among those signatures , electron bremsstrahlung produces continuum emission , nuclear excitations produce a variety of lines , most prominently @xmath49ne , @xmath33c , and @xmath46o de - excitation at 1.634 , 4.439 , and 6.129 mev , respectively , and spallation - produced neutrons capture on hydrogen and produce a characteristic line at 2.223 mev .
most - detailed gamma - ray spectra had been measured for the 4 jun 1991 solar flare with the osse spectrometer on cgro , confirming these spectral characteristics @xcite .
the modest spectral resolution of osse ( @xmath7 8% fwhm ) implied that the most - valuable diagnostics were temporal changes of spectral - line intensities .
for example , the relative intensities of the @xmath49ne and @xmath46o lines encode the particle spectrum , because the nuclear - excitation cross sections for these isotopes are very different , @xmath49ne activation having a much lower excitation threshold energy . analyzing this line ratio , and comparing with electron bremsstrahlung
, it was found that the charged - particle energy is about equally distributed between electrons and nuclei , and that the ion energy spectrum is steep with typical power - law slopes between -3.5 and -5.5 @xcite .
important diagnostics from high - resolution spectroscopy employs doppler shifts of gamma - ray lines or attenuation between source and observer , which modify the position and shape of the gamma - ray lines ( fig .
[ fig : solflare_c_o ] ) .
flare - accelerated particles propagate downward , spiraling around magnetic field lines . depending on the importance of scattering , and on convergence of field lines , particles may propagate deeper into the chromosphere , or get magnetically mirrored at higher altitudes when pitch angles are large .
therefore the line shape details encode how accelerated particles lose their energy after the acceleration event and before thermalization . the _ ramaty high - energy solar spectrometer instrument ( rhessi ) _
@xcite included a gamma - ray spectrometer with 9 ge detectors , located behind a rotation - modulation collimator for high - resolution imaging at 2.3 arcsec resolution .
rhessi always points to the sun .
solar - flare data collected with spi / integral typically occur with the instrument pointed away from the sun , and gamma - rays entering ge detectors through the side and penetrating the bgo scintillation detector which forms spi s anti - coincidence system .
a bright flare occurred on 28 oct 2003 , early in both those missions , and resulted in bright gamma - ray line signals extending over a period of @xmath7 15 min .
the gamma - ray flare started off with intense continuum emission , but within @xmath7 a minute , line emission set in , with nuclear de - excitation lines preceding the neutron capture line at 2.223 mev , as expected from the required slowing - down processes .
nuclear lines were identified from neutron capture on hydrogen , and from excited @xmath33c , @xmath46o , @xmath50 mg , @xmath49ne , and @xmath51si nuclei .
the line intensity ratios evolve somewhat over the period of the flare .
in particular the c and o line ratios are sensitive to the ratio of @xmath19 particles over protons in the flare , and suggest that the @xmath19/p ratio also evolves during the flare itself , reducing the @xmath19-particle content in the later flare phase .
spi data from these lines detected of kinematic doppler shifts / distortions ( see fig .
[ fig : solflare_c_o ] ) .
these were interpreted as a beam geometry of accelerated particles being downward - directed and rather fan - like , which implies that pitch - angle scattering in the solar atmosphere is large .
since mid 2010 , again more solar flare events are being observed as solar activity begins to increase towards the next solar maximum in mid 2013 .
observations now also include measurements from the gbm scintillation detectors aboard the fermi satellite .
studies focus on discrimination of different energy ranges , as they determine nuclear lines , the 2.23 mev neutron capture line , and the spectral signature related to pions ( in the order of increasing accelerated - particle energies ) , so try and learn about the energetics of the solar - flare particle acceleration process .
most of the spectroscopic signatures which are exploited in astronomy are related to absorption processes , which shape a continuum background light source in ways characteristic for the absorbing atoms or molecules .
atomic nuclei should , by analogy , provide absorption signatures which are characteristic for nuclei
. this will be particularly useful where matter is fully ionized , such as expected in the vicinity of gamma - ray burst sources or active galactic nuclei .
both these sources generate intrinsic photon emission with a featureless continuum spectrum extending to gamma - ray energies .
candidate nuclear absorption from early - universe cosmic matter could arise from nuclear excitation of abundant nuclei such as c and o ( see above ) , but probably more significantly from he excitation and dissociation through the giant dipole resonance with a threshold energy of @xmath7 25 mev .
high redshifts of the most - interesting objects move such features into the @xmath7 mev regime .
gamma - ray telescopes so far could not find these signals , which should , however , be detectable with next - generation gamma - ray telescopes @xcite .
the pioneering balloone - borne experiments ( @xmath7 1976 ) and the comptel @xcite and osse @xcite instruments aboard the compton observatory ( 1991 - 2000 ; @xcite ) provided a first sky survey and identified several prominent gamma - ray lines , thus confirming expectations from theory . the integral mission @xcite deepened the exploration of the nuclear - radiation sky since 2002 .
gamma - ray spectroscopy was thus consolidated with the measurements at sufficient spectral resolution of ge detectors through esa s integral mission and its spectrometer spi @xcite .
the sensitivity and astronomical resolution of all these instruments is comparable ( fluxes must be above 10@xmath11ph @xmath12s@xmath13 and sources can only be resolved on the degree scale ) , and only the brightest sources can be seen due to backgrounds intrinsic to all instruments . the predicted and known gamma - ray lines from radioactive isotopes have been confirmed , measurements have been enriched with spectroscopic detail which add significant astrophysical aspects .
the characteristics of @xmath1ti emission from the cas a supernova remnant and candidate sources in the galaxy show that nucleosynthesis in core - collapse supernovae occurs under a diversity of environmental conditions , which result from substantial deviations from spherical symmetry in these explosions @xcite .
exploitation of gamma - rays from the @xmath0ni decay chain which is responsible for supernova light still awaits a sufficiently - nearby supernova of type ia @xcite ; the independent information carried by gamma - rays from this decay chain could help to understand details of explosion physics , which happens in the dense and dynamic initial phase of the exploding star that is otherwise occulted to observations .
diffuse gamma - ray emission from @xmath2al decay @xcite as well as positron annihilation gamma - rays at 511 kev @xcite have established their own astronomical windows , where unique measurements are being made .
the detailed multi - messenger comparison of their emission in intensity and spatial morphology with model predictions for each specific candidate sources provides insights into massive - star groups and high - energy sources throughout the galaxy .
the puzzle of the bright bulge - like 511 kev image holds a promise for learning new aspects both of cosmic - ray propagation and of dark - matter interactions in our galaxy .
@xmath3fe radioactivity from ( probably ) the same sources which create @xmath2alprovide a global massive - star population diagnostic of the interior structure of massive stars .
the discovery of @xmath3fe also in ocean crust material @xcite underlines the role of such long - lived radioactive isotopes as tools to study cosmic nucleosynthesis , in particular also addressing the aspect of how core - collapse supernovae spread their ejecta in their surroundings .
@xmath2alemission from nearby regions is bright enough so it can be located relative to its sources .
fine spectroscopy of the @xmath2alline in the inner galaxy shows doppler shifts from the large - scale motion of galactic rotation with surprisingly - high velocities .
this may teach us new aspects through nucleosynthesis ejecta of how massive stars shape their surroundings and provide feedback that regulates star formation in disk galaxies .
positron annihilation throughout the galaxy shows puzzling brightness of the bulge region of our galaxy , which may find explanations in the complex propagation of positrons in tangled , but possibly also rather regular magnetic - field configurations in interstellar space of the galaxy s disk and halo .
it may also shed light on past activity in the center of our galaxy , or even interactions of dark matter .
the acceleration of particles to the relativistic energies of cosmic - rays are still unclear a hundred years after cosmic rays had been discovered @xcite . how thermal particles develop high - energy tails , how these are then boosted towards relativistic energies , and where this happens , this remains to be one of the big astrophysical questions .
gamma - ray spectroscopy can address the low - energy part of physical processes herein , and detailed spectra of these same lines are being studied from solar flares and their gamma - ray spectra . nuclear line emission is expected from more and different sources than the ones known so far , such as nova explosions or the surfaces of compact stars in binary systems , and from other isotopes than the ones discovered so far . the unique and different processes which generate nuclear line features are only partly exploited with respect to their astrophysical messages .
an advance in sensitivity is needed , to reach other nearby galaxies in the known lines , and to find several more of the expected .
ideas for such next - generation telescopes have been presented , but not given priority by evaluating committees .
exploitation of the current instruments and missions can still add important detail , on the galactic sources , and on instrument building and data analysis approaches .
the nuclear - physics signatures of cosmic processes can teach us aspects that are difficult to reveal by other means .
* acknowledgements . *
the integral / spi project has been completed under the responsibility and leadership of cnes ; we are grateful to asi , cea , cnes , dlr ( grants 50 og 1101 and 50 or 0901 ) , esa , inta , nasa and ostc for support of this esa space science mission and its science analysis .
this research was supported also by the german dfg cluster of excellence _ origin and structure of the universe_. h. bloemen , a. m. bykov , r. diehl , w. hermsen , _ et al . _ . in c. d. dermer , m. s. strickman , and j. d. kurfess , editors , _ proceedings of the fourth compton symposium _ ,
volume 410 of _ american institute of physics conference series _ , pages 1074+ , nov .
1997 .
r. diehl , d. h. hartmann , and n. prantzos . .
in r. diehl , d. h. hartmann , & n. prantzos , editors , _ lecture notes in physics , berlin springer verlag _ , volume 812 of _ lecture notes in physics , berlin springer verlag _ , pages 345438 , 2011 .
j. isern , m. hernanz , and j. jos . .
in r. diehl , d. h. hartmann , & n. prantzos , editor , _ lecture notes in physics , berlin springer verlag _ , volume 812 of _ lecture notes in physics , berlin springer verlag _ , pages 233308 , 2011 .
a. f. iyudin , r. diehl , g. g. lichti , _ et al . _ . in a.
gimenez , v. reglero , and c. winkler , editors , _ exploring the gamma - ray universe _ , volume 459 of _ esa special publication _ , pages 4146 , sept
. 2001 .
p. jean , j. kndlseder , m. hernanz , _ et al . _ . in v.
schoenfelder , g. lichti , & c. winkler , editor , _
5th integral workshop on the integral universe _
, volume 552 of _ esa special publication _
, page 119 , oct .
p. jean , j. kndlseder , m. hernanz , _ et al . _ . in v.
schoenfelder , g. lichti , & c. winkler , editor , _
5th integral workshop on the integral universe _ , volume 552 of _ esa special publication _
, page 119 , oct . 2004 .
j. kndlseder , m. valsesia , m. allain , _ et al .
_ . in v. schoenfelder , g. lichti , and c. winkler , editors , _
5th integral workshop on the integral universe _ ,
volume 552 of _ esa special publication _ , pages 33+ , oct .
2004 .
s. plschke , r. diehl , v. schnfelder , _ et al . _ . in a.
gimenez , v. reglero , and c. winkler , editors , _ exploring the gamma - ray universe _ , volume 459 of _ esa special publication _ , pages 5558 , sept .
2001 .
thielemann , r. hirschi , m. liebendrfer , and r. diehl . .
in r. diehl , d. h. hartmann , & n. prantzos , editor , _ lecture notes in physics , berlin springer verlag _ , volume 812 of _ lecture notes in physics , berlin springer verlag _ , pages 153232 , 2011 .
a. zoglauer , r. andritschke , s. e. boggs , _ et al .
_ . in _ society of photo - optical instrumentation engineers ( spie ) conference series _ ,
volume 7011 of _ society of photo - optical instrumentation engineers ( spie ) conference series _ , aug . 2008 .
|
penetrating gamma - rays require complex instrumentation for astronomical spectroscopy measurements of gamma - rays from cosmic sources .
multiple - interaction detectors in space combined with sophisticated post - processing of detector events on ground have lead to a spectroscopy performance which is now capable to provide new astrophysical insights .
spectral signatures in the mev regime originate from transitions in the nuclei of atoms ( rather than in their electron shell ) .
nuclear transitions are stimulated by either radioactive decays or high - energy nuclear collisions such as with cosmic rays .
gamma - ray lines have been detected from radioactive isotopes produced in nuclear burning inside stars and supernovae , and from energetic - particle interactions in solar flares .
radioactive - decay gamma - rays from @xmath0ni directly reflect the source of supernova light .
@xmath1ti is produced in core - collapse supernova interiors , and the paucity of corresponding @xmath1ti gamma - ray line sources reflects the variety of dynamical conditions herein . @xmath2al and @xmath3fe are dispersed in interstellar space from massive - star nucleosynthesis over millions of years .
gamma - rays from their decay are measured in detail by gamma - ray telescopes , astrophysical interpretations reach from massive - star interiors to dynamical processes in the interstellar medium .
nuclear de - excitation gamma - ray lines have been found in solar - flare events , and convey information about energetic - particle production in these events , and their interaction in the solar atmosphere .
the annihilation of positrons leads to another type of cosmic gamma - ray source .
the characteristic annihilation gamma - rays at 511 kev have been measured long ago in solar flares , and now throughout the interstellar medium of our milky way galaxy .
but now a puzzle has appeared , as a surprising predominance of the central bulge region was determined .
this requires either new positron sources or transport processes not yet known to us . in this paper
we discuss instrumentation and data processing for cosmic gamma - ray spectroscopy , and the astrophysical issues and insights from these measurements .
|
one of the major goals of spintronic applications has always been to manipulate electron s spin degree of freedom to create a new paradigm @xcite in the fields of quantum information processing .
spin-@xmath3 particles are a natural choice for a qubit in quantum computers .
so , generation of spin polarized beam is a highly significant issue as far as spintronic applications are concerned . a more or less usual way of realization @xcite of spin filtering action is by using ferromagnetic leads or by external magnetic field .
but , in the first case , spin injection from ferromagnetic lead is difficult due to large resistivity mismatch and for the second one , the difficulty is to confine a very strong magnetic field into a small region like a quantum ring .
therefore , attention is being paid for modeling of spin filter using the intrinsic properties of mesoscopic systems such as spin - orbit ( so ) interaction @xcite .
originating from the relativistic correction to the schrdinger equation , so interaction provides an all electrical way to generate and manipulate spin current in a far precise way rather than the usual magnetic field based spin control .
the main source of so coupling in mesoscopic systems comes either from magnetic impurities ( extrinsic type ) , or from bulk asymmetry or structural inversion asymmetry in the confining potential of the system ( intrinsic type ) yielding dresselhaus or rashba type of so interaction @xcite .
studies on rashba or dresselhaus kind of interactions has made a significant impact in semiconductor spintronics as far as the control of spin dynamics is concerned .
since so interaction couples the spin degree of freedom with the momentum of an electron , so it is possible to achieve spin polarized currents in output terminals of a multi - terminal mesoscopic ring when an unpolarized electron beam is injected into its input terminal , though kramer s degeneracy suggests that due to preservation of time reversal symmetry only so interaction can never induce in a two - terminal system @xcite , whereas , in case of multi - terminal system the condition gets relaxed . additionally , in a multi - terminal system the degree of spin coherence can also be manipulated even when a polarized beam is injected . till date
a lot of theoretical work has been done to model spin selective transmission . in 2003 ,
kislev and kim have proposed that a planar t - shaped structure with a ring resonator can be highly efficient in producing spin polarized currents in different output arms in presence of rashba so interaction @xcite . following that in 2005 , shelykh _ et al . _
investigated analytically by s - matrix theory both the effects of magnetic flux and rashba coupling on charge and spin transport in a multi - terminal quantum ring @xcite .
then , in 2006 , peeters _
et al . _ have shown that a mesoscopic semiconductor ring with one input and two outputs can act as an electron spin beam splitter due to quantum interference effect and rashba so interaction @xcite . in another work , rabani _
_ has designed a spin filter and spin splitter considering a two - terminal device in presence of external magnetic field @xcite .
thus , the studies involving spin dependent transport in multi - terminal mesoscopic rings have already generated a wealth of literature , but to the best of our knowledge there is still a need to look deeper into the problem to address several important issues those have not been well analyzed earlier , for example , the understanding of the effect of rashba and dresselhaus so couplings on all three components @xmath0 , @xmath1 and @xmath2 of the polarization vector associated with the charge current through the outgoing leads , and also the sensitivity of the polarization components on different electrode - ring interface configurations . in the present work we mainly concentrate on these issues .
here we analyze the dependence of the polarization components @xmath0 , @xmath1 and @xmath2 in the two output terminals of a three - terminal mesoscopic ring subjected to rashba and dresselhaus so interactions . within a tight - binding ( tb ) framework we evaluate these quantities ( @xmath0 , @xmath1 and @xmath2 ) using a general spin density matrix formalism , which was first approached by nikolic and co - workers @xcite in 2004 .
they presented an elegant way of expressing all three components @xmath0 , @xmath1 and @xmath2 of outgoing spin polarization vector in terms of spin resolved transmission matrices within the framework of landauer - bttiker formalism .
this formalism encompasses both pure and mixed states as incoming beam and provides knowledge to investigate how polarization evolves @xcite in outgoing current due to rashba or dresselhaus interaction , scattering at lead - conductor interface and spin dependent scattering off impurities in both the weak and strong disordered regimes .
the sensitivity of the polarization components on the electrode - ring interface geometry is also described in detail to make the present communication a self contained study .
the rest of the paper is organized as follows . in section
ii , we present the model and theoretical formulation to obtain spin polarization components of the output currents in terms of transmission probabilities .
the numerical results are illustrated in section iii .
finally , in section iv , the results are summarized .
in this section we describe the model quantum system within a tb framework and express the spin polarization components @xmath0 , @xmath1 and @xmath2 in terms of transmission coefficients of outgoing electrons through the quantum ring following spin density matrix formalism .
let us start with fig .
[ multiring ] where a mesoscopic ring subjected to rashba and dresselhaus so interactions is attached with one input and two output terminals . a simple lattice model within the framework of tb approximation assuming only nearest - neighbor coupling is used to describe the ring and side - attached leads .
the tb hamiltonian describing the entire system gets the form : @xmath4 the first term @xmath5 describes the hamiltonian of the ring and for a @xmath6-site ring it reads , @xmath7 where , + @xmath8 @xmath9 and @xmath10 .
+ 0.2 cm here , @xmath11 and @xmath12 are the creation and annihilation operators , respectively , at the @xmath13-th site for an electron with spin @xmath14 ( @xmath15 , @xmath16 ) .
@xmath17 being the on - site energy .
the factor @xmath18 is the sum of three terms as follows .
@xmath19 here @xcite , + @xmath20 , + @xmath21 , + @xmath22 .
+ 0.1 cm in these above expressions @xmath23 is the isotropic nearest - neighbor coupling strength , whereas , @xmath24@xmath25 and @xmath24@xmath26 correspond to the spin dependent terms with @xmath27 and @xmath28 being the nearest - neighbor hopping integrals due to rashba and dresselhaus so interactions , respectively , which introduce spin flipping in the system and @xmath29 is the azimuthal angle for the @xmath13-th site .
mathematically , it can be expressed as @xmath30 . in our formulation
we assume that the one - dimensional ( @xmath31d ) semi - infinite leads are free from any kind of disorder and so interactions .
they can be expressed as , @xmath32 where , @xmath33 similarly , the ring - to - lead coupling is described by the following hamiltonian .
@xmath34 here , @xmath35 .
\label{eqn7}\ ] ] in these above equations ( eqs . [ eqn4]-[eqn7 ] )
, the index @xmath36 signifies the number of leads attached to the ring .
it can be two or three or even more depending on the number of outgoing leads in addition to the incoming one .
@xmath37 and @xmath38 stand for the site energy and nearest - neighbor hopping between the sites of the leads .
the coupling between the leads and the ring is denoted by the hopping integral @xmath39 . in eq .
[ eqn7 ] , @xmath40 and @xmath41 belong to the boundary sites of the ring and the leads , respectively .
most of the quantum interference phenomena observed in different experiments , e.g. , aharonov - bohm effect , weak localization effect , etc .
, within the mesoscopic regime deal with the aspect of orbital quantum coherence of electronic states
. at much low temperatures ( @xmath42k ) and for the systems having @xmath43 ( @xmath44 m ) , inelastic scattering processes get suppressed so that an electron can be described by a single orbital wave function within the system .
now if the spin degree of freedom of the electron is taken into account then two separate vector spaces are multiplied tensorially to get the full hilbert space of the quantum states i.e. , @xmath45 here , @xmath46 spans over the orbital degrees of freedom while @xmath47 operates in spin space only .
therefore , any arbitrary state @xmath48 can be written as a linear combination of @xmath49 , where @xmath50 s are the basis vectors of @xmath46 and @xmath51 s are the eigenstates of @xmath52 . @xmath53
being the pauli spin matrix and @xmath54 is the unit vector along the direction of spin quantization axis .
thus , we can write the most general form of @xmath55 as , @xmath56 here , we choose the quantization direction along the @xmath57ve @xmath58 direction , and accordingly , @xmath51 s are the eigenstates of @xmath59 operator i.e. , @xmath60 @xmath61 @xmath62 and @xmath63 @xmath61 @xmath64 .
the corresponding density matrix operator @xcite for the state @xmath65 becomes , @xmath66 this state can describe a pure one or a mixture of different quantum states .
below we consider both these two cases to get a complete picture .
+ 0.01 cm @xmath67 * coherent beam of electrons : * first , we consider the case where the state @xmath55 is a pure one indicating a coherent beam of electrons .
if the electron is free from any kind of spin - dependent interactions , then spin and charge coherences are independent of each other , resulting a separable state @xmath55 . in that case
@xmath68 can be written as , @xmath69 where , @xmath70 and @xmath71 are the orbital and spin parts , respectively , of the total wave function @xmath55 and they are expressed as follows . @xmath72 the corresponding density matrix operator is thus given by , @xmath73 here , @xmath74 and @xmath75 are the reduced density matrices in the orbital and spin spaces , respectively .
now , in presence of so interaction , the state @xmath55 is no longer separable , and hence , individual orbital and spin parts lose their coherence and become entangled and the state @xmath55 can be written in a form as expressed in eq .
[ eqn9 ] .
in this situation , @xmath76 @xmath67 * incoherent beam of electrons : * next , we consider the case where the electronic beam is an incoherent mixture i.e. , statistical superposition of different quantum states . here
, the state can not be expressed like eqs .
[ eqn9 ] and [ eqn11 ] .
in this situation the state is best represented by the density matrix operator as , @xmath77 where , @xmath78 gives the probability for the ensemble to be found in the quantum state @xmath79 .
in the present work , we want to determine the components @xmath0 , @xmath1 and @xmath2 of output currents propagating through the leads attached to the ring subjected to so interactions . using pauli spin operators they are expressed as follows .
@xmath80 now , following quantum statistics , measurement of any spin observable @xmath81 is accomplished by the following way , @xmath82 \label{eqn16}\ ] ] where @xmath83@xmath84 is the matrix form of the operator @xmath81 ans
@xmath85@xmath84 is the spin density matrix which is obtained by taking partial trace over the orbital degrees of freedom of the full density matrix @xmath85 , @xmath86 \nonumber\\ & = & \sum_{\alpha } \langle \phi_{\alpha}|\mbox{\boldmath$\rho$}| \phi_{\alpha } \rangle .
\label{eqn17}\end{aligned}\ ] ] thus , in our case spin density matrix plays the central role in understanding the quantum dynamics of a spin sub - system , subjected to so interactions and attached to the environment through ideal ( free from any kind of charge or spin dependent interaction ) leads . for electrons
( i.e. , spin-@xmath87 particles ) @xmath75 has a simple @xmath88 representation in a chosen basis @xmath60 ,
@xmath63 @xmath89 @xmath47 which reads , @xmath90 here , the diagonal elements represent the probabilities of finding an electron with spin @xmath60 or @xmath63 , whereas the off - diagonal elements describe the probabilities of coherent superposition of @xmath60 and @xmath63 states due to quantum interference effect .
it can also be represented as , @xmath91 here @xmath92 is the polarization vector whose components are evaluated from the following relations .
@xmath93 , \nonumber\\ p_y & = & tr[\mbox{\boldmath$\rho$}_s \mbox{\boldmath$\sigma_y$ } ] , \nonumber\\ p_z & = & tr[\mbox{\boldmath$\rho$}_s \mbox{\boldmath$\sigma_z$ } ] .
\label{eqn20}\end{aligned}\ ] ] for a completely unpolarized electron beam i.e. , an incoherent mixture of up and down spin electrons , @xmath94 = @xmath95 . in this case
, the spin density operator ( @xmath75 ) becomes , @xmath96 it is seen that for an incoherent mixture of up and down spin electrons the off - diagonal elements of @xmath75 are zero .
therefore , to determine the polarization components of outgoing currents we need to construct the spin density matrix for the outgoing charge current .
in this sub - section we present a general scheme for evaluating spin polarization components of the charge current through the outgoing leads , considering a mesoscopic ring subjected to so interactions , where each lead contains @xmath97 number of channels . in our problem , we assume that a beam of incoherent i.e. , unpolarized electrons is injected from the source to the ring .
therefore , the incident beam is best represented by the spin density operator as , @xmath98 for a fully unpolarized beam , @xmath99=@xmath100=@xmath3 and hence the polarization components are evaluated as- ( using en .
[ eqn20 ] ) : @xmath101 now , in presence of so interaction , individual spin and orbital parts become entangled , and thus , the corresponding outgoing beam through the lead j , can be described in terms of reduced spin density matrix @xcite by taking partial trace over all the orbital degrees of freedom . @xmath102 where , 0.1 cm @xmath103 , 0.1 cm @xmath104_{n'n}|^2 + n_{\downarrow } |[\mbox{\boldmath$t_{ji}^{\uparrow \downarrow}$}]_{n'n}|^2 $ ] , 0.1 cm @xmath105_{n'n } [ \mbox{\boldmath$t_{ji}^{\downarrow \uparrow}$}]_{n'n}^ { * } + n_{\downarrow } [ \mbox{\boldmath$t_{ji}^{\uparrow \downarrow}$}]_{n'n } [ \mbox{\boldmath$t_{ji}^{\downarrow \downarrow}$}]_{n'n}^{*}$ ] , 0.1 cm @xmath106_{n'n}^ { * } [ \mbox{\boldmath$t_{ji}^{\downarrow \uparrow}$}]_{n'n } + n_{\downarrow } [ \mbox{\boldmath$t_{ji}^{\uparrow \downarrow}$}]_{n'n}^ { * } [ \mbox{\boldmath$t_{ji}^{\downarrow \downarrow}$}]_{n'n}$ ] , 0.1 cm @xmath107_{n'n}|^2 + n_{\downarrow } |[\mbox{\boldmath$t_{ji}^{\downarrow \downarrow}$}]_{n'n}|^2 $ ] .
0.3 cm following the prescription of en .
( [ eqn20 ] ) , the spin polarization components of the outgoing through lead j , current can be expressed as , @xmath108 [ \mbox{\boldmath$t^{\downarrow \uparrow}_{ji}$}]^{\dag } + [ \mbox{\boldmath$t^{\uparrow \downarrow}_{ji}$ } ] [ \mbox{\boldmath$t^{\downarrow \downarrow}_{ji}$}]^{\dag } \right]\right ] , \nonumber\\ p_j^{y } & = & \frac{2 e^2/h}{g_{ji } } \mbox{im}\left[\mbox{tr}\left[[\mbox{\boldmath$t^{\uparrow \uparrow}_{ji}$}]^{\dag } [ \mbox{\boldmath$t^{\downarrow \uparrow}_{ji}$ } ] + [ \mbox{\boldmath$t^{\uparrow \downarrow}_{ji}$}]^{\dag } [ \mbox{\boldmath$t^{\downarrow \downarrow}_{ji}$}]\right]\right ] , \nonumber\\ p_j^{z } & = & \frac{g^{\uparrow \uparrow}_{ji } + g^{\uparrow \downarrow}_{ji } - g^{\downarrow \uparrow}_{ji } - g^{\downarrow \downarrow}_{ji } } { g^{\uparrow \uparrow}_{ji } + g^{\uparrow \downarrow}_{ji } + g^{\downarrow \uparrow}_{ji } + g^{\downarrow \downarrow}_{ji}}. \label{eqn24}\end{aligned}\ ] ] where , @xmath109=@xmath110 . here ,
@xmath111 $ ] is the transmission matrix , having dimension @xmath112 , for an electron injected from the lead i , with spin @xmath14 and transmitted through the lead j with @xmath113 .
thus , for a single channel lead @xmath111 $ ] becomes a simple element rather than being a matrix .
@xmath114 is the conductance of the ring , defined by the landauer formula @xcite , and at low bias voltage it gets the form , @xmath115 [ t_{ji}^{\sigma ' \sigma}$}]^{\dag}\right ] .
\label{eqn25}\ ] ] below we describe the way of determining the transmission matrix and its relation to the conductance .
the transmission matrix is obtained from the well known relation @xcite , @xmath116 here , @xmath117 is the retarded green s function ( in matrix form ) in the reduced dimension @xmath112 , connecting @xmath40-th and @xmath118-th leads ( each lead contains @xmath97 channels ) i.e. , @xmath117 = @xmath119 .
@xmath120 s are the coupling matrices in the same reduced dimension . the single particle green s function describing the complete system
i.e. , the ring with side - attached leads for an electron with energy @xmath121 is defined as , @xmath122 where , @xmath123 .
the problem of finding in the full hilbert space of taking the matrix forms of and can be mapped exactly to a green s function which represents an effective hamiltonian in the reduced hilbert space of the mesoscopic ring .
it is expressed like , @xmath124 where , @xmath125 these s are the self - energies introduced to incorporate the effect of coupling of the ring to the leads .
it is evident from eq .
[ eqn29 ] that the form of the self - energies are independent of the ring itself through which spin transmission is studied .
s describe the coupling between the ring and the leads and they are mathematically defined as , @xmath126 , \label{eqn30}\ ] ] where , and are the retarded and advanced self - energies associated with the @xmath36-th lead , respectively .
this self - energy term is again expressed @xcite as a sum of real and imaginary parts , @xmath127 where , they describe the energy shift and broadening of the energy levels of the ring , respectively .
the finite imaginary part is obtained due to inclusion of semi - infinite leads having continuous energy spectrum .
therefore , the coupling matrices can easily be determined from the self - energy expression and is expressed in the form , @xmath128 from this relation ( eq .
[ eqn32 ] ) the reduced coupling matrices s are constructed . following the reference @xcite , the self - energy matrices ( s )
are evaluated in the reduced hilbert space of the ring .
the details are also available in other articles @xcite .
although the theoretical formulation presented above is based on a general approach considering finite width leads having @xmath97 number of channels , for completeness of the theoretical description , but in this article we present all the numerical results ( sec .
iii ) considering single - channel leads .
in what follows we restrict ourselves to absolute zero temperature and use the units where @xmath129 . throughout the numerical calculations we choose @xmath17 = @xmath37 = @xmath95 and @xmath130 .
the energy scale is measured in unit of @xmath23 and the so coupling strengths ( @xmath27 and @xmath28 ) are also scaled in unit of @xmath23 . before addressing the central problem i.e. , the possibilities of getting spin polarized currents in two output terminals of a three - terminal mesoscopic ring from a completely unpolarized incident electron beam , first we analyze the results for a simple system where a mesoscopic ring , subjected to rashba and dresselhaus so interactions , is coupled to two leads . in fig .
[ poltwosym1 ] we show the variation of spin polarization components @xmath0 , @xmath1 and @xmath2 as a function of energy @xmath121 for a symmetrically connected two - terminal mesoscopic ring in presence of a non - zero value of rashba so coupling strength . here we take a @xmath131-site ring and set @xmath132 . the dresselhaus so coupling is fixed at zero . from this spectrum
it is clearly observed that all three polarization components of the outgoing current become exactly zero for the entire energy band region .
we also carry out extensive numerical work considering different values of rashba so coupling ( the results are shown in fig . [ poltwosym2 ] ) and for other possible lead - ring interface geometries i.e. , for asymmetrical connections and find the identical behavior of these three components .
thus we can emphasize that in absence of any magnetic impurity or external magnetic field , only so coupling can not induce spin polarization in a two - terminal quantum system and this feature is independent of the lead - ring interface geometry .
the reason is that in presence of so interaction the time - reversal symmetry is not broken , and accordingly , the kramer s degeneracy between the states @xmath133 and @xmath134 gets preserved . in a work kim _
et al . _
@xcite have argued on the basis of symmetry of the s - matrix elements that a two - terminal time - reversal invariant system is incapable of producing spontaneous spin polarization .
et al . _
@xcite have also shown in their recent work that only so interaction can not remove the degeneracies of the transmission eigenvalues .
thus , our numerical results exactly corroborate with these arguments .
a similar nature of the polarization components in this two - terminal geometry is also be obtained when they are plotted as a function of dresselhaus so coupling setting @xmath135 . in this sub - section
we discuss the central results of our present investigation i.e. , the interplay of rashba and dresselhaus so couplings and lead - ring interface geometry on the polarization components of outgoing currents in a three - terminal mesoscopic ring .
we analyze the results for two distinct configurations of lead - ring geometry as schematically illustrated in fig .
[ multiring ] . in one configuration
the outgoing leads are coupled symmetrically with respect to the source lead , while in the other case they are connected asymmetrically .
we start by discussing the variation of spin polarization components @xmath0 , @xmath1 and @xmath2 of outgoing currents in a three - terminal mesoscopic ring with rashba so coupling only , that is , setting the dresselhaus so coupling to zero . the results for a @xmath131-site ring with @xmath132 and @xmath136 are shown in fig .
[ poleng1 ] , where the red lines describe the results for the output lead @xmath137 , while for the other lead ( lead @xmath138 ) they are presented by the blue lines . from the spectra
it is observed that the @xmath139 and @xmath58 components of the spin polarization vectors in two symmetrically coupled output leads are exactly identical in magnitude but they carry opposite signs for each value of the injecting electron energy @xmath121 . on the other hand , the component @xmath1 exhibits identical sign in both these two leads providing vanishingly small amplitudes .
these features can be explained from the following arguments .
the @xmath58 component of spin polarization essentially describes the normalized difference between the charge currents of up and down spin electrons flowing through the output leads ( see eq .
[ eqn24 ] ) , since in our present scheme we choose the quantization direction along the @xmath57ve @xmath58 axis .
now , it is well known that the up and down spin electrons scatter in opposite directions when they traverse through a conductor subjected to a so interaction which is the aspect of visualizing mesoscopic spin hall effect and accumulation of opposite spins on the opposite edges .
therefore , spin polarization with opposite signs for the component @xmath2 is expected in two output leads those are attached symmetrically to the mesoscopic ring with respect to the input lead .
the above argument can not be given to explain the characteristic features of the other two components @xmath0 and @xmath1 since we select @xmath57ve @xmath58 axis as the quantization direction . in our theoretical framework
we have already stated that @xmath0 and @xmath1 are evaluated from the expectation values of @xmath140 and @xmath141 .
now , we know from the rashba hamiltonian that @xmath140 is coupled with the @xmath142-component of the momentum of an electron which gets opposite signs at two equivalent atomic sites in the two arms of the ring resulting opposite spin polarization @xmath0 in the output leads .
but , for @xmath1 the situation is somewhat different as @xmath141 is coupled with the @xmath139-component of the momentum which shows identical sign at the equivalent points ( see fig.[diagram ] ) .
hence , a destructive interference takes place and it leads to a vanishingly small spin polarization at the output leads . based on the symmetry arguments of the s - matrix elements kim _ et al . _
@xcite have shown analytically that in a y - shaped conductor subjected to so interaction the transmission amplitudes for the @xmath139 and @xmath58 components get equal magnitude and opposite phases for symmetrically connected leads which provide opposite spin polarization for these two components . while , for the @xmath142 component almost zero polarization is achieved with identical phase in the leads .
our numerical results for a three - terminal mesoscopic ring match very well with their analytical findings .
next , we focus our attention on the behavior of spin polarization components of outgoing currents considering a mesoscopic ring with only dresselhaus so interaction . the results are presented in fig .
[ poleng2 ] , where @xmath0 , @xmath1 and @xmath2 are computed for the same ring size ( @xmath143 ) and identical lead - ring interface geometry used in fig .
[ poleng1 ] , setting @xmath135 and @xmath144 .
interestingly , we see that the @xmath58 component alternates its sign in two output leads keeping the magnitude unchanged , while the other two components @xmath0 and @xmath1 interchange their features compared to the previous case i.e. , where the ring is described with rashba so interaction only .
below we justify these phenomena through simple analytical arguments . the tb hamiltonians ( @xmath145 and @xmath146 ) describing the rashba and dresselhaus so interaction terms can be transformed into each other by a simple unitary transformation using the matrix @xmath147 , i.e. , @xmath148 = @xmath146 , where @xmath149 = @xmath150 . therefore ,
if @xmath65 is the eigenstate of @xmath145 corresponding to a particular energy eigenvalue and @xmath151 is the eigenstate of the transformed hamiltonian @xmath146 , then we can write @xmath152 = @xmath153 . following this transformation the z component of spin polarization ( eq .
[ eqn20 ] ) in presence of only dresselhaus so interaction gets the form : @xmath154 this equation ( eq . [ eqn33 ] ) clearly illustrates the reason behind the sign reversal of @xmath2 of outgoing currents in two leads when the ring is described with only dresselhaus so term compared to the other case i.e. , the ring with only rashba so interaction . following the above prescription we can also get the relations @xmath155 since , @xmath156 these expressions ( eqs . [ eqn34 ] and [ eqn35 ] ) yield the reason for interchanging the features of @xmath0 and @xmath1 in the mesoscopic ring with only dresselhaus so coupling . finally , in figs .
[ polxr]-[polzr ] we present the variations of @xmath0 , @xmath1 and @xmath2 as a function of rashba so coupling for a typical energy @xmath157 considering the identical ring size used in figs . [ poleng1 ] and [ poleng2 ] .
the results are computed for two different values of @xmath28 to explore the combined effect of rashba and dresselhaus so interactions on the spin polarization components .
when @xmath136 , the components @xmath0 and @xmath2 provide outgoing currents with equal magnitude and opposite phases , while the other component @xmath1 almost drops to zero .
the situation is somewhat interesting when dresselhaus so coupling is included in addition to the rashba term . for the non - zero value of @xmath28 ,
all these three components get finite values in output leads and the magnitudes of individual components also differ in these two leads . these phenomena can be well explained from the analysis described above .
very interestingly we see that , at the particular case when the strengths of these two so couplings are identical i.e. , @xmath158 , the three polarization components drop to zero .
( shown by the green dashed lines in figs .
[ polxr]-[polzr ] ) .
it is already said that @xmath2 in presence of rashba coupling is just equal in magnitude but opposite in sign with @xmath2 in presence of dresselhaus coupling , so when both the interactions are present in equal strength net polarization must vanish .
for the other components we argue that when @xmath27 = @xmath28 , the total hamiltonian commutes with @xmath159 , so ( @xmath160 + @xmath161 ) i.e. , ( @xmath0+@xmath1 ) is a conserved quantity , i.e. , ( @xmath0+@xmath1 ) should have the same sign and same magnitude in both leads , which is possible only when both @xmath0 and @xmath1 are individually zero at both the outputs .
this phenomenon emphasizes that any one of the so fields can be predicted precisely if the other one is known .
needless to say , the precise determination of the so coupling strengths is extremely crucial in the field of spintronics .
one can control rashba so interaction by a suitable gate voltage , and hence , it can be measured . on the other hand , the possible techniques for the determination of dresselhaus
so coupling are relatively few .
recently , we have put forward some ideas of estimating so coupling strengths in a single sample which include the measurement of the minimum in drude weight @xcite which describes the conducting nature of the material , the observation of transmission resonance or anti - resonance of outgoing electrons @xcite and the determination of persistent spin current @xcite in an isolated loop geometry in presence of these two so interactions . in the present work
we give a separate proposal for it . for a single sample
if a finite polarization is achieved for a particular value of @xmath28 , then one can tune rashba so coupling by means of gate voltage to get vanishing spin polarization at the output currents .
thus , knowing the rashba so coupling , an accurate measurement of dresselhaus term is possible .
the spin polarization turns out to be sensitive to the lead - ring interface geometry . to this end
, we analyze the behavior of spin polarization components in a three - terminal mesoscopic ring where the output leads are coupled asymmetrically ( see fig . [ multiring ] ) with respect to the source lead . in fig .
[ polasymeng1 ] we display the variation of spin polarization components @xmath0 , @xmath1 and @xmath2 for a @xmath131-site ring described with rashba so interaction only , i.e. , setting @xmath28 at zero .
the red curves represent the results for the lead @xmath137 , while for the lead @xmath138 they are shown by the green lines . from the spectra
we see that , unlike the symmetric configuration , the magnitudes of @xmath0 and @xmath2 in two output leads are no longer identical even when one so coupling is set equal to zero .
this is solely due to the effect of quantum interference among the electronic waves passing through different arms of the mesoscopic ring .
all the other properties , for examples , the phase reversals of the polarization components @xmath0 and @xmath2 in two output leads and the vanishingly small amplitude of @xmath1 remain exactly same as discussed earlier in the case of symmetric configuration .
the same quantities are also analyzed for this asymmetric configuration considering the ring with only dresselhaus so interaction , i.e. , setting @xmath135 .
the results are given in fig .
[ polasymeng2 ] . except getting different amplitudes of polarizing currents in two output leads ,
all the other physical phenomena remain unchanged as discussed in fig .
[ poleng2 ] .
before we end this section , in figs .
[ polx]-[polz ] we describe the variations of @xmath0 , @xmath1 and @xmath2 as a function of rashba so coupling considering two different values of @xmath28 for this asymmetric lead - ring interface geometry to make the present communication a self contained study .
the amplitudes of the individual components of spin polarization in two output terminals get changed , as expected , and their magnitudes can also be tuned by including other so coupling .
the physical picture about the possibilities of determining so fields by observing the vanishing spin polarization of out going currents in the limit @xmath158 remains also invariant for this lead - ring interface geometry , like the symmetric configuration .
to conclude , in the present work we have described spin dependent transport through a multi - terminal mesoscopic ring in presence of rashba and dresselhaus so interactions . within a tb framework we have determined the polarization components @xmath0 , @xmath1 and @xmath2 of outgoing currents using a general spin density matrix formalism . the sensitivity of these components on the lead - ring interface geometry has also been analyzed in detail to make the communication is self contained study . from our extensive numerical work
we have established that a two - terminal mesoscopic ring with only so coupling , in absence of any magnetic impurity or external magnetic field , can not induce spin polarization in the output lead . on the other hand ,
a multi - terminal geometry , subjected to so interaction , containing at least two out put leads can generate polarized spin currents from a completely unpolarized electron beam even in absence of any magnetic field or magnetic like impurities .
finally , we have also provided a possible realization of determining dresselhaus so coupling knowing the rashba term in a single sample by observing the vanishing spin polarization of the outgoing currents , and hence facilitates a possible experimental measurement in this line .
finally we point out that , the presented results in this communication are also valid even for non - zero temperatures ( @xmath162k ) as the broadening of the energy levels caused by side - attached leads is much higher than the thermal broadening @xcite .
|
spin dependent transport in a multi - terminal mesoscopic ring is investigated in presence of rashba and dresselhaus spin - orbit interactions . within a tight - binding framework
we use a general spin density matrix formalism to evaluate all three components ( @xmath0 , @xmath1 and @xmath2 ) of the polarization vector associated with the charge current through the outgoing leads .
it explores the dynamics of the spin polarization vector of current propagating through the system subjected to the rashba and/or the dresselhaus spin - orbit couplings .
the sensitivity of the polarization components on the electrode - ring interface geometry is discussed in detail .
our present analysis provides an understanding of the coupled spin and electron transport in mesoscopic bridge systems .
|
it is well known that thermodynamic quantities exhibit singularities asymptotically close to the critical point .
the power - law behavior of these singularities , characterized by critical exponents and the concept of universality and scaling , have been successfully described by renormalization - group ( rg ) theory .
away from the asymptotic region , thermodynamic quantities of real physical systems deviate from simple power - law behavior
. however , rg theory can still provide insight in understanding critical crossover behavior .
there are two main field - theoretical renormalization - group schemes to treat critical - to - classical crossover phenomena .
dohm and co - workers developed the minimal - subtraction renormalization ( msr ) scheme @xcite while bagnuls and bervillier developed the massive renormalization ( mr ) scheme @xcite .
both of these theories used the borel resummation technique to describe the crossover behavior of the @xmath0 model in any @xmath3 universality class and in three dimensions .
the difference between the two schemes was discussed in ref.@xcite .
these field - theoretical crossover theories were improved over the years as asymptotic theories became more accurate @xcite .
recently , larin _
et al_. improved the msr expressions for the specific heat and compared their results with the superfluid helium @xmath4 system @xcite .
bagnuls and bervillier have also improved their theory to match the recent asymptotic values for exponents and leading amplitude ratios @xcite .
both renormalization schemes can provide crossover functional forms for thermal properties with a minimal set of fluid - dependent adjustable parameters .
however , a direct comparison between these recent theoretical predictions and different experimental measurements near the liquid - vapor critical point @xmath5 has been lacking . in this paper
we will present a direct comparison between the msr field theoretical crossover functions and various experimental measurements near the liquid - gas critical point of @xmath1he .
the comparison using the mr theory will be published elsewhere .
the paper is divided into two parts . in the first part
, we briefly summarize the msr functional expressions for susceptibility , specific heat , coexistence curve , and correlation length from previous work @xcite .
in addition , we derived within the msr framework new rg functional expressions for the asymptotic critical amplitudes of the susceptibility and coexistence curve as well as the first coefficient in a wegner expansion for susceptibility , specific heat , coexistence curve , and correlation length . from these expressions , universal amplitude ratios for the @xmath6 , 3-dimensional system are calculated and compared with the most recent values from bagnuls _
et al_. @xcite and fisher _ et al_. @xcite . the second part of the paper includes the results of msr functional fits to experimental measurements . in our previous work
, we analyzed the isothermal susceptibility of @xmath1he along the critical isochore above @xmath7 using theoretical expressions based upon the minimal - subtraction scheme @xcite . in this work
, we combine that analysis with susceptibility measurements along the coexistence curve and specific heat measurements along the critical isochore @xcite .
measurements of coexistence curve and the light scattering intensity near the critical point of @xmath1he @xcite are also analyzed .
the hamiltonian for the @xmath0 model in three dimensions @xmath8 is @xmath9where @xmath10 is the order parameter field , whose statistical mean value is the physical order parameter of a given system .
the parameter @xmath11 is the fourth - order coupling constant .
the parameter @xmath12 is related to the reduced temperature @xmath13 by @xmath14where @xmath15 is a nonuniversal constant .
it is important to note that the total hamiltonian is the sum of @xmath16 , where @xmath17 is the analytic background free energy .
since the liquid - vapor critical point has a single component order parameter and belongs to the @xmath6 universality class , we have @xmath18 . the dimensionless bare order parameter field @xmath10 and the bare coupling parameters @xmath11 and @xmath12 are renormalized to [ eqs .
( s2.11)and ( s2.12 ) @xcite ] @xmath19where @xmath20 is a geometric factor and @xmath21 for dimension @xmath22 .
the @xmath23 factors are associated with their respective field - theoretic functions @xcite @xmath24 \ , , \label{eq : beta_u_0}\end{aligned}\]]where the index 0 means differentiation at fixed @xmath12 , @xmath10 , and @xmath11 . by introducing a flow parameter @xmath25
, the effective coupling @xmath26 satisfies the flow equation @xmath27the flow parameter @xmath25 is related to the correlation length by @xmath28with @xmath29 being an arbitrary reference length .
the flow parameter @xmath30 corresponds to the ising fixed point @xmath31 , which is determined from @xmath32 .
the effective coupling @xmath33 satisfies the flow equation @xmath34the flow parameter @xmath35 is an arbitrary reference point , at which the nonuniversal initial values are @xmath36 and @xmath37 . the field - theoretic functions @xmath38 and @xmath39 in eqs .
( [ eq : zeta_r_0])-([eq : beta_u_0 ] ) are known up to five - loop order in perturbation expansions around @xmath40 @xcite .
however , the expansions do not converge away from @xmath41 .
hence borel resummations were used on the expansions to calculate the values of these functions over the range @xmath42 . for most investigations ,
see for instance ref.@xcite , only the function values at the fixed point @xmath43 were calculated using the borel resummations on the five - loop expansions .
the function values over the range @xmath44 were obtained using up to two - loop order expressions with extrapolation terms added in order to reproduce the values at the fixed point @xcite . for a system of dimension @xmath22 and single component order - parameter @xmath18 ,
one obtains @xmath45here @xmath46 through @xmath47 are the coefficients for the extrapolation terms with values listed in appendix [ dix : const ] . using these functions and the flow equations , thermal properties along the critical isochore and coexistence curve can be calculated from the asymptotic to crossover regions using the initial values for eqs .
( [ eq : flow_eqn_u ] ) and ( [ eq : flow_eqn_r ] ) , @xmath48 and @xmath49 , and the arbitrary length scale @xmath29 in eq .
( [ eq : l2xi ] ) . within the msr scheme
, the expression for the reduced temperature in terms of the flow parameter @xmath25 can be derived as follows .
the reduced temperature @xmath50 and the flow parameter @xmath25 can be linked using eqs .
( s4.25 ) , and ( s4.26 ) in ref.@xcite and eq .
( h2.9 ) of ref.@xcite , together with the solution of eq .
( [ eq : flow_eqn_r ] ) @xmath51with @xmath52here ` + ' is for @xmath53 and ` @xmath54 ' is for @xmath55 .
et al_. @xcite determined a one - loop expression plus a higher - order extrapolation for @xmath56 given by [ eq .
( k3.5 ) ] @xmath57where @xmath58 and @xmath59 are the extrapolation coefficients with the values given in table [ table : extrapolation ] in appendix [ dix : const ] . by adding and subtracting @xmath60 in the integrand of eq .
( [ eq : t2la ] ) and using the identity @xmath61 , where @xmath62 is the critical exponent of correlation length @xmath63 , one arrives at @xmath64by rearrange eq .
( [ eq : t2 lb ] ) , one obtains the following expression for the reduced temperature @xmath65\,,\]]and @xmath66 \\
\nonumber & = & \int_{u^*}^{{u}(l ) } { { \zeta_r(u^\prime ) - \zeta_r(u^ * ) } \over { \beta_u(u^\prime ) } } d u^\prime \,.\end{aligned}\ ] ] the following expressions for the dimensionless susceptibility @xmath67 were given in refs .
@xcite respectively for @xmath68 [ eq .
( k2.7 ) ] and @xmath69 [ eq . ( h2.16 ) ] , @xmath70the amplitude functions , @xmath71 , were expressed to two - loop order plus a higher - order extrapolation , [ eqs .
( k3.1 ) and ( h4.2 ) ] , to give @xmath72 & = & 1 - { 92\over 9 } u^2 ( 1 + b_\chi u ) \,\,\,\,\,\,\,\,\,\,\,\,\,\ , ( t > t_c ) , \\ \nonumber f_-[u ] & = & \left[1 - 18u + 159.56 u^2 ( 1 + d_\chi u)\right]^{-1 } \ , ( t < t_c),\,\end{aligned}\]]where @xmath73 and @xmath74 are the extrapolation coefficients with the values given in table [ table : extrapolation ] in appendix [ dix : const ] . the minimal renormalization factor , @xmath75 , in eq . ( [ eq : chi1 ] ) is given by [ eq .
( ka12 ) ] @xmath76the expression @xmath77 } \over { f_\pm ( { u}(l))}}\,\]]can be obtained by adding and subtracting @xmath78 in the integrand of eq .
( [ eq : chi1 ] ) , and using the relations @xmath79 @xcite and @xmath80 , where @xmath81 is the critical exponent of the fluctuation correlation at the critical point and @xmath82 is the critical exponent of susceptibility . in eq .
( [ eq : chi2 ] ) @xmath83 \,,\]]and @xmath84 \\ \nonumber & = & \int_{u^*}^{{u}(l ) } { { \zeta_\phi(u^\prime ) - \zeta_\phi(u^ * ) } \over { \beta_u(u^\prime ) } } d u^\prime \,.\end{aligned}\]]using eq .
( [ eq : l2xi ] ) and @xmath85 , eq .
( [ eq : chi2 ] ) can be rewritten as @xmath86 } \over { f_\pm ( { u}(l))}}\,.\]]thus , in the asymptotic regime ( @xmath87 ) @xmath88with a nonuniversal proportionality constant @xmath89 . within the pure @xmath0 model , the standard wegner expansion for the susceptibility is given by @xmath90 are the leading asymptotic critical amplitudes , @xmath91 are the first wegner expansion amplitudes above and below the transition , and @xmath92 is the correction - to - scaling exponent .
the details of the derivations of the leading and first wegner critical amplitudes are given in appendix [ dix : chi_amplitudes ] .
here we list the derived expressions for the critical amplitudes , @xmath93with @xmath94 and @xmath95
. the total specific heat is usually separated as @xmath96where the term @xmath97 represents an analytic background " contribution from the analytic background free energy @xmath17 , and @xmath98 represents the critical contribution from order parameter fluctuations . here ` + ' is for the specific heat above @xmath7 along the critical isochore , ` @xmath54 ' is for below @xmath7 in coexisting phases .
the critical specific heat @xmath98 derived from the hamiltonian expressed in eq .
( [ eq : hamiltonian ] ) has two representations within the msr scheme .
these two representations are derived via multiplicative and additive renormalization as detailed in ref .
the most recent work by larin _
et al_.@xcite used the representation via additive renormalization that we will use in this paper .
the critical specific heat @xmath98 per unit volume near @xmath7 is expressed by [ eqs .
( s2.36 ) or ( l3.3)]@xcite @xmath99the amplitude functions @xmath100 are given by @xmath101the functions @xmath102 for @xmath18 can be expressed by a two - loop calculation plus a higher - order extrapolation , [ eqs .
( k3.4 ) @xcite and ( h4.4)@xcite ] @xmath103 = \cases { -1 -6u(1+b_f u ) & ( $ t > t_c$ ) \cr ( 2u)^{-1 } -4(1+d_f u ) & ( $ t <
t_c$ ) \cr } \,,\]]where @xmath104 and @xmath105 are the extrapolation coefficients with the values given in table [ table : extrapolation ] in appendix [ dix : const ] .
the function @xmath106 in eq .
( [ eq : ku ] ) is governed by @xmath107with @xmath108 . the function @xmath109 has been calculated to @xmath110 for any given @xmath111 [ eq . ( l2.21 ) ] @xcite . however
the five - loop borel resummation of @xmath109 was only performed for @xmath112 at @xmath43 .
hence a new extrapolation term with a coefficient @xmath113 is added to the two - loop expression @xcite in order to satisfy @xmath114 , @xmath115at the fixed point @xmath116 , @xmath117 since @xmath32 , and eq . ( [ eq : au_diff ] ) leads to @xmath118where @xmath119 @xcite is used with @xmath120 being the critical exponent for specific heat at constant volume . the integral in the exponential of eq .
( [ eq : c_bare ] ) can be rewritten as @xmath121using @xmath119 .
the expression for the specific heat from the additive renormalization can now be rewritten as @xmath122 \exp \left[-2f_r(1 ) \right ] \\ & = & c_0 l^{-\alpha /\nu } \exp \left [ 2f_r(l ) \right]k_\pm({u}(l ) ) \,,\end{aligned}\]]where @xmath123 is given by eq .
( [ eq : fr ] ) and @xmath124 is defined as @xmath125 = \frac{\mu^3}{16\pi t_0 ^ 2}\,.\ ] ] the standard wegner expansion within the pure @xmath0 model for specific heat can be written as @xmath126where @xmath127 is a constant background induced by long - range correlations between the fluctuations .
the experimentally measured constant background is the sum of @xmath127 and the analytic background , @xmath128 .
the expression for the wegner expansion of the specific heat via multiplicative renormalization was derived and given in eqs .
( s4.23 ) and ( s4.24 ) of ref .
@xcite . in appendix
[ dix : cv_amplitudes ] , we derive the expressions for the critical amplitudes and the critical background , @xmath127 , for the representation via additive renormalization , using the technique that is consistent with the one used for susceptibility .
the results of these derivations are , @xmath129@xmath130\right|_{u^ * } { { u^ * - u}\over { \left(b^*_\pm\ , t_0\right)^{\delta}}}\,,\end{aligned}\]]@xmath131the variables with a prime in eqs .
( [ eq : a_1 ] ) and ( [ eq : bcr_full ] ) are derivatives with respect to @xmath132 . the right hand side of eq .
( [ eq : bcr_full ] ) is negative for any given @xmath132 .
hence one has @xmath133 since @xmath134 from eq .
( [ eq : c_0 ] ) . in the liquid - vapor
coexisting phases below @xmath7 , the density difference @xmath135 is the statistical mean of the order parameter field @xmath136 .
there is no asymmetry between @xmath137 and @xmath138 within the @xmath0 model .
schloms and dohm have given an expression for the square of the physical order parameter [ eq .
( s3.10 ) ] @xcite , @xmath139here @xmath140 is given in eq .
( [ eq : zphi ] ) .
the amplitude function @xmath141 is expanded in one - loop with an extrapolation term , [ eq . ( h4.1)]@xcite , to yield @xmath142the correlation length below @xmath7 is linked to the flow parameter by @xmath143 . by combining these expressions and following the derivation of eq .
( [ eq : chi2 ] ) , one has @xmath144\,,\]]where @xmath145 is used and @xmath146\,,\]]with @xmath147 being the critical exponent of the order parameter . using eq .
( [ eq : t2lc ] ) to replace @xmath148 in eq .
( [ eq : phisqr2 ] ) and the scaling relations @xmath149 , @xmath150 , and @xmath151 , one has @xmath152\exp [ -f_\phi(l_-)/2]\,.\end{aligned}\]]expanding @xmath153 , @xmath154 , @xmath155 , and @xmath156 in the same manner as described in appendix [ dix : chi_amplitudes ] , one obtains the wegner expansion for coexistence curve , @xmath157with the leading critical amplitude and the first wegner amplitude being respectively @xmath158 using eq .
( [ eq : t2lc ] ) to express @xmath25 in terms of @xmath159 , the expression for dimensionless correlation length ( if @xmath29 is taken dimensionless ) is derived from eq .
( [ eq : l2xi ] ) as @xmath160^{\nu } \ , \exp[-\nu f_r(l ) ] \,.\]]an expansion of eq .
( [ eq : correlation_length ] ) around @xmath161 to @xmath162}$ ] leads to @xmath163 ) to the standard wegner expansion form , @xmath164one obtains the leading amplitudes and first wegner correction amplitudes of the correlation length @xmath165 even though the leading amplitude and subsequent wegner expansion coefficients are fluid - dependent , certain combination ratios of these amplitudes are universal . from the equations for the first wegner amplitudes of the specific heat ,
susceptibility , coexistence curve , and correlation length , one notices that the system - dependent part , @xmath166 , is the same in every expression
. therefore the ratio of any of these first wegner amplitudes is universal based on the msr @xmath0 model .
these universal ratios have been given for the specific heat in ref .
@xcite , [ eqs .
( [ eq : ratio_a_0 ] ) , ( [ eq : ratio_a_1 ] ) , and ( [ eq : a_0_xi_0 ] ) ] . in this paper
we derive the other universal ratios based on the msr @xmath0 model . from eqs .
( [ eq : gamma_0 ] ) and ( [ eq : gamma_1 ] ) , one has the universal amplitude ratios for susceptibility , @xmath167@xmath168from eqs .
( [ eq : a_0 ] ) and ( [ eq : a_1 ] ) , one obtains the universal amplitude ratios for specific heat , @xmath169@xmath170 use of the scaling relation @xmath151 and the combination of eqs .
( [ eq : gamma_0 ] ) , ( [ eq : a_0 ] ) , and ( [ eq : b_0 ] ) leads to a universal ratio @xmath171 from eqs .
( [ eq : xi_0 ] ) and ( [ eq : xi_1 ] ) , the universal amplitude ratios for the correlation length are , @xmath172@xmath173 from eqs .
( [ eq : a_0 ] ) and ( [ eq : xi_0 ] ) , one has the universal relation between the amplitude of specific heat and correlation length @xmath174where the scaling relation @xmath175 has been used .
equation ( [ eq : a_0_xi_0 ] ) is identical to eq .
( s4.22 ) of ref.@xcite .
the evaluation of the right - hand side uses the constants given in appendix [ dix : const ] . a natural extension of eq .
( [ eq : a_0_xi_0 ] ) is the universal relation between specific heat and the correlation length throughout the crossover region . using eqs .
( [ eq : l2xi ] ) , ( [ eq : t2lc ] ) , and ( [ eq : c_bare_2 ] ) , and the scaling relation @xmath175 , one has @xmath176since there is no fluid - dependent parameters appearing on the right hand side of eq .
( [ eq : cvxiuniversal ] ) , the product of the critical specific heat and the cubic of the correlation length is universal for any given temperature throughout the crossover region .
table [ table : ratios ] lists the various universal amplitude ratios derived from the minimal - subtraction renormalization scheme , bagnuls and bervillier s massive - renormalization scheme @xcite , and other methods , such as @xmath177-expansion , summarized by fisher and zinn @xcite .
the values given by bagnuls and bervillier are closely matched to the values given by guida and zinn - justin @xcite after the readjustment of the borel resummation criteria @xcite .
noticeable differences exist in table [ table : ratios ] among various theories . in attempting to explain these differences , two factors are important to note .
first , we are unable to evaluate the uncertainties of the universal ratios since the uncertainties on the borel resummations at the fixed point @xmath43 for most of the amplitude functions were not given in previous studies .
secondly , eqs .
( [ eq : gamma_1 ] ) , ( [ eq : a_1 ] ) , and ( [ eq : xi_1 ] ) use the derivatives of eqs .
( [ eq : zeta_r ] ) , ( [ eq : zeta_phi ] ) , ( [ eq : qu1loop ] ) , ( [ eq : amp_chit ] ) , ( [ eq : amp_cv ] ) , and ( [ eq : b_u ] ) which could have sizable systematic uncertainties .
these equations were only obtained from two - loop calculations and extrapolated to the five - loop fixed point values with adjustable constants .
hence it is desirable to have these derivatives calculated at the fixed point with borel resummations
. then the extrapolation coefficients can be more accurately reconstructed , leading to the estimates of the first wegner coefficients with less uncertainties .
.[table : ratios]the values of various universal amplitude ratios .
the calculation for this work uses the values of the amplitude functions at the fixed point @xmath43 given in table and the values of the critical exponents given by guida and zinn - justin @xcite . [ cols="^,^,^,^ " , ] close to @xmath7 , the susceptibility data for @xmath178 and the specific heat data for both @xmath55 and @xmath53 deviate slightly from the theoretical prediction .
these deviations can be attributed to a gravity - induced density stratification . since the specific heat was measured as an average of the whole cell while the susceptibility was measured locally across a density sensor
, there was a stronger gravity effect in the measured @xmath179 than @xmath180 .
the gravity effect on @xmath181 is about a factor of five smaller than that on @xmath182 because of the difference in @xmath180 magnitudes .
when @xmath7 is used as an adjustable parameter , the individual fits of susceptibility and specific heat tend to skew @xmath7 such that the difference between the experimental measurements and theoretical prediction is minimized because of the shift in reduced temperature for the measurements .
the @xmath7 determined from the fits of the specific heat ( fig .
[ fig : cv_leveled ] ) and susceptibility data ( fig .
[ fig : chi_t_leveled ] ) tends to be higher and lower , respectively , than it should be .
this tendency was approximately cancelled out in the joint fit shown in fig .
[ fig : triple_joint_leveled ] . the slight gravity effect on the experimental measurements for @xmath183
can be clearly seen in fig .
[ fig : triple_joint_leveled ] .
we mention that in ref .
@xcite earlier measurements of the susceptibility of @xmath1he , both above and below @xmath7 , were compared with the present data .
also in table i of that reference , the amplitudes of susceptibility and coexistence curve data and their ratios , such as @xmath184 and @xmath185 , obtained from individual fits , were presented . by using @xmath186 , @xmath187 , and @xmath188
given in table [ table : joint_fit ] , the dimensionless correlation length can be calculated for any given @xmath159 using eq .
( [ eq : correlation_length ] ) . figure [ fig : xi_leveled ] shows the dimensionless correlation length versus @xmath50 calculated from the msr @xmath0 model for @xmath1he .
the length scale to recover the dimensional @xmath189 is @xmath190 , given by eq .
( [ eq : l0 ] ) .
thus one has a dimensional @xmath191 .
this value can be directly compared with @xmath192 measured in an acoustic experiment ref .
@xcite . considering that the experimental @xmath189 had 10% uncertainty , the agreement is very good .
model for @xmath1he .
the dot - dashed straight lines represent the asymptotic predictions from the fit.,width=297 ] the correlation length can also be determined from a light scattering experiment .
miura , meyer , and ikushima measured the intensity of scattered light of @xmath1he fluid near its critical point @xcite . the intensity scattered per unit beam length per unit solid angle in the fluid , @xmath193 , is given by @xmath194where @xmath195 is the beam intensity in the scattering region , @xmath196 is the angle between the electric field of the incident light and the wave vector of the scattered light , @xmath180 is the susceptibility , and @xmath197 . here
@xmath111 is the index of refraction of the fluid , and @xmath198 is the vacuum wavelength of the incident light .
the function @xmath199 is , for @xmath200 , very accurately given by the ornstein - zermike approximation @xmath201 , where @xmath202 is the scattering wave vector , @xmath63 is the correlation length , and @xmath81 is the critical exponent of the fluctuation correlation at the critical point . in ref .
@xcite , @xmath203 @xmath204 . at @xmath205 ,
the value of correlation length can be estimated from @xmath206 cm , hence the condition @xmath207 was satisfied for @xmath208 .
since @xmath209 is essentially a constant for the experimental condition , one can use the knowledge of @xmath180 and @xmath63 , based on the msr @xmath0 model , to fit experimental data of the scattered intensity , with @xmath210 as an adjustable parameter . as it can be seen in fig .
[ fig : light_scattered ] , the agreement between the experimental data and theoretical calculation is reasonably good . the intensity of light scattered by @xmath1he versus reduced temperature .
the theoretical calculated @xmath193 , using the msr @xmath0 model , is adjusted with a constant amplitude for @xmath211 while @xmath186 , @xmath187 , and @xmath188 were fixed from the values given in table .,width=297 ]
in this paper we have used parametric expressions to calculate the isothermal susceptibility , specific heat , coexistence curve , and correlation length along the critical isochore and coexistence curve from the asymptotic region to the crossover region .
all the critical leading amplitude ratios were contained in the model as listed in table [ table : ratios ] . using only two adjustable parameters in these theoretical expressions for the critical contributions , we fit the theory to recently obtained experimental data for the isothermal susceptibility , specific heat , and early experimental data of the coexistence curve and light scattering intensity .
the agreement between the theory and experimental measurements is good .
further improvements to the minimal renormalization scheme are desired , especially the five - loop borel resummations throughout the whole range of @xmath212 .
more accurate borel resummations at the fixed point should also lead to improved calculations of @xmath213 and @xmath214 so that the resultant critical exponents can be compared with other published values ( see appendix [ dix : const ] ) .
theoretical insights on non - critical contributions are also needed in order to formulate more accurate analytical expressions for the background contributions . while the present minimal subtraction renormalization model describes quite well to experimental measurements along the critical isochore and coexistence curve , it is not as yet a model for a complete equation - of - state .
recently , agayan _
et al_. have developed a phenomenological crossover parametric model ( cpm ) equation - of - state that is also based upon rg theory @xcite . within this model ,
the internal constants were adjusted such that the critical leading amplitude ratios agreed with the values in the fisher and zinn column in table [ table : ratios ] .
this cpm model was developed to fit simple fluids as well as complex fluid systems that exhibit non - monotonic crossover behavior .
this non - monotonic crossover behavior could be described by the cpm approach using a finite cut - off wavelength as an additional fitting parameter . however , in simple fluid systems , like @xmath1he , crossover behavior of different physical quantities can be well described within the framework of the field theoretical @xmath0 model without the finite cut - off wavenumber .
nasa supported microgravity flight experiments @xcite ( http://miste.jpl.nasa.gov ) , which are under preparation , will take experimental data of the susceptibility , specific heat , and coexistence curve in the asymptotic region .
combining these microgravity measurements in the asymptotic region with ground - based measurements in the crossover region should permit a rigorous test of the predictions of recent renormalization theories .
we are indebted to dr .
r. haussmann and prof .
j. rudnick for supporting the early development of this work and for many stimulating discussions .
we are also grateful to prof .
h. meyer for a critical reading of the manuscript and to dr .
m. weilert for his contribution in performing the experiments .
the research described in this paper was carried out at the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the field - theoretic functions , @xmath215 , @xmath216 , and @xmath39 , and the amplitude functions , @xmath217 , @xmath56 , @xmath218 , @xmath219 , @xmath106 , and @xmath109 are known up to five - loop order from expansions around @xmath40
. however , these expansions do not converge . to overcome this difficulty ,
these quantities were expanded to two - loop order and then extrapolation terms were added to have the functions agree with the calculations of high - order borel resummations at the fixed point @xcite .
all these functions have at least one extrapolation term to match the function s value at the fixed point @xmath43 ; some functions also have a second extrapolation term in order to match the value of its derivative at the fixed point .
listed in this appendix are the values of these extrapolation coefficients , their origins , and recent improvements .
the effects of these coefficient values on the critical exponents and the fitting quality in this work are discussed .
the extrapolation coefficients for the field - theoretic functions , @xmath215 , @xmath216 , and @xmath39 in eqs .
( [ eq : zeta_r_0 ] ) , ( [ eq : zeta_phi_0 ] ) , ( [ eq : beta_u_0 ] ) are @xmath220 , @xmath221 , @xmath222 , @xmath223 , and @xmath224 .
they are taken from table 2 of ref.@xcite .
the fixed point value for @xmath225 is solved from the condition @xmath32 using the given values for @xmath226 and @xmath47 .
the latest published @xmath43 value for @xmath112 is @xcite @xmath227 the asymptotic critical exponents are linked to the exponent functions @xmath228 , @xmath229 , and @xmath230 by @xmath231 @xmath232^{-1 } = \left(2 - \zeta_r^*\right)^{-1 } = 0.629\ , , \label{eq : nu}\ ] ] @xmath233 once the critical exponents , @xmath81 , @xmath62 , and @xmath234 are known , the remaining important critical exponents can be obtained from scaling using @xmath235 @xmath236 @xmath237 @xmath238 for @xmath112 , the latest theoretically calculated critical exponents given by guida and zinn - justin @xcite are , @xmath239 @xmath240 @xmath241 @xmath242 @xmath243 @xmath244 @xmath245 a clear difference exits for the value of the critical exponent @xmath82 which warrants further efforts from the theoretical community for improvements in the msr @xmath0 model calculation .
the amplitude function @xmath246 for reduced temperature is expressed as @xmath247at the fixed point @xmath43 , there is an identity @xmath248 that simplifies eq .
( [ eq : qu ] ) and leads to @xmath249krause _ et al_. @xcite obtained the expression [ eq .
( k_a28 ) ] @xmath250with @xmath251 .
they also provided a one - loop expression for @xmath252 using a higher - order approximation [ eq .
( k3.2 ) ] @xcite @xmath253the latest calculation by larin _
et al_. @xcite for @xmath112 gives @xmath254if the theoretically calculated critical exponent @xmath62 for @xmath255 , given by guida and zinn - justin @xcite , is used @xcite , one has @xmath256for the extrapolation coefficients in the expression for @xmath56 , eq .
( [ eq : qu1loop ] ) , krause _ et al_. @xcite determined @xmath257 and @xmath258 such that eq .
( [ eq : qustar ] ) and ( [ eq : dqdu ] ) were satisfied with the then calculated @xmath259 .
the values of the extrapolation coefficients @xmath260 and @xmath261 have been readjusted to agree with the new @xmath262 .
there is no value for @xmath263 , so the new @xmath264 has been fixed to its old value @xcite . for the amplitude function @xmath218 in the expression of the susceptibility , eq .
( [ eq : chi2 ] ) , @xmath265 comes from table 1 of ref.@xcite and @xmath266 comes from table 4 of ref .
@xcite . for the amplitude function in the expression of the specific heat , eqs .
( [ eq : c_bare_2 ] ) , ( [ eq : ku ] ) , and ( [ eq : amp_cv ] ) , the five - loop approximation with a borel resummation gives @xcite @xmath267 by combining eqs .
( [ eq : amp_cv])b , ( [ eq : ustar ] ) , and ( [ eq : f_minus ] ) , the old interpolation coefficient , @xmath268 ( table 4 of ref.@xcite ) , becomes @xmath269 .
the latest five - loop calculation also gives @xcite @xmath270= 0.4170 \pm 0.0036\ , .
\label{eq : f_difference}\ ] ] using eqs .
( [ eq : f_minus ] ) and ( [ eq : f_difference ] ) , one has @xmath271 by combining eqs .
( [ eq : amp_cv])a , ( [ eq : ustar ] ) , and ( [ eq : f_plus ] ) , the old interpolation coefficient , @xmath272 ( table 1 of ref.@xcite ) , becomes @xmath273 .
we modify eq .
( [ eq : b_u ] ) to be @xmath274with @xmath275 in order to satisfy the five - loop borel resumed results [ eq . ( l2.34 ) ] @xcite @xmath276all the calculations use the value of @xmath43 derived in this paper .
table [ table : extrapolation ] lists the values of the various extrapolation coefficients for the amplitude functions in the msr @xmath0 model .
table [ table : fixedpoint ] lists the values of the various amplitude functions at the fixed point @xmath43 .
[ c]c|c|c coefficient & value & appeared in +
@xmath46 & @xmath277 & @xmath215 for @xmath278 + @xmath279 & @xmath280 & @xmath215 for @xmath278 + @xmath281 & @xmath282 & @xmath216 for @xmath140 + @xmath226 & @xmath283 & @xmath39 for @xmath284 + @xmath47 & @xmath285 & @xmath39 for @xmath284 + @xmath58 & @xmath286 & @xmath56 for @xmath287 + @xmath59 & @xmath288 & @xmath56 for @xmath287 + @xmath73 & @xmath289 & @xmath290 for @xmath291 + @xmath74 & @xmath292 & @xmath293 for @xmath294 + @xmath104 & @xmath295 & @xmath296 for @xmath297 + @xmath105 & @xmath298 & @xmath299 for @xmath300 + @xmath113 & @xmath301 & @xmath109 for @xmath302 + @xmath303 & @xmath304 & @xmath141 for @xmath305 + [ c]c|c|c coefficient & value & appeared in + @xmath306 & @xmath307 & @xmath287 in eq .
( ) + @xmath308 & @xmath309 & @xmath287 in eq .
( ) + @xmath310 & @xmath311 & @xmath312 for @xmath291 in eq .
( ) + @xmath313 & @xmath314 & @xmath315 for @xmath294 in eq .
( ) + @xmath316 & @xmath317 & @xmath318 for @xmath319 in eq .
( ) + @xmath320 & @xmath321 & @xmath322 for @xmath323 in eq .
( ) + @xmath324 & @xmath325 & @xmath326 for @xmath327 in eq .
( ) + @xmath328 & @xmath329 & @xmath330 for @xmath305 in eq .
( ) + equations ( [ eq : t2lc ] ) and ( [ eq : chi2 ] ) provide a clear identification of the leading critical divergence and crossover contribution in a multiplicative form . in the original expressions , the critical divergence
is contained implicitly in the integrals of @xmath228 and @xmath229 in eqs .
( [ eq : t2la ] ) and ( [ eq : chi1 ] ) . the calculated @xmath214 and @xmath213 using borel resummations at the fixed point lead to the critical exponents @xmath62 and @xmath81 that are slightly different from the latest values given by guida and zinn - justin . because of the expressions in eqs .
( [ eq : t2 lb ] ) and ( [ eq : chi2 ] ) , the critical exponent values given by guida and zinn - justin are used for the leading divergence .
the inconsistency is only in the crossover part in the integrands of @xmath331 $ ] and @xmath332 $ ] that go to zero as the fixed point is approached .
expressions for the wegner expansion of the susceptibility will be derived in this appendix that were not presented in previously published work .
multiplying eq .
( [ eq : chi2 ] ) by eq .
( [ eq : t2lc ] ) to the power @xmath82 yields @xmath333^\gamma \ , { { \exp [ -f_\phi(l ) - \gamma f_r(l ) ] } \over { f_\pm ( l)}}\,.\]]in order to expand the exponent functions , @xmath334 and @xmath335 , based on eqs .
( [ eq : fr ] ) and ( [ eq : fphi ] ) , one needs to expand first the function for the flow equation , @xmath336 , to the first order in @xmath337 $ ] , @xmath338 + o[({u}(l ) - u^*)^2]\,,\]]where @xmath95 and @xmath32 . since @xmath339 and @xmath340 , one obtains @xmath341 } } \ , [ { u}(l ) - u^ * ] \\ & + & o[({u}(l ) - u^*)^2 ] \\
\nonumber & = & { { \zeta_r^\prime(u^*)}\over{\omega}}[{u}(l ) - u^ * ] + o[({u}(l ) - u^*)^2]\,,\end{aligned}\]]@xmath342 } } \ , [ { u}(l ) - u^ * ] \\ & + & o[({u}(l ) - u^*)^2 ] \\
\nonumber & = & { { \zeta_\phi^\prime(u^*)}\over{\omega}}[{u}(l ) - u^ * ] + o[({u}(l ) - u^*)^2]\,,\end{aligned}\]]@xmath343 \\ & + & o[({u}(l ) - u^*)^2]\,,\end{aligned}\]]@xmath344 \\ & + & o[({u}(l ) - u^*)^2]\,.\end{aligned}\]]the expression @xmath345 / f_\pm ( { u}(l))$ ] is then expanded in terms of @xmath346 $ ] , dropping the higher orders , to give @xmath347\right|_{u^ * } \ , [ { u}(l ) - u^ * ] \right\}\,.\nonumber\end{aligned}\]]the solution of the flow equation with @xmath39 approximated by eq .
( [ eq : beta_expanded ] ) is @xmath348by expressing @xmath25 in terms of @xmath159 and dropping higher order terms , one has @xmath349 and @xmath350\,.\end{aligned}\]]comparing eq .
( [ eq : chi5 ] ) to the standard wegner expansion to the first term , see eq .
( [ eq : chi_wegner ] ) , one obtains the critical amplitudes of the susceptibility expressed analytically in eqs .
( [ eq : gamma_0 ] ) and ( [ eq : gamma_1 ] ) with @xmath351 .
a derivation of the critical amplitudes and constant background of the specific heat in the additive renormalization form will be given in this appendix .
this derivation is consistent with the one for susceptibility given above and is different from the one given by schloms and dohm @xcite .
first an expansion expression for the function @xmath352 will be derived that is an approximate solution of eq .
( [ eq : au_diff ] ) . by expanding @xmath109 and
@xmath215 around @xmath43 and omitting higher order terms beyond the linear term , eq .
( [ eq : au_diff ] ) becomes @xmath353 a(l)\,.\end{aligned}\]]then eq .
( [ eq : l_expanded ] ) is used to replace @xmath354 with @xmath355 , yielding @xmath356where @xmath357with a variable change of @xmath358eq .
( [ eq : au_diff_approx_2 ] ) becomes @xmath359the solution of eq .
( [ eq : au_diff_approx_3 ] ) is @xmath360\,,\end{aligned}\]]where @xmath361 is a constant to be determined through the initial condition .
expanding eq .
( [ eq : au_solution_1 ] ) in @xmath362 and keeping only the linear terms of @xmath363 and @xmath355 , one obtains @xmath364where eq . ( [ eq : a_star ] )
is used for @xmath365 and @xmath366
\\\nonumber & = & \frac{2\nu}{\delta - \alpha}\left[2b^\prime(u^ * ) - a^*\zeta _ r^\prime(u^*)\right]\,.\end{aligned}\]]@xmath367 in eq . (
[ eq : au_solution_2 ] ) will be eliminated through initial condition at the reference point @xmath368 @xmath369where @xmath370 .
since @xmath371 has not been given as a fitting parameter , it is calculated from the numerical solution of eq .
( [ eq : au_diff ] ) with @xmath372 . substituting eq .
( [ eq : au_solution_2 ] ) into eq .
( [ eq : c_bare_2 ] ) , one has @xmath373 k_3 + \exp [ 2 f_r(l ) ] l^{-\alpha/\nu}\\ & \times & \left [ f_\pm(l ) - a^ * - k_2 ( u - u^*)l^\omega \right]\,.\end{aligned}\]]replacing @xmath374 in eq .
( [ eq : c_bare_4 ] ) with @xmath375 from eq .
( [ eq : t2lc ] ) leads to @xmath376k_3 \\ \nonumber & + & \exp [ ( 2-\alpha ) f_r(l ) ] \left[b_\pm(l ) t_0\right]^\alpha u^*)l^\omega \right]\,.\end{aligned}\]]@xmath123 , @xmath377 , and @xmath378 are expanded according to eqs .
( [ eq : fr_expanded ] ) , ( [ eq : ampl_chi_expanded ] ) , and ( [ eq : ampl_t_expanded ] ) respectively .
higher order terms than @xmath379 $ ] or @xmath380 are dropped in the expansion , and @xmath381 is replaced using eq .
( [ eq : l_expanded ] ) . by using the approximation of @xmath382 , one finally has @xmath383\right|_{u^ * } \\ & \times & \left .
( u - u^*){{|t|^\delta}\over{\left(b_\pm^ * t_0\right)^\delta } } \right\ } \,.\end{aligned}\]]by comparing eq .
( [ eq : c_bare_expanded ] ) to the standard wegner expansion to the first term , see eq .
( [ eq : cv_wegner ] ) , one obtains the analytical expressions for the critical amplitudes of the specific heat given in eqs .
( [ eq : a_0 ] ) and ( [ eq : a_1 ] ) .
the critical background specific heat is also identified as @xmath384
we list in this appendix the dimensionless experimental measurements of isothermal susceptibility , specific heat , and coexistence curve of @xmath1he .
the its90 temperature standard was used in the following tables . + @xmath387 & @xmath388 & @xmath385 & index + 3.3157200 & 5.288e-05 & 2.763e+04 & 0 + 3.3158500 & 9.209e-05 & 1.524e+04 & 0 + 3.3158770 & 1.002e-04 & 1.278e+04 & 0 + 3.3159900 & 1.343e-04 & 9.468e+03 & 0 + 3.3161700 & 1.886e-04 & 6.006e+03 & 0 + 3.3161770 & 1.907e-04 & 5.600e+03 & 0 + 3.3163600 & 2.459e-04 & 4.477e+03 & 0 + 3.3164000 & 2.580e-04 & 4.457e+03 & 0 + 3.3166570 & 3.355e-04 & 2.966e+03 & 0 + 3.3173270 & 5.376e-04 & 1.670e+03 & 0 + 3.3175600 & 6.078e-04 & 1.469e+03 & 0 + 3.3179100 & 7.134e-04 & 1.256e+03 & 0 + 3.3179170 & 7.155e-04 & 1.180e+03 & 0 + 3.3179400 & 7.225e-04 & 1.194e+03 & 0 + 3.3180170 & 7.457e-04 & 1.120e+03 & 0 + 3.3185400 & 9.034e-04 & 9.472e+02 & 0 + 3.3188970 & 1.011e-03 & 7.790e+02 & 0 + 3.3202400 & 1.416e-03 & 5.268e+02 & 0 + 3.3221970 & 2.006e-03 & 3.450e+02 & 0 + 3.3239900 & 2.547e-03 & 2.587e+02 & 0 + 3.3259870 & 3.149e-03 & 1.970e+02 & 0 + 3.3295500 & 4.224e-03 & 1.407e+02 & 0 + 3.3341869 & 5.623e-03 & 9.850e+01 & 0 + 3.3404500 & 7.512e-03 & 7.046e+01 & 0 + 3.3487469 & 1.001e-02 & 4.980e+01 & 0 + 3.3487600 & 1.002e-02 & 4.979e+01 & 0 + 3.3596699 & 1.331e-02 & 3.535e+01 & 0 + 3.3745467 & 1.780e-02 & 2.480e+01 & 0 + 3.3897402 & 2.238e-02 & 1.884e+01 & 0 + 3.3999300 & 2.545e-02 & 1.642e+01 & 0 + 3.3999967 & 2.547e-02 & 1.620e+01 & 0 + 3.4402800 & 3.762e-02 & 1.014e+01 & 0 + 3.4999362 & 5.561e-02 & 6.540e+00 & 0 + 3.4999899 & 5.563e-02 & 6.540e+00 & 0 + 3.6000160 & 8.580e-02 & 3.990e+00 & 0 + 3.6000502 & 8.581e-02 & 3.939e+00 & 0 + 3.8000406 & 1.461e-01 & 2.126e+00 & 0 + 3.3149265 & -1.864e-04 & 1.151e+03 & 19 + 3.3144595 & -3.273e-04 & 8.291e+02 & 19 + 3.3137295 & -5.475e-04 & 3.271e+02 & 19 + 3.3135883 & -5.901e-04 & 3.665e+02 & 19 + 3.3125410 & -9.059e-04 & 2.145e+02 & 19 + 3.3122627 & -9.899e-04 & 1.868e+02 & 19 + 3.3096911 & -1.765e-03 & 9.418e+01 & 19 + 3.3050875 & -3.154e-03 & 4.708e+01 & 19 + 3.2969714 & -5.602e-03 & 2.344e+01 & 19 + 3.2824717 & -9.975e-03 & 1.225e+01 & 19 + 3.2630551 & -1.583e-02 & 7.408e+00 & 19 + 3.2566732 & -1.776e-02 & 6.555e+00 & 19 + 3.2323114 & -2.510e-02 & 4.455e+00 & 19 + 3.1836186 & -3.979e-02 & 2.696e+00 & 19 + 3.1064086 & -6.308e-02 & 1.717e+00 & 19 + 3.3144595 & -3.273e-04 & 6.347e+02 & 8 + 3.3137295 & -5.475e-04 & 3.326e+02 & 8 + 3.3135883 & -5.901e-04 & 3.080e+02 & 8 + 3.3125410 & -9.059e-04 & 1.820e+02 & 8 + 3.3122627 & -9.899e-04 & 1.436e+02 & 8 + 3.3096911 & -1.765e-03 & 9.300e+01 & 8 + 3.3051276 & -3.142e-03 & 4.865e+01 & 8 + 3.3050875 & -3.154e-03 & 4.533e+01 & 8 + 3.2969591 & -5.606e-03 & 2.937e+01 & 8 + 3.2824590 & -9.979e-03 & 1.586e+01 & 8 + 3.2630538 & -1.583e-02 & 6.585e+00 & 8 + 3.2323290 & -2.510e-02 & 4.100e+00 & 8 + 3.1836229 & -3.979e-02 & 2.486e+00 & 8 + 3.1064148 & -6.308e-02 & 1.600e+00 & 8 + 2.9840784 & -9.997e-02 & 9.556e-01 & 8 + @xmath387 & @xmath388 & @xmath386 + 3.014570 & -9.078e-02 & 7.881 + 3.021686 & -8.863e-02 & 7.946 + 3.028699 & -8.652e-02 & 7.959 + 3.035582 & -8.444e-02 & 8.038 + 3.042383 & -8.239e-02 & 8.088 + 3.049135 & -8.035e-02 & 8.069 + 3.055774 & -7.835e-02 & 8.173 + 3.059940 & -7.709e-02 & 8.183 + 3.064174 & -7.582e-02 & 8.200 + 3.070900 & -7.379e-02 & 8.231 + 3.077578 & -7.177e-02 & 8.312 + 3.081723 & -7.052e-02 & 8.331 + 3.085888 & -6.927e-02 & 8.332 + 3.092441 & -6.729e-02 & 8.541 + 3.099071 & -6.529e-02 & 8.471 + 3.103159 & -6.406e-02 & 8.428 + 3.104792 & -6.357e-02 & 8.498 + 3.106435 & -6.307e-02 & 8.512 + 3.108071 & -6.258e-02 & 8.498 + 3.109702 & -6.208e-02 & 8.543 + 3.111329 & -6.159e-02 & 8.578 + 3.112949 & -6.110e-02 & 8.624 + 3.114572 & -6.062e-02 & 8.528 + 3.115721 & -6.027e-02 & 8.561 + 3.116924 & -5.991e-02 & 8.622 + 3.120951 & -5.869e-02 & 8.631 + 3.130580 & -5.579e-02 & 8.680 + 3.143237 & -5.197e-02 & 8.798 + 3.155746 & -4.820e-02 & 8.902 + 3.165016 & -4.540e-02 & 9.008 + 3.171098 & -4.357e-02 & 9.061 + 3.177146 & -4.174e-02 & 9.118 + 3.182946 & -3.999e-02 & 9.177 + 3.188913 & -3.819e-02 & 9.248 + 3.190517 & -3.771e-02 & 9.231 + 3.191638 & -3.737e-02 & 9.257 + 3.198331 & -3.535e-02 & 9.419 + 3.210298 & -3.174e-02 & 9.590 + 3.222044 & -2.820e-02 & 9.760 + 3.235858 & -2.403e-02 & 9.677 + 3.237019 & -2.368e-02 & 9.888 + 3.237827 & -2.344e-02 & 9.861 + 3.238887 & -2.312e-02 & 9.836 + 3.245184 & -2.122e-02 & 10.083 + 3.268550 & -1.417e-02 & 10.452 + 3.269553 & -1.387e-02 & 10.481 + 3.286051 & -8.896e-03 & 11.023 + 3.287624 & -8.421e-03 & 10.997 + 3.288580 & -8.133e-03 & 11.091 + 3.296766 & -5.664e-03 & 11.477 + 3.297686 & -5.386e-03 & 11.591 + 3.298910 & -5.017e-03 & 11.697 + 3.301335 & -4.286e-03 & 11.856 + 3.304018 & -3.477e-03 & 12.119 + 3.305181 & -3.126e-03 & 12.209 + 3.306067 & -2.859e-03 & 12.272 + 3.307219 & -2.511e-03 & 12.498 + 3.308159 & -2.228e-03 & 12.715 + 3.308833 & -2.024e-03 & 12.833 + 3.309386 & -1.858e-03 & 12.824 + 3.310214 & -1.608e-03 & 13.081 + 3.311035 & -1.360e-03 & 13.029 + 3.311577 & -1.197e-03 & 13.314 + 3.313179 & -7.135e-04 & 13.795 + 3.313675 & -5.639e-04 & 14.128 + 3.314163 & -4.167e-04 & 14.827 + 3.314846 & -2.107e-04 & 15.637 + 3.316036 & 1.482e-04 & 8.389 + 3.316098 & 1.669e-04 & 8.595 + 3.316160 & 1.856e-04 & 8.159 + 3.316217 & 2.028e-04 & 8.289 + 3.316279 & 2.215e-04 & 7.875 + 3.316339 & 2.396e-04 & 7.907 + 3.316400 & 2.580e-04 & 7.823 + 3.316405 & 2.595e-04 & 7.992 + 3.316461 & 2.764e-04 & 7.654 + 3.316512 & 2.918e-04 & 7.572 + 3.316562 & 3.068e-04 & 7.689 + 3.316627 & 3.264e-04 & 7.617 + 3.316690 & 3.454e-04 & 7.643 + 3.316757 & 3.656e-04 & 7.274 + 3.316827 & 3.868e-04 & 7.473 + 3.316893 & 4.067e-04 & 7.407 + 3.316943 & 4.217e-04 & 7.611 + 3.316995 & 4.374e-04 & 7.202 + 3.317088 & 4.655e-04 & 7.302 + 3.317206 & 5.011e-04 & 6.963 + 3.317399 & 5.593e-04 & 7.138 + 3.317663 & 6.389e-04 & 6.968 + 3.317885 & 7.059e-04 & 6.913 + 3.317972 & 7.321e-04 & 6.858 + 3.318097 & 7.698e-04 & 6.693 + 3.318225 & 8.084e-04 & 6.562 + 3.318429 & 8.699e-04 & 6.581 + 3.318670 & 9.426e-04 & 6.702 + 3.318830 & 9.909e-04 & 6.573 + 3.318919 & 1.018e-03 & 6.624 + 3.319077 & 1.065e-03 & 6.661 + 3.319217 & 1.108e-03 & 6.493 + 3.319447 & 1.177e-03 & 6.449 + 3.319738 & 1.265e-03 & 6.279 + 3.320283 & 1.429e-03 & 6.528 + 3.320536 & 1.505e-03 & 6.285 + 3.320809 & 1.588e-03 & 6.270 + 3.320989 & 1.642e-03 & 6.261 + 3.321325 & 1.743e-03 & 6.209 + 3.321702 & 1.857e-03 & 6.096 + 3.322174 & 1.999e-03 & 5.919 + 3.322638 & 2.139e-03 & 6.104 + 3.323109 & 2.281e-03 & 5.910 + 3.323586 & 2.425e-03 & 5.944 + 3.324054 & 2.566e-03 & 5.853 + 3.324387 & 2.667e-03 & 5.711 + 3.324605 & 2.733e-03 & 5.867 + 3.324881 & 2.816e-03 & 5.874 + 3.325280 & 2.936e-03 & 5.802 + 3.325758 & 3.080e-03 & 5.776 + 3.326317 & 3.249e-03 & 5.741 + 3.327445 & 3.589e-03 & 5.682 + 3.328419 & 3.883e-03 & 5.649 + 3.329076 & 4.081e-03 & 5.485 + 3.329729 & 4.278e-03 & 5.666 + 3.331770 & 4.894e-03 & 5.507 + 3.332809 & 5.207e-03 & 5.440 + 3.333808 & 5.508e-03 & 5.465 + 3.334811 & 5.811e-03 & 5.397 + 3.335822 & 6.116e-03 & 5.382 + 3.336837 & 6.422e-03 & 5.337 + 3.337850 & 6.727e-03 & 5.329 + 3.338592 & 6.951e-03 & 5.308 + 3.339364 & 7.184e-03 & 5.301 + 3.340384 & 7.492e-03 & 5.277 + 3.341933 & 7.959e-03 & 5.241 + 3.344007 & 8.585e-03 & 5.199 + 3.346105 & 9.217e-03 & 5.161 + 3.348203 & 9.850e-03 & 5.142 + 3.350301 & 1.048e-02 & 5.112 + 3.352403 & 1.112e-02 & 5.085 + 3.354507 & 1.175e-02 & 5.051 + 3.356612 & 1.239e-02 & 5.022 + 3.357923 & 1.278e-02 & 5.015 + 3.358726 & 1.302e-02 & 4.987 + 3.359605 & 1.329e-02 & 5.006 + 3.361045 & 1.372e-02 & 4.992 + 3.363198 & 1.437e-02 & 4.967 + 3.365356 & 1.502e-02 & 4.949 + 3.367521 & 1.568e-02 & 4.931 + 3.369688 & 1.633e-02 & 4.909 + 3.371862 & 1.699e-02 & 4.884 + 3.374039 & 1.764e-02 & 4.875 + 3.375402 & 1.805e-02 & 4.917 + 3.376038 & 1.825e-02 & 4.857 + 3.376954 & 1.852e-02 & 4.862 + 3.378597 & 1.902e-02 & 4.845 + 3.380724 & 1.966e-02 & 4.836 + 3.382861 & 2.030e-02 & 4.820 + 3.385006 & 2.095e-02 & 4.812 + 3.387156 & 2.160e-02 & 4.803 + 3.389310 & 2.225e-02 & 4.789 + 3.391472 & 2.290e-02 & 4.774 + 3.393640 & 2.355e-02 & 4.766 + 3.395811 & 2.421e-02 & 4.765 + 3.397990 & 2.487e-02 & 4.738 + 3.400172 & 2.552e-02 & 4.736 + 3.402356 & 2.618e-02 & 4.727 + 3.404543 & 2.684e-02 & 4.720 + 3.406731 & 2.750e-02 & 4.718 + 3.408924 & 2.816e-02 & 4.710 + 3.411116 & 2.883e-02 & 4.701 + 3.412433 & 2.922e-02 & 4.666 + 3.413044 & 2.941e-02 & 4.661 + 3.413649 & 2.959e-02 & 4.737 + 3.414220 & 2.976e-02 & 4.723 + 3.414872 & 2.996e-02 & 4.681 + 3.415805 & 3.024e-02 & 4.678 + 3.417482 & 3.075e-02 & 4.674 + 3.419719 & 3.142e-02 & 4.661 + 3.421956 & 3.209e-02 & 4.660 + 3.424190 & 3.277e-02 & 4.651 + 3.426428 & 3.344e-02 & 4.645 + 3.428664 & 3.412e-02 & 4.631 + 3.430897 & 3.479e-02 & 4.630 + 3.433131 & 3.547e-02 & 4.632 + 3.435362 & 3.614e-02 & 4.647 + 3.437601 & 3.681e-02 & 4.628 + 3.439850 & 3.749e-02 & 4.626 + 3.442101 & 3.817e-02 & 4.620 + 3.444355 & 3.885e-02 & 4.609 + 3.446613 & 3.953e-02 & 4.597 + 3.448872 & 4.021e-02 & 4.592 + 3.451131 & 4.089e-02 & 4.594 + 3.453395 & 4.158e-02 & 4.584 + 3.455659 & 4.226e-02 & 4.576 + 3.459058 & 4.328e-02 & 4.579 + 3.463590 & 4.465e-02 & 4.570 + 3.468123 & 4.602e-02 & 4.557 + 3.472658 & 4.739e-02 & 4.562 + @xmath388 & @xmath389 & index + 2.848e-04 & 7.192e-02 & 0 + 3.128e-04 & 7.460e-02 & 0 + 3.445e-04 & 7.705e-02 & 0 + 3.836e-04 & 7.972e-02 & 0 + 4.460e-04 & 8.365e-02 & 0 + 5.319e-04 & 8.859e-02 & 0 + 5.322e-04 & 8.826e-02 & 0 + 5.963e-04 & 9.177e-02 & 0 + 6.638e-04 & 9.512e-02 & 0 + 7.279e-04 & 9.822e-02 & 0 + 8.332e-04 & 1.028e-01 & 0 + 9.630e-04 & 1.089e-01 & 0 + 9.658e-04 & 1.084e-01 & 0 + 1.279e-03 & 1.198e-01 & 0 + 1.399e-03 & 1.235e-01 & 0 + 1.430e-03 & 1.239e-01 & 0 + 1.518e-03 & 1.270e-01 & 0 + 2.171e-03 & 1.438e-01 & 0 + 2.421e-03 & 1.494e-01 & 0 + 3.255e-03 & 1.657e-01 & 0 + 3.862e-03 & 1.761e-01 & 0 + 4.066e-03 & 1.793e-01 & 0 + 4.868e-03 & 1.908e-01 & 0 + 6.266e-03 & 2.090e-01 & 0 + 7.261e-03 & 2.207e-01 & 0 + 9.366e-03 & 2.420e-01 & 0 + 1.319e-02 & 2.738e-01 & 0 + 1.834e-02 & 3.081e-01 & 0 + 2.285e-02 & 3.340e-01 & 0 + 2.854e-02 & 3.626e-01 & 0 + 3.135e-02 & 3.755e-01 & 0 + 3.496e-02 & 3.911e-01 & 0 + 3.783e-02 & 4.027e-01 & 0 + 4.348e-02 & 4.232e-01 & 0 + 5.230e-02 & 4.463e-01 & 0 + 5.320e-02 & 4.474e-01 & 0 + 5.559e-02 & 4.636e-01 & 0 + 6.988e-02 & 5.038e-01 & 0 + 7.140e-02 & 5.124e-01 & 0 + 7.389e-02 & 5.135e-01 & 0 + 8.380e-02 & 5.381e-01 & 0 + 9.589e-02 & 5.656e-01 & 0 + 1.025e-01 & 5.749e-01 & 0 + 1.034e-01 & 5.800e-01 & 0 + 1.339e-01 & 6.363e-01 & 0 + 1.710e-01 & 6.887e-01 & 0 + 2.227e-01 & 7.524e-01 & 0 + 3.106e-01 & 8.389e-01 & 0 + 4.970e-01 & 9.499e-01 & 0 + 5.850e-01 & 9.813e-01 & 0 + 6.980e-01 & 1.007e+00 & 0 + 7.580e-01 & 1.016e+00 & 0 + 8.190e-01 & 1.018e+00 & 0 + 8.790e-01 & 1.019e+00 & 0 + 9.390e-01 & 1.019e+00 & 0 + 9.980e-01 & 1.017e+00 & 0 + 8.433e-04 & 1.039e-01 & 0 + 8.995e-04 & 1.062e-01 & 0 + 9.060e-04 & 1.058e-01 & 0 + 1.221e-03 & 1.172e-01 & 0 + 1.515e-03 & 1.264e-01 & 8 + 1.611e-03 & 1.300e-01 & 8 + 1.647e-03 & 1.302e-01 & 8 + 1.729e-03 & 1.330e-01 & 8 + 1.791e-03 & 1.342e-01 & 8 + 1.931e-03 & 1.375e-01 & 8 + 2.075e-03 & 1.417e-01 & 8 + 2.315e-03 & 1.473e-01 & 8 + 2.382e-03 & 1.478e-01 & 8 + 2.661e-03 & 1.545e-01 & 8 + 2.849e-03 & 1.585e-01 & 8 + 3.085e-03 & 1.625e-01 & 8 + 3.693e-03 & 1.731e-01 & 8 + 4.713e-03 & 1.890e-01 & 8 + 5.169e-03 & 1.950e-01 & 8 + 6.004e-03 & 2.057e-01 & 8 + 6.811e-03 & 2.153e-01 & 8 + 7.311e-03 & 2.205e-01 & 8 + 8.153e-03 & 2.294e-01 & 8 + 1.314e-02 & 2.727e-01 & 8 + 1.500e-02 & 2.857e-01 & 8 + 1.934e-02 & 3.140e-01 & 8 + 2.485e-02 & 3.438e-01 & 8 + 2.880e-02 & 3.617e-01 & 8 + 3.157e-02 & 3.747e-01 & 8 + 3.505e-02 & 3.897e-01 & 8 + 3.827e-02 & 4.021e-01 & 8 + 4.388e-02 & 4.226e-01 & 8 + 5.599e-02 & 4.619e-01 & 8 + 7.054e-02 & 5.023e-01 & 8 + 7.474e-02 & 5.123e-01 & 8 + 8.388e-02 & 5.347e-01 & 8 + 9.639e-02 & 5.623e-01 & 8 + 1.467e-01 & 6.502e-01 & 8 + 2.096e-01 & 7.309e-01 & 8 + 2.627e-01 & 7.854e-01 & 8 + 3.267e-01 & 8.368e-01 & 8 + 3.927e-01 & 8.802e-01 & 8 + 5.547e-01 & 9.589e-01 & 8 + 6.773e-01 & 9.766e-01 & 8 + 8.272e-01 & 9.618e-01 & 8 + 9.960e-01 & 9.989e-01 & 8 +
|
parametric expressions are used to calculate the isothermal susceptibility , specific heat , order parameter , and correlation length along the critical isochore and coexistence curve from the asymptotic region to crossover region .
these expressions are based on the minimal - subtraction renormalization scheme within the @xmath0 model . using two adjustable parameters in these expressions , we fit the theory globally to recently obtained experimental measurements of isothermal susceptibility and specific heat along the critical isochore and coexistence curve , and early measurements of coexistence curve and light scattering intensity along the critical isochore of @xmath1he near its liquid - vapor critical point .
the theory provides good agreement with these experimental measurements within the reduced temperature range @xmath2 .
|
for metal element @xmath122@xcite , @xmath123 is the state density which have some spin directions on the fermi surface .
@xmath124 according to the assumption@xmath125 formalism([renormal ] ) can be written as@xmath126 so from bcs theory , the cooper pair lies in the attraction area , ie @xmath127 . on substitution of @xmath128ev
@xcite we can estimate @xmath129 , so @xmath108ev .
wu , m. s. byrd , and d. a. lidar , _ phys rev lett . _
* 89 * , 057904 ( 2002 ) ; j. dukelsky , j. m. romn , and g. sierra , _ phys rev lett . _ * 90 * , 249803 ( 2003 ) ; l .- a .
wu , m. s. byrd , and d. a. lidar , _ phys rev lett . _ * 90 * , 249804 ( 2003 )
|
we propose a new simulation computational method to solve the reduced bcs hamiltonian based on spin analogy and submatrix diagonalization .
then we further apply this method to solve superconducting energy gap and the results are well consistent with those obtained by bogoliubov transformation method .
the exponential problem of @xmath0-dimension matrix is reduced to the polynomial problem of @xmath1-dimension matrix .
it is essential to validate this method on a real quantum computer and is helpful to understand the many - body quantum theory .
bcs theory@xcite and its subsequent extension is a well established theory to explain the mechanism of superconducting property . with two gross simplifications : the free electron approximation and the effective interaction approximation@xcite@xmath2 @xcite , a simplified bcs model is obtained and described by the reduced bcs hamiltonian . there has been much work on solving this hamiltonian .
the mean field method is exact in the limit of large number of electrons where fluctuation can be neglected but disabled in the case of small number of electrons . since richardson s work@xcite in the 60 s to now , the exactly solvable bcs hamiltonian attracts much attention in connection with the problems in different areas of physics such as superconductivity , nuclear physics , physics of ultrasmall metallic grains . recently in _
l .-
a .
wu et al . _
s paper@xcite an nmr experiment scheme performing a polynomial - time simulation of pairing model was reported .
based on this work we propose an explicit theory method to diagonalize the reduced bcs hamiltonian through the spin analogy and submatrix diagonalization . compared with the conventional method it is more useful in solving practical problem .
the problem is solved in the spin space , which is convenient related to the qubit system .
it gives a senseful alive method , quantum simulation , instead of the numerical diagonalization calculation .
and it shows the potential to solve many - body problem by quantum computer .
in fact more and more people concentrate on the research of simulating other physics systems by quantum computer@xcite .
the experimental quantum simulations about quantum harmonic oscillator@xcite , three - spin artifical hamiltonian@xcite and migration of excitation in a one - dimensional chain@xcite _ et al .
_ have been realized .
recently a relative experiment is performed to get the eigenvalues of the bcs hamiltonian through selecting a proper initial state and realizing hamiltonian evolution.@xcite@xcite the exact solvable model , _
i.e. _ the reduced bcs hamiltonian considered in this paper is@xcite@xmath2@xcite@xmath2@xcite : @xmath3 where @xmath4 are the electron number operators , @xmath5 is the fermionic creation ( annihilation ) operator .
the coupling coefficient is simplified as a constant @xmath6@xcite@xmath2@xcite .
note that the summation indexes @xmath7 represent all of relevant quantum numbers , and the electron pairs are labelled by the the quantum number @xmath8 and @xmath9 , according to the cooper pair situation where the paired electrons have equal energies but opposite momenta and spins : @xmath10 and @xmath11 .
introduce the pair creation operator @xmath12 and the pair annihilation operator @xmath13 .
so one can write the hamiltonian ( [ bcsh ] ) as@xcite : @xmath14 where @xmath15 is the free electron kinetic energy from fermi surface ( @xmath16 is the fermi energy ) . there are two possible cases for every pair state @xmath8 : occupation `` and empty '' , which are denoted respectively by : @xmath17 where the spin up state @xmath18 indicates occupation `` and the spin down state @xmath19 indicates empty '' .
obviously , @xmath20 and @xmath21 , then we can get the so - called spin - analogy corresponding of the pair annihilation operator @xmath22 as @xmath23 in the same way , the spin - analogy corresponding of the pair creation operator @xmath24 becomes @xmath25 from the pair number operator @xmath26 has the eigenvalue 2 ( which represents the electron number in every cooper pair ) when operating on @xmath18 , and 0 when operating on @xmath19 , it follows that @xmath27 in fact , the fermionic pair operators satisfy the commutation algebra : @xmath28 , _ i.e. _ @xmath29 . from formulas ( [ bx])-([nk ] )
one can express @xmath30 in terms of the spin operators : @xmath31 where @xmath32 .
in fact the spin analogy of the bcs hamiltonian is well known and exact diagonalization of the pairing model in the spin space has been carried out in several previous works@xcite . in this paper
we propose a computational simulation method which is potential to realize in future with the development of quantum computer .
the primary advantage of our method lies in the practical realization in experiments .
especially we can solve superconducting energy gap by this simulation method conveniently as following paragraphs .
it is more practicably than other solution of energy gap , because it can simplify a @xmath0-dimension problem to an @xmath1-dimension problem .
we know eigenvalues may not be solvable for high dimension matrix in principle .
now a @xmath0-dimension problem , exponential problem(ep ) can be simplified an @xmath1-dimension problem , polynomial problem(pp ) . in the following part
we will describe how to transform ep to pp in detail .
firstly the total hamiltonian @xmath33 will be expressed as the direct - sum of a set of submatrices@xcite .
@xmath34 the system states with the same spin - up state number form an absolute subspace .
the subscripts @xmath35 representing the number of the spin - up state in the corresponding subspace are respectively @xmath36 .
secondly we will prove that the eigenvalues of @xmath37 in sub1 submatrix justly are the eigenvalues of @xmath38 note the @xmath39 operator as@xmath40 whose @xmath41-th diagonal element is noted as @xmath42 $ ] @xmath43 .
it is easy to see the non - diagonal elements of @xmath44 are zero . for the hamiltonian as @xmath45 , the @xmath46-th element satisfies@xmath47=\delta_{im}\text { \ \ \ \ } ( i , m=1,2,\cdots , n )
\label{delta im}%\ ] ] we will prove this lemma by mathematic induction . when @xmath48 , it is easy to get three matrices , @xmath8 value is 1 , 2 , 3 respectively . @xmath49 here the non - diagonal elements are all 0 .
these diagonal elements with @xmath50 @xmath50 are
@xmath51 $ ] .
it is easy to validate @xmath51=\delta_{im}$ ] and the last diagonal element @xmath52 $ ] is zero .
if when @xmath53 , @xmath54=\delta_{im}\text { \ } ( i , m=1,2,\cdots , l ) \label{l(1)}%\ ] ] is right and @xmath55=0 \label{l(2)}%\ ] ] then we should examine whether @xmath56=\delta_{im}$ ] @xmath57 is right when @xmath58 . we will discuss it in two cases : @xmath59 and @xmath60 .
\(1 ) @xmath59 : @xmath61 from eq.([h(l+1,m ) ] ) it is easy to see@xmath62=h(l+1,m)[2x-1]=h(l+1,m)[x ] \label{h(2x)}%\ ] ] so@xmath63=h(l , m)[2^{l}-2^{l - i}]=\delta_{im}\text { \ } ( i=1,2,\cdots , l)\ ] ] next we should also know the value of @xmath64 $ ] . from eq.([l(2 ) ] ) and eq.([h(2x ) ] ) there is @xmath65=h(l , m)[2^{l}]=0 $ ] .
from above discussion for @xmath66 , the equality @xmath56=\delta_{im}$ ] is valid .
\(2 ) @xmath60:@xmath67 from the above expressions the @xmath68-th diagonal element @xmath69=1 $ ] , that is to say@xmath70=1\text { } ( i = l+1)\ ] ] and then for @xmath71 , @xmath70=0\ ] ] because the even - th diagonal elements are all zero obviously .
so for @xmath60 there is also @xmath72=\delta_{im}$ ] . from the discussion ( 1 ) and ( 2 )
we have proved that when @xmath58 , @xmath56=\delta_{im}$ ] @xmath57 is valid .
so for @xmath73 ( @xmath1 is the natural number ) , there is the equality@xmath47=\delta_{im}\text { \ \ \ \ } ( i , m=1,2,\cdots , n)\ ] ] after the preparation we will prove that the eigenvalues of @xmath37 in sub1 subspace justly are the energy spectrum of quasiparticle excitation of @xmath30 .
firstly set a diagonal hamiltonian as @xmath74 , @xmath75 and @xmath76 are the quasiparticle number operators .
according to the previous analogy rule of number operators eq.([nk ] ) the spin - analogy form of @xmath77 is@xmath78 @xmath79 s submatrix in sub1 subspace is denoted as @xmath80 and the @xmath41-th diagonal element of @xmath81 as @xmath82 $ ] . we can find the @xmath41-th diagonal element of @xmath80 is the @xmath46-th diagonal element of the total hamiltonian @xmath79 . according to eq.([delta i m ] ) @xmath83 & = h_{\mathrm{spin}}^{\mathrm{diag}}% [ 2^{n}-2^{n - i}]\\ & = \sum_{m=1}^{n}e_{m}h(n , m)[2^{n}-2^{n - i}]\\ & = \sum_{m=1}^{n}e_{m}\delta_{im}\\ & = e_{i}%\end{aligned}\ ] ] so the spin - analogy hamiltonian of a diagonal bcs hamiltonian @xmath77 in sub1 subspace has the same eigenvalues as the diagonal bcs hamiltonian s energy spectrum .
we can deduce further that the eigenvalues of @xmath37 are justly the eigenvalues of @xmath30 whether which is diagonal or not , because @xmath30 can be written as the diagonal form like @xmath77 generally , i.e. @xmath84 , while we need not care about how to obtain the diagonal form hamiltonian .
consequently it implies that if we have diagonalized @xmath37 we can get the energy spectrum of quasiparticle excitation of @xmath30 and get energy gap further avoiding complex computation .
one of the classical methods to diagonalize @xmath30 , bogoliubov transformation method is available under the mean - field approximation , which is not exact especially in the case of limited @xmath1 .
another famous method about the exact solution of bcs hamiltonian has been proposed in the 60 s by richardson@xcite .
he considered the system with @xmath85 pairs electrons and constructed a set of operators @xmath86 @xmath87 commuted with @xmath30 , finally gave the expression of the eigenvalues @xmath88 of @xmath86 through solving the @xmath89 coupled algebraic equations .
@xmath88 is not yet the eigenvalue of @xmath30
. we can more directly and simply give the energy spectrum of @xmath30 .
thus it can be seen the idea of diagonalizing @xmath37 instead of solving the eigenvalues of @xmath30 directly is better than those classical ones .
now a key problem that how to get the submatrix @xmath37 is placed to the front .
the correlative work@xcite in our group has proved that the general form of @xmath37:@xmath90=\epsilon_{i},\text { \ \ \ } h_{sub1}[i , j]=-v\text { \ } % ( i , j=1,2,\cdots , n;\text { } i\neq j ) \label{wang}%\ ] ] @xmath91 $ ] is the matrix element of @xmath37 .
finally it is necessary to check the methods in numerical computation . here
we will compare our solution with the result of the mean - field approximation by the value of superconducting energy gap @xmath92 @xmath93 . according to the physics meaning of @xmath92 ,
the energy required to excite at least a quasiparticle from the fermi surface , the energy of the element excitation is written as @xmath94.@xcite in fact the element excitation energy is also the eigenvalue of @xmath30 . after getting the eigenvalue of @xmath37 , _
i.e. _ the eigenvalue of @xmath30 , we can get the value of @xmath92 by solving equation @xmath94 .
but in order to get rid of the effect of energy zero the equation@xmath95 is used to solve @xmath92 in practice , because the difference of eigenvalues dosent depend on the energy zero and that we find the energy difference between the ground and the first excited state @xmath96 is by far larger than @xmath97 in the course of the numerical computation . here
@xmath98 and @xmath99 .
the another kind of solution used in comparing is the following energy gap equation@xcite:@xmath100 here we consider the reduced bcs model whose energies are given for simplicity by @xmath101@xcite@xmath2@xcite@xmath2@xcite , here @xmath102 is the average level spacing which is inversely proportional to the size of the grains . in the strong coupling regime , corresponding to large grains or strong coupling constants , @xmath103 . in the weak coupling region , corresponding to small grains or small coupling constants , @xmath104 @xcite . from much research about ultrasmall superconducting grains@xcite
, the mean - field theory is not suitable in the weak coupling region .
it has been proved that the corrections to the mean - field results are small in large grains become important in the opposite limit@xcite .
so in this paper we carry out the numerical computation in the first case , @xmath103 .
we also take the coupling constants @xmath105 in order to discuss conveniently . in order to give the numerical pictures ,
we suppose the value of @xmath6 by the rough estimate . from bcs theory
, the cooper pair lies in the attraction area , _
i.e. _ @xmath106@xcite . for metal debye energy @xmath107ev
, we can set @xmath108ev by rough estimate .
the estimate process is put to the later appendix .
another two variables @xmath109 and @xmath1 are taken as the independent variables of the energy gap .
we list our results in diagrams . in fig.[gap - m ] setting @xmath110 , the energy gap is plotted as the function of energy level number , which is the mono - increasing function of the energy level number @xmath1 . as a function of the number of energy level , @xmath1 .
we give the comparison between two results by the different methods .
the solid line is the solution of the energy - gap equation ; the dashed line is our result by spin analogy and diagonalizing submatrix .
here we choose @xmath108ev , @xmath110 .,scaledwidth=50.0% ] the relative error between our result and that of the eq.([gapeqold ] ) is not more than 5% in the range from @xmath111 to @xmath112 and fixed @xmath113 .
it shows that the result from our method is well consistent with the solution of the energy gap equation . in order to check the universality of this new method
, we also give the dependence relation between the energy gap and the level spacing , see fig.[gap - namda ] .
, energy gap @xmath92 is plotted as the function of @xmath109 .
we give the comparison between two results : the solid line is the solution of the energy - gap equation ; the dashed line is our result by spin analogy and diagonalizing submatrix . here
we choose @xmath108ev.,scaledwidth=50.0% ] it is clear to see @xmath92 changes gently with @xmath109 when @xmath109 is large enough .
that is to say , @xmath92 is almost independent of @xmath102 when @xmath102 is small enough .
it shows the rationality of @xmath114 on the inverse hand . obviously , in fig.[gap - namda ]
, when @xmath109 is larger than 80 , the relative error is less than 1.1% .
we also consider a small departure from the fermi surface , that is @xmath115 .
@xmath116 is a small value , @xmath117 @xmath102 . in the following discussion we note @xmath118
, @xmath119 is the natural number .
the small departure from the fermi surface reduces the energy gap and energy gap is not a real root when the departure reaches a critical value , see fig.[gap - b ] .
ev , @xmath110 , @xmath120 the energy gap @xmath92 is ploted as the function of @xmath119 through the quantum simulation . when @xmath121 , there is not the real root for @xmath92 .
the result is obtained by the quantum simulation .
similar result will be obtained by the energy gap equation eq.([gapeqold]).,scaledwidth=50.0% ] according to above comparison we know two results are consistent well , while our method to solve the energy spectrum does nt include approximation , which indicates that our result includes that obtained by mean field theory and is superior to it . in summary
, we have proposed an exact numerical simulation method to calculate the energy spectrum of the reduced bcs hamiltonian by spin analogy and diagonalizing submatrix .
a numerical computation to verify the validity of our computational method is given .
we make a comparison between our method and energy gap equation eq.([gapeqold ] ) , and two results are well consistent in numerical computation . by examining the change of the energy gap value under the change of the parameter
, we include the excellent consistency between the two results by the different methods is independent on the particular parameter .
it implies that one can implement this quantum simulation on a quantum computer and the result will be believable . currently a new experiment about 2-qubit simulation of the pairing hamiltonian on an nmr quantum computer has been realized and get the energy spectrum of the pairing hamiltonian successfully@xcite . with the development of quantum computer , especially the manipulation and control of multi - qubit system
, this new simulation computation method has the great potential in practical application .
we are grateful xiaosan ma , wanqing niu , zhao ningbo , zhu rengui and su xiao - qiang for helpful discussion .
this work was founded by the national fundamental research program of china with no .
2001cb309310 , partially supported by the national natural science foundation of china under grant no .
60173047 and the natural science foundation of anhui province .
|
the x - ray source was discovered in february , 1999 , independently with the wide field cameras ( wfcs ) on _ bepposax _ ( in t zand et al .
1999a ) and with the proportional counter array ( pca ) on _ rossixte _ ( ergo , its alternative designation xte j1819254 , markwardt et al .
the wfc detection involved an hour - long flare with a peak of 80 mcrab on feb . 20
the first pca detection occurred during regular scans of the galactic center field , on feb .
18 . for some time
there was confusion over the identification of the optical counterpart .
the variable star gm sgr is within the 2-radius wfc error box of the x - ray source as noted by in t zand et al .
( 1999a ) .
however , it was later discovered that the actual counterpart is another , previously unknown , variable star only @xmath11 north of gm sgr : ( williams 1999 , samus et al .
1999 ) . was found to show a large outburst in visual magnitude @xmath2 from about 11 on sep .
14.8 up to 8.8 on sep
. 15.4 ( stubbings 1999 ) .
x - ray observations with the all - sky monitor on rossixte also showed a bright outburst of up to a peak intensity of 12.2 crab on sep .
15.7 ( 2 - 12 kev , smith et al .
the giant x - ray outburst was preceded by a 4.5 crab precursor 0.8 d earlier . between both peaks
the flux decreased to a quiescent level .
wijnands & van der klis ( 2000 ) report that during the tail of the giant x - ray outburst , strong variability was observed of factor - of - four on seconds time scales to factor - of-500 on minutes time scales .
the giant outburst was also seen at radio wavelengths ( hjellming et al .
1999a , 1999b ) .
the radio source was resolved and had the appearance of a twin jets structure extending 0.25 each jet .
the giant and precursor outburst were , furthermore , seen in 20 - 100 kev ( mccollough et al . 1999 ) . within 2 d after the giant outburst , settled down to a quiescent brightness of @xmath3 .
combined with the failure of any x - ray detection since the giant outburst ( c.b .
markwardt , priv . comm . )
, this shows that the flaring activity faded considerably , although the balmer lines continued to show additional activity up to at least sep .
30 ( wagner 1999 ) .
if one assumes that the emission is due to a main sequence a star ( garcia et al .
1999 ) , one obtains with b
v=0.2 and e(b v)=0.24 ( wagner 1999 ) a distance of 0.4 kpc
. we will adhere to this distance in the present paper but acknowledge its uncertainty . from the magnitudes
alone , we estimate the uncertainty to be about 50% . with a 0.4 kpc distance , the giant outburst would have a peak x - ray luminosity of @xmath4 in 2 - 10 kev .
this luminosity , and the spectrum discussed in the present paper , clearly suggest that there is a compact object at play that is either a neutron star or black hole .
it also identifies the object as a low - mass x - ray binary ( lmxb ) , and the short distance implies it is possibly the _ nearest _ lmxb known so far ( c.f . , van paradijs & white 1995 ) .
this opens up interesting perspectives for as a case study of the quiescent emission from transient low - mass x - ray binaries .
we note that recently rutledge et al .
( 1999 ) introduced diagnostics for the nature of the compact object from the quiescent emission . in this paper , we present the observations of with the wfcs and narrow field instruments ( nfi ) on _ bepposax _ in the spring and fall of 1999 .
this includes the longest semi - continuous observation of the system so far .
was observed in two states : a flaring state with peak fluxes that are one to two orders of magnitude less than that of the giant outburst , and a calm state with an average flux that is about three orders of magnitude smaller ( though still considerably higher than that of the quiescent emission ) .
the _ bepposax _ platform carries two sets of astrophysical x - ray and @xmath5-ray devices in space ( boella et al .
one pertains to two identical wide field cameras ( wfcs , jager et al .
1997 ) that view the sky with @xmath6 square degrees field - of view in opposite directions with 5 spatial resolution in the 2 to 26 kev bandpass .
the other set includes the narrow field instruments ( nfi ) that are co - aligned in a direction that is perpendicular to that of both wfcs .
the nfi include 2 imaging instruments that are active below 10 kev , the low - energy and the medium - energy concentrator spectrometer ( lecs and mecs respectively , see parmar et al .
1997 and boella et al .
1997b respectively ) , and two non - imaging instruments that have bandpasses of @xmath7 to 300 kev ( the phoswich detector system or pds , frontera et al .
1997 ) and 4 to 120 kev ( the high - pressure gas scintillation proportional counter or hp - gspc , manzo et al .
the mecs has a photon energy resolution of 8% ( full width at half maximum ) at 6 kev . since mid 1996 ,
the wfcs carry out a long - term program of monitoring observations in the field around the galactic center .
the program consists of campaigns during the spring and fall of each year .
each campaign lasts about two months and typically comprises weekly observations .
was detected twice during these campaigns so far , in hourly exposures above a @xmath150 mcrab detection limit , on feb .
20 and sep . 10 , 1999 .
the first wfc - detection triggered a target - of - opportunity observation ( too ) with the nfi on march 13.22 - 14.02 , 1999 ( this is 186 d before the giant outburst ) .
was strongly detected in three nfi ( the hp - gspc was not turned on ) , and the average intensity was found to be about 12 mcrab ( 2 - 10 kev ) .
the nfi net exposure times on are 10.2 ks for lecs , 27.7 ks for mecs and 14.7 ks for pds .
the lecs and mecs images show a single bright source , the position as determined from the mecs image is r.a . 18@xmath819@xmath9222 , decl .
= 252403 ( eq . 2000.0 , error radius 08 ) which is 10 from that determined with the wfc ( in t zand et al .
1999a ) and 04 from the optical counterpart .
the background - subtracted light curves of the two flares as measured with the wfcs are presented in figs .
[ figwfclc1 ] and [ figwfclc2 ] . besides these two occasions
, the source was never detected in either short 1-hr or long 24-hr wfc exposures with typical upper limits of 50 and 12 mcrab respectively .
the light curves of both detections are characterized by sporadic flaring activity up to 0.2 , which corresponds to 0.1 crab units , with a larger and longer flare on sep . 10 with a peak of 0.3 crab units .
the large flare was above the detection limit for about 6 hr and occurred in the middle of a 34 hr long observation .
it started on sep .
10.0 , which is 5.4 d before the giant outburst .
its duration is comparable to that of the giant flare .
the flaring activity on a time scale of hours is in line with the behavior as measured with the proportional counter array ( pca ) on _ rossixte _ ( markwardt et al .
the pca covers bi - weekly since feb . 5 , 1999 , for individual snapshot exposures of about 60 s with a sensitivity of about 1 mcrab .
was first seen on feb .
18 and subsequently showed an erratic variable behavior until immediately after the giant outburst , with fluxes ranging from 0.5 to 30 mcrab ( markwardt et al .
1999b ) . since the giant outburst
, no emission was detected anymore ( c.b .
markwardt , priv . comm . ) .
the wfc data do not reveal any type i x - ray burst from in about 0.8 ms of source coverage for the year 1999 when it was seen to be active with the pca ( markwardt et al .
1999b ) , or for @xmath13 ms over all wfc observations since 1996 .
[ fignfilc ] shows the evolution of the background - corrected photon count rates in various bandpasses of the three nfi , in 400 s time resolution . for an explanation of the method of data extraction
, we refer to in t zand et al .
( 1999b ) .
the light curves show slow variability on a time scale of a few hours with an amplitude of about 50% .
the average flux level is 12 mcrab ( 2 - 10 kev ) which , relative to the flares , can be regarded as calm emission .
however , we note that this can not be regarded as quiescent emission because the pca has seen flux levels from this source at least one order of magnitude smaller ( markwardt et al .
1999b ) .
a power density spectrum of the mecs photon count rate , generated with a timing resolution of 0.02 s and averaged over 256 s time intervals , reveals no measurable narrow features nor broad - band noise .
the upper limit on the variability , integrated between 0.01 and 10 hz with power laws with indices of 0.5 , 1.0 and 2.0 , are 18 , 8 and 3.5% fractional rms respectively ( 90% confidence ) .
the upper limit on the pulsed amplitude is 1.3% fractional rms ( 95% confidence ) in 0 to 25 hz .
.best - fit parameter values of the comptonized model to the nfi spectrum .
the last line specifies the @xmath10 value for the fit without a bb ( black body ) component .
this value applies after re - fitting the remaining parameters .
ew means equivalent width .
[ cols= " < , < " , ] [ tabnfifit ]
we first extracted a spectrum for the average emission over the whole nfi observation .
the spectral channels were rebinned so as to sample the spectral full - width at half - maximum resolution by three bins and to accumulate at least 20 photons per bin .
the bandpasses were limited to 0.4 - 4.0 kev ( lecs ) , 1.8 - 10.5 kev ( mecs ) and 15 - 120 kev ( pds ) to avoid photon energies where either the spectral calibration of the instruments is not complete or no flux was measured above the statistical noise .
we tried to model the spectrum with various descriptions . in all models
, an allowance was made to leave free - within reasonable limits - the relative normalization of the spectra from lecs and pds to that of the mecs spectrum , to accommodate cross - calibration uncertainties in this respect .
publicly available instrument response functions and software were used ( version november 1998 ) .
the continuum could best be fitted ( @xmath11 for 96 dof ) with a comptonization model ( titarchuk 1994 ) plus black body radiation , see table [ tabnfifit ] .
next to that there is a strong emission line at 7 kev .
a fit with a single narrow line results in a centroid energy of 6.85@xmath12 kev .
we identify this as k@xmath13 fluorescence in strongly ionized iron .
the centroid energy is between the expected fe - k lines for helium - like ( 6.68 kev ) and hydrogen - like fe ( 6.96 kev ) .
we included in the model narrow lines at these fixed energies .
the best - fit parameter values are given in table [ tabnfifit ] and a graph of the spectrum and the model fit in fig .
[ fignfispectrum ] .
the fits with other continuum models , in combination with a black body and two narrow line components results in fit qualities of @xmath14 for 98 dof ( thermal bremsstrahlung ) , @xmath15 for 96 dof ( broken power law ) and @xmath16 for 96 dof ( power law with high - energy cut off ) .
terada et al . ( 1999 ) reported about the 1996 transient ax j1842.80423 which exhibited an fe - k line at 6.80 kev with a large equivalent width of 4 kev .
the 0.5 - 10 kev continuum plus line spectrum was successfully fitted with a thin hot thermal plasma emission model of temperature 5.1 kev .
we fitted such a model according to the mekal code implementation ( mewe et al .
the fit was reasonable , provided two additional components were included . with a power law and black body as additional components , @xmath17 ( 98 dof ) , which is a worse fit than the comptonization model in table [ tabnfifit ] .
the resulting plasma temperature is @xmath18 kev .
the fitted contribution of the thin plasma to the flux is of order 10% .
the emission measure of the thermal plasma is @xmath19 @xmath20 .
e(b v)=+0.24 ( wagner 1999 ) implies @xmath21 @xmath22 according to the relationship defined by predehl & schmitt ( 1995 ) .
if we assume that the uncertainty in e(b
v ) is 0.10 , where most of the uncertainty comes from the uncertainty in the calibration of the relationship used by wagner ( 1999 ) between the equivalent width of the 578.0 nm interstellar absorption line to e(b v ) ( see herbig 1975 ) , then the error in @xmath23 is 0.05@xmath24 @xmath22 .
if we fix @xmath23 to @xmath25 @xmath22 and leave free the remaining parameters of the comptonized model in table [ tabnfifit ] , @xmath10 is 1.31 ( 97 dof ) . we conclude that @xmath23 as determined from the x - ray spectrum is consistent with e(b v)=+0.24@xmath26 . to determine whether the variability as illustrated in fig .
[ fignfilc ] is accompanied by strong spectral changes , we extracted a spectrum for times when the source was relatively faint and one for times when the source was relatively bright .
these times are indicated by hatched areas in fig .
[ fignfilc ] .
subsequently , we employed the same comptonized model as for the whole observation , leaving free only the normalizations of the different contributions .
the resulting values for @xmath10 are 1.36 for the bright data ( 104 dof ) and 1.22 ( 103 dof ) for the faint data .
the 44% difference in the 0.4 - 10 kev flux between the faint and bright data is due to approximately equal changes in blackbody and comptonized components ( i.e. , 25 and 19% respectively ) .
the flux of the emission lines scales with the integral flux : the combined equivalent widths of both lines is identical in both cases at 0.26 kev .
this spectral behavior is illustrated in fig .
[ fig2mecsspectra ] which zooms in on the mecs part of the spectrum ( including the emission lines ) for the two extremes . in fig .
[ fignfiratio ] the ratio between both spectra is presented .
this is consistent with a constant ratio of @xmath27 throughout the spectrum ( @xmath28 for 65 dof ) .
the apparent bump between 4 and 5 kev is statistically not significant .
we extracted a 2 - 26 kev spectrum of the complete sep .
10 flare which amount to an elapsed time of 5.5 hr and an exposure time of 2.3 hr , and attempted to fit the same comptonization model as applies to the nfi data when the source is @xmath29 times fainter . leaving free only the normalizations of the different spectral components ,
the fit was unsatisfactory with @xmath30 for 24 dof .
however , if @xmath23 was allowed to vary , the fit became dramatically better with @xmath31 ( 23 dof ) and @xmath32 @xmath22 .
if instead the black body temperature was allowed to vary , an even better improvement occurred with @xmath33 ( 23 dof ) and k@xmath34 kev .
if the fe - k emission lines at 6.68 and 6.96 kev are replaced for a single one whose energy is left free , the fit improves further to @xmath35 for 23 dof .
the line energy then is @xmath36 kev .
this suggests a shift of the ionization balance to a lower degree ( fe i - xvii ) .
we determined an average flux over the 3 hr duration of the peak of @xmath37 .
the equivalent width of the emission line is @xmath38 kev .
the flare spectrum is presented in fig .
[ figwfcspe2 ] .
we have analyzed the spectrum at basically two flux levels which correspond to 2 - 10 kev luminosities of @xmath39 and @xmath40 , where @xmath41 is the distance to .
these compare to a maximum luminosity during the giant outburst of @xmath42 ( smith et al .
1999 ) and a minimum of @xmath43 ( see below ) .
these numbers may be multiplied by a factor of 2 to obtain a rough extrapolation of the x - ray luminosity to the 0.4 - 120 kev range .
they show that , though bright , in fact has a peak luminosity that is at least an order of magnitude below the eddington limit for a mass of the compact object that is equal to or larger than that of a canonical 1.4 m@xmath44 neutron star .
however , one should bear in mind that the sampling of the asm light curve is sparse ( about ten 90 s observations per day ) and we do not know what the source did between asm measurements . the eye - catching feature of the _ bepposax _ data is the strong fe - k emission line complex .
the line energies during the calm phase point to a large ionization degree of the iron .
two origins for the line can be considered : thermal emission from a thin hot plasma , where iron atoms are collisionally excited , or fluorescent emission from a plasma illuminated by the continuum x - radiation .
the thin hot plasma model gave an unsatisfactory fit to the nfi data .
also , the fact that no difference was seen in equivalent width between the low- and high - flux intervals of the nfi data seems to be at odds with a thin thermal plasma origin of the emission line .
therefore , we believe it is more probable that we are dealing with fluorescent emission in a photo - ionized medium .
the ionization degree of the medium is fairly high , the ionization parameter @xmath45 ( kallman & mccray 1982 ) , where @xmath46 is the size of the fluorescent material and @xmath47 the atom density , is of order 10@xmath48 to 10@xmath49 which implies that the medium must be close to the source of the continuum radiation .
the 2 to 26 kev spectrum of the 2nd wfc - detected flare ( fig . [ figwfcspe2 ] ) shows an ionization balance of iron that is at much lower degrees than that of the calm emission half a year earlier .
apparently , @xmath50 is at least two orders of magnitude smaller .
this implies that @xmath47 and/or @xmath46 increased substantially .
therefore , the cloud of fluorescent material must have been , during the flare , denser and/or farther away from the source of the fluorescing radiation ( probably near to the compact object ) .
the same probably applies to the giant outburst 5 days later because the iron line was seen at an energy of 6.5 kev then ( markwardt et al .
. we do not detect k - edge absorption of iron at the appropriate energies of 8.8 ( fe xxv ) and 9.3 kev ( fe xxvi ) .
the 3@xmath51 upper limit on the relative depth of an edge at 9.0 kev is 0.2 which does not appear to be very constraining ( e.g. , makishima 1986 ) .
the equivalent width of the fe - k line complex is relatively large . in the wfc - detected flare in september
, it was 3 to 4 times as large as during the march nfi observation .
markwardt et al .
( 1999b ) also observed the fe - k complex during the tail of the giant flare , at a photon energy that is close to that measured with wfc six days earlier and with an equivalent width of about 500 ev .
model calculations of the structure of the accretion disk coronae and data analyses of fe - k lines in other x - ray binaries by vrtilek et al .
( 1993 ) suggest that large equivalent widths in fe - k lines may be due to high ( edge - on ) inclination angles .
this appears to be in line with the high inclination angle inferred from the radio jets . hjellming et al .
( 1999b ) reported a double - sided jet structure from in a vla image taken on sep .
16.02 ut , with sizes of roughly 0.25 in both directions . a subsequent vla image taken on sep .
17.93 only shows one of the two jets at the same position .
if @xmath52 is the intrinsic velocity of the jets , @xmath53 the angle between the jets and the line of sight , and @xmath54 the velocity of light , then the angular proper motion @xmath55 of the approaching ( ) and receding jet ( + ) are given by : @xmath56\end{aligned}\ ] ] ( e.g. , hjellming & rupen 1995 ) .
we can only make a rough estimate of @xmath57 and @xmath58 .
the uncertainty results from the facts that no moving radio blobs were seen , like in gro j165540 ( c.f .
, hjellming & rupen 1995 ) , and that no difference was measured between @xmath57 and @xmath58 ( this is partly due to insufficient accuracy of the position of the optical counterpart ) .
if we assume that the difference between @xmath57 and @xmath58 is less than 10% , we find that @xmath59 .
the value for @xmath60 is determined by the time of jet ejection . depending on whether one takes this to be at the time of the measured optical peak ( stubbings 1999 ) , x - ray peak or x - ray pre - peak ( see smith et al .
1999 ) , the travel time of the jets on sep .
16.02 was between 0.3 and 1.1 d. this results in an allowed range for @xmath60 of 224 to 806 mas d@xmath61 .
this implies @xmath62 .
the above constraint on @xmath63 then results in @xmath64 .
additionally , we can obtain a constraint on the distance from @xmath65 and @xmath66 . for @xmath67
mas d@xmath61 , @xmath68 kpc .
conversely , a distance of 0.4 kpc implies a jet ejection on sep . 15.4 for @xmath69 or sep .
14.9 for @xmath70 the angle @xmath53 is probably close to the inclination angle of the binary orbit , since one may expect the jets to be ejected close to perpendicular to the binary s orbital plane .
the high value of @xmath53 then suggests that at times the compact object should be eclipsed by the companion star .
our nfi data involve the longest continuous observation ever performed of while it was above the detection threshold ( all _ rossixte _ toos lasted shorter than 3 hr ) . during 19 hr
no eclipse was seen .
since the earth obscured the view to each _ bepposax _ orbit , there is some uncertainty in the non - detection of an eclipse . with this reservation ,
the non - detection suggests that either the binary has an orbital period in excess of 19 hr , that the jets are misaligned from the orbital rotation axis by more than @xmath110 ( like seems to be the case for the micro quasar gro j165540 , see van der hooft et al .
1997 ) , or that the companion star is not an a - type main sequence star .
the x - ray flux of behaved quite erratic .
whenever was detected , it exhibited at least 50% variability on hourly time scales .
the variability is not ( quasi ) periodic .
the scan observations with the pca show that the source turned on probably on feb .
18 and disappeared immediately after the giant outburst on sep .
15 . during the seven months in between
, the snapshot measurements revealed fluxes varying at least between an upper limit of 0.4 and detections of up to 12@xmath71 ( 2 - 60 kev ; markwardt et al .
1999b , c.b .
markwardt priv .
the two wfc - detected flares show that the flux sometimes reached flux levels one order of magnitude higher .
where does this variability come from and why did it turn off immediately after the giant outburst ?
the mass transfer from the companion star should be mediated either through an ( irregular ) wind from the companion star or an accretion disk .
which one applies to is unclear , given the observations thus far .
the high variability is perhaps most easily explained by an irregular wind from the companion star .
such wind accretion is very common among the high - mass x - ray binaries ( see review by white et al .
we note that shows similarities to ci cam .
ci cam was fast ( e - folding decay time 0.6 d , harmon et al .
1998 ) , exhibited a bright fe - k line ( ew up to 597 ev , orr et al .
1998 ) , was a radio jet source ( hjellming & mioduszewski 1998 ) , and was detectable in the 20 - 70 kev band ( harmon et al .
the optical counterpart of ci cam is a symbiotic b star with an irregular wind .
the alternative , in the accretion disk interpretation , is that disk or thermal instabilities are continuously important .
the apparent 0.3 to 0.9 day head start in the giant outburst of the optical to the x - ray emission suggests that a disk instability may be at work during this particular flare which moves from the outside to the inside of the disk , like has been seen in dwarf novae ( e.g. , meyer & meyer - hofmeister 1994 ) and in the lmxb black - hole transient gro j165540 ( orosz et al . 1997 ) . during the 5 days
before the giant outburst is continuously 2 mag brighter than immediately after the giant outburst when there is hardly any x - ray emission ( kato et al .
this can be explained as optical emission from a disk which disappeared with the giant outburst .
the alternative in the wind accretion interpretation is that the 2 mag come from the companion star .
this too is not unreasonable , increased levels of mass loss are likely to go hand in hand with such brightening . finally , there is the observation that the largest flare forced the system to go in quiescence .
again , this can be explained either way . in the accretion disk interpretation ,
the giant outburst drained the disk in such a manner that no accretion occurs anymore .
the implied time scale of this drainage is about 10 hr .
this can only be explained if the drained mass is small though enough to induce a strong flare .
the short distance enables one to set a rather strict limit on the flux for quiescent emission .
the pca monitoring program reveals that the flux went below an upper limit of 0.4@xmath71 ( 2 - 60 kev ) or 1 mcrab ( c.b .
markwardt , priv .
comm . ) . assuming the same spectrum holds as during our nfi observation ( when the flux was approximately 25 times higher ) , the 0.4 - 120
kev luminosity upper limit is @xmath72 .
unfortunately , this flux limit is too high to say something sensible about the nature of the compact object ( see , e.g. , rutledge et al .
1999 ) .
we reiterate the point made by wijnands & van der klis ( 2000 ) that this transient may be an example of a separate class of fast and faint x - ray transients whose existence has been proposed by heise et al .
( 1999 ) and is characterized by peak luminosities that are two orders of magnitude less than the typical bright soft x - ray transient and by outburst durations of order one week instead of months .
they estimate a rate of 18 such transients per year within 20 from the galactic center .
such transients are easily over - seen by surveying instruments with small duty cycles if they are at the usual few - kpc distances .
it is only because of its brightness due to its nearness that was easily noticed . at the distance to the galactic center
, the source would only have had a peak flux of 30 mcrab .
it is not clear whether the compact object is a neutron star or a black hole candidate .
no x - ray characteristics that unambiguously identify the nature , such as type i x - ray bursts or pulsations , are present .
optical doppler measurements during quiescence will give the best opportunity to obtain constraints on the mass of the compact object .
we thank frits paerels for useful discussions .
this research has made use of linearized and cleaned event files from the asi - bepposax sdc on - line database and archive system .
_ bepposax _ is a joint italian and dutch program .
|
is a nearby x - ray transient which exhibited a fast and large x - ray outburst on sep . 15 , 1999 ( smith et al . 1999 ) .
the wide field cameras and the narrow field instruments ( nfi ) on board _ bepposax _ observed at various stages of its activity before that , in the spring and fall of 1999 .
the fluxes range between 0.012 and 0.3 crab units ( 2 - 10 kev ) .
the nfi observation is unique because it is the longest semi - continuous observation of so far , and it offers a study of the spectrum at a relatively high resolution of 8% full width at half maximum at 6 kev .
we discuss the observations with emphasis on the x - ray spectrum .
a strong fe - k emission line was detected in with an equivalent width between 0.3 and 1 kev .
the line energy is up to [email protected] kev and suggests the presence of highly ionized iron .
we identify this as fluorescent emission from a photo - ionized plasma .
the continuum spectrum is typical for a low - mass x - ray binary in which emission from an accretion disk corona plays an important role .
there is no sign of an eclipse or periodic signal due to the binary orbit in this exposure , despite the fact that the twin jets seen at radio wavelengths suggest a high inclination angle .
|
in the context of the large effort performed within the lhc @xcite experiments in order to finalize the construction of the standard model ( sm ) and to shed some light on the building blocks of its future extension , the b physics program plays a very important role . in a complementary way to the direct searches of higgs boson(s ) and new particles , the new physics ( np ) processes can be studied by making precision measurements in the flavour sector , where we do expect significant effects
if the np scale is not completely decoupled from the tev energy range . in this framework , the b meson system is a natural place where to look .
the @xmath0 mesons properties have already been studied in detail by the b - factories experiments ( cleo , babar , belle ) yielding and impressive agreement with the sm expectations .
the properties of the @xmath1 mesons , not available at e@xmath2e@xmath3 colliders(5s ) .
the small statistics of @xmath1 mesons collected , will mainly be used to measure some absolute branching ratios .
] , are much less constrained and there s still room to account for significant contributions from np effects . moreover
, the study of the b system with high statistics samples of different b hadrons can be used to further constrain the sm cp violation mechanism described by the cabibbo - kobayashi - maskawa ( ckm ) matrix .
three lhc experiments , lhcb @xcite atlas @xcite and cms @xcite , have defined a research program that includes the search of rare decays and np signatures trough precision measurements of the meson decays .
while lhcb @xcite is devoted and optimized for the measurement of the b meson properties , the other two are general purpose experiments where the study of b mesons proceeds in parallel with the higgs boson and supersymmetry searches .
the doability of such b physics programs at the hadron machines is demonstrated by the recent results of tevatron experiments ( cdf , d0
@xcite ) with their nice measurements of @xmath4 @xcite and @xmath5 and @xmath6 @xcite .
the lhc environment , the expected resolutions of the various detectors and their triggering strategies are reviewed in section [ sec : lhc ] , while section [ sec : bprog ] contains the detail description of the physics program and of the expected performances that can be achieved with the data samples collected during the first year of operation of lhc .
the parameters summarizing the lhc environment for b physics studies ( center of mass energy , cross sections , luminosity ) are reported in table [ tab : kpar ] for the lhcb , atlas and cms experiments . the very large cross section for @xmath7 production , yielding @xmath8(10@xmath9 )
@xmath7 pairs per second ( l = 10@xmath10 @xmath11 s@xmath12 ) with all the b - hadrons species being produced , extend down to 10@xmath13 the branching ratio ( br ) measurement capability for the rare b decays .
for this aim a huge effort is needed in order to reduce the large background coming from the pp interactions , whose production cross section is two orders of magnitude higher than the b quarks production one .
attention should also be paid to the choice of running luminosity ( l ) : while larger l provides larger data samples , the number of interactions per crossing increase rapidly with l , creating problems to the b hadron reconstruction and identification ( see table [ tab : kpar ] ) .
the main difference in the visible cross section between the lhcb detector , a single arm forward spectrometer accepting events with 1.8@xmath145 , and the atlas / cms detectors , that are accepting events with @xmath152.5 , comes from the lower p@xmath16 threshold that can be achieved in the former .
the production of @xmath7 pairs in lhc , according to pythia simulations , occurs with the @xmath7 pairs highly collimated ( as shown in figure [ fig : bbcorr ] ) : this also favor the lhcb forward design with respect to the 4@xmath17 design of the others two experiments .
.center of mass energy , cross sections , luminosity and number of interactions per crossing expected in the lhcb , atlas and cms experiments . [ cols="<,^,^",options="header " , ] [ tab : btodpres ] the current prediction of @xmath18(@xmath7 ) @xmath19 500 @xmath20b comes from the extrapolations of tevatron results : a precise measurement of such cross section is of major importance in order to test the mc simulations , the nlo qcd calculations used for the extrapolations and the parton density functions ( pdf ) knowledge .
for the rare decays , np and new particle searches the knowledge of @xmath18(@xmath7 ) is also important to achieve a good estimate of the background level .
inclusive and exclusive strategies can be used to measure the cross section .
all the lhc experiments plan to use j/@xmath21 events to measure the production rate : the muons in the decay ensure an high efficiency in triggering and reconstructing those channels . in atlas and cms
experiments an additional requirement on the transverse energy ( e@xmath16 ) is used in order to enhance the b quark component in the data sample .
the statistical error that can be achieved is @xmath19 1% , already with small data samples ( l @xmath22 ) .
more detailed analyses , aiming at a full p@xmath16 scan , achieving a resolution of @xmath1910% in all bins , have been performed in cms assuming l @xmath19 10 fb@xmath12 .
the lhc program is proceeding without major delays : first beam is expected in september and the first collisions soon afterwards .
there are currently three experiments getting ready for the b physics challenge : with the expected performance they will be able to test sm and beyond sm ( np ) effects .
the first analysis approach reported here is related to the b@xmath23s observables , that can cleanly reveal some np effects .
a first goal of the lhc experiments is the measurement of the @xmath24 br down to its current sm prediction and of the @xmath6 angle with a 0.04 absolute precision .
these goals can be achieved with data collected during 2009 .
the lhcb detector will also address the study of b@xmath23dk and b@xmath23hh decay modes that can be used to reduce the ckm @xmath25 angle uncertainty down to 10@xmath26 , with data collected during 2009 .
the lhc experiments , whose b physics program highlights have been reported here , are ready to provide new measurements that will enrich the lhc program by exploring a new whole experimental region in the b quark phenomenology . the cdf collaboration , `` first flavor tagged determination of bounds on mixing induced cp violation in @xmath28 j / psi phi decays '' , http://arxiv.org/abs/0712.2397v1 . the d0 collaboration , `` measurement of @xmath29 mixing parameters from the flavor - tagged decay '' , http://arxiv.org/abs/0802.2255 .
d. atwood , i. dunietz , a. soni , phys .
* 78 * , 3257 ( 1997 ) and and phys .
d * 63 * , 036005 ( 2001 ) ; m. gronau and d. wyler , phys . lett .
b * 265 * , 172 ( 1991 ) ; m. gronau and d. london , phys .
b * 253 * , 483 ( 1991 ) ; a. giri , yu . grossman , a. soffer and j. zupan , phys .
d * 68 * , 054018 ( 2003 ) .
|
three experiments , among the lhc project , are getting ready to explore the b quark flavour sector . while atlas and cms are general purpose experiments , where the study of b mesons is going to proceed in parallel with the higgs boson and supersymmetry searches , the lhcb experiment is devoted to b physics studies .
the key parameters entering the physics analyses and the performances achieved in all the three experiments are presented .
given the large b physics program foreseen in the lhc experiments , the studies reported in this paper have been selected as those with higher likelihood to provide solid and interesting new results on standard model validation and new physics processes search with early data .
|
asca observations of the sgr b2 giant molecular cloud revealed very hard x ray spectrum with a very prominent iron fluorescent line at 6.4 kev ( koyama et al . , 1996 ) .
this discovery provided an important confirmation of the hypothesis of sunyaev , markevitch and pavlinsky ( 1993 ) that the diffuse emission from the giant molecular clouds in the galactic centre region is at least partly due to reprocessed emission of a powerful x ray flare from the supermassive black hole sgr a*. the geometry of the problem suggests that such flare could have happened few hundred years ago .
the morphology and the spectrum of the reprocessed emission have been modeled by sunyaev & churazov 1998 and murakami et al .
recent sax ( sidoli et al . , 2001 ) and chandra data ( murakami , koyama , maeda 2001a ) are broadly consistent with the assumption that reprocessed ( reflected ) emission is due to the past flare from sgr a*. main observational arguments in favor of this interpretation are : * remarkably hard shape of the sgr b2 x ray spectrum * extremely high flux in the neutral iron fluorescent line at 6.4 kev ( equivalent width @xmath012 kev ) . * the side of the cloud towards sgr a * is brighter in x
rays than the opposite side .
we argue below that the most convincing test of the origin of the diffuse x ray emission from the sgr b2 cloud would be a detection of polarized emission from this object .
recent progress in the development of the x ray polarimeters for space missions ( costa et al . ,
2001 ) implies a drastic increase of the detector sensitivity and makes the sgr b2 cloud a natural target for the polarimetric studies .
we consider a uniform cloud with the radius of 10 pc and the thomson optical depth of 0.5 exposed to the hard ( power law photon index @xmath1 ) unpolarized x
ray radiation from an external source . as we argue below the particular values of the cloud size and its optical depth
do not strongly affect the degree of the polarization of the reprocessed / reflected emission .
the external source was assumed to be steady on the time scales comparable with the light crossing time of the cloud ( the morphology of the reprocessed emission from the cloud illuminated by a short flare is considered in sunyaev & churazov 1998 ) .
the cloud consists of hydrogen ( in molecular form ) , helium and admixture of heavy elements with the solar abundance .
the reprocessed radiation was calculated via monte
carlo method .
the following three processes have been taken into account : i ) photoelectric absorption , ii ) compton scattering by bound electrons and iii ) fluorescent emission of heavy elements .
the cross sections for compton scattering by electrons in hydrogen molecules and helium atoms were taken from sunyaev & churazov 1996 , vainshtein , sunyaev & churazov 1998 and sunyaev , uskov & churazov 1999 .
the fluorescent yields and the energies of the fluorescent lines were taken from kaastra & mewe 1993 .
in particular case of the sgr b2 cloud ( as a reflection nebula ) and sgr a * ( as an external source ) the projected distance of @xmath0100 pc is known .
the cloud was assumed to be located at the same distance as sgr a * and outgoing photons , used for calculating degree of polarization , were accumulated over small solid angle towards the observer s direction . the resulting spectrum and the degree of polarization
are shown in fig.[sp ] .
for comparison the energy spectrum and the degree of polarization have been calculated for the case when scattering cloud is located at the same projected distance , but further 100 pc away from the observer than the illuminating source ( fig.[sp135 ] ) .
the spectra shown in fig.[sp ] and [ sp135 ] have a typical `` reflection '' shape of the continuum and a set of strong fluorescent lines of the astrophysically abundant elements like s , ar , ca , fe and ni . at low energies ( below @xmath07 kev ) the shape of the continuum emission
is determined by the ratio of the scattering cross section to the photoabsorption cross section . due to the coherent scattering by the electrons bound in hydrogen molecules and helium atoms
the scattered continuum is higher ( by a factor of 2 at lowest energies ) than the reflected continuum calculated for free electrons .
the `` compton shoulders '' on the red side of the strong lines are smooth and do not show sharp features typical for scattering by the free cold electrons ( sunyaev and churazov 1996 ) . for the iron 6.4 kev line the flux in the shoulder is @xmath01020 % of the line flux .
additional modification of the spectra shown in fig.[sp ] and [ sp135 ] should be due to interstellar absorption along the line of sight .
for the hydrogen column densities typical for the galactic centre region ( i.e. @xmath2 of @xmath3 ) the emission below 4 - 5 kev will be strongly suppressed . as is clear from fig.[sp ]
the continuum emission is strongly polarized , while the emission in the fluorescent lines is not .
this is of course a natural result given the geometry of the problem . for the neutral ( or molecular ) gas with solar abundance of heavy elements
the first scattering gives the dominant contribution to the reflected continuum at energies below @xmath07 kev . for a thomson thick cloud the contribution from second scattering
is attenuated by a factor @xmath4 , where @xmath5 is the thomson cross section and @xmath6 is the photoelectric cross section at given energy @xmath7 . for a cloud with a very small optical depth @xmath8 the suppression factor for second scattering
is @xmath9 .
therefore contribution of second scattering is almost always modest ( at least at low energies ) and reasonable estimate of the degree of polarization can be made under single scattering approximation .
the scattering cross section is proportional to the factor @xmath10 , where @xmath11 and @xmath12 are the polarization vectors of the photon before and after the scattering .
since the angular size of the reflecting cloud ( as seen from the illuminating source ) is small then the degree of polarization @xmath13 is related to the scattering angle @xmath14 via well known expression : @xmath15 where @xmath16 .
in particular case of the scattering by @xmath090simulated above ( fig.[sp ] ) , the degree of polarization should be close to unity .
the deviation of the degree of polarization from unity in simulations is caused mainly by the contribution of the second and subsequent scatterings to the total spectrum .
for the second simulation ( fig.[sp135 ] ) the scattering angle is around 135and the degree of polarization is close to 33% in agreement with the expression ( [ pol ] ) for rayleigh law of scattering . using again the argument that the angular size of the cloud ( as seen from the primary source ) is relatively small one can treat the incident radiation as a narrow unpolarized beam and get simple estimate ( see appendix ) of the degree of polarization for photons experienced two or more scatterings : @xmath17^{n-1}-1 ) } } , \label{poln}\end{aligned}\ ] ] where @xmath18 is the number of scatterings . from this equation
it is clear that the degree of polarization rather slowly decreases with the increase of the number of scatterings
. therefore even although the degree of polarization decreases with energy ( see fig.[sp ] ) due to increasing contribution from multiple scatterings , the reflected continuum radiation remains nevertheless strongly polarized at all energies .
the direction of polarization remains the same ( perpendicular to the direction from the primary source to the cloud ) after any number of scatterings .
fluorescent photons , which are born inside the cloud and are emitted isotropically should not be polarized as indeed seen in the simulations .
the compton shoulder of the fluorescent lines can be polarized for some specific geometries of the scattering clouds .
e.g. for an optically thin , strongly elongated cloud ( in the direction perpendicular to the line of sight ) the shoulder can be strongly polarized .
but in simple cases like considered here the shoulder can be polarized only weakly ( compared to the continuum ) .
thus sensitive polarimetric observations of the sgr b2 cloud should reveal strongly polarized continuum and unpolarized fluorescent lines .
the polarization direction must be perpendicular to the direction from the reflection nebula towards the primary source .
actual degree of polarization is a function of the scattering angle and therefore can be used to infer the mutual location of the sgr b2 cloud and the sgr a * source along the line of sight . for sgr b2 cloud the total continuum flux in the 4 - 10 kev band is 1.52 @xmath19 as observed by asca . for a 4% efficient detector
( see costa et al . ,
2001 ) placed under @xmath0 1000 @xmath20 effective area mirror this source would yield about @xmath21 in polarized flux . therefore ( assuming that the detector background can be neglected ) a significant detection can be achieved in few @xmath22 s. the sgr b2 is not the only object which is suspected to be the `` reflection nebula '' .
murakami et al .
, 2001b found strong fluorescent line emission from another rich molecular complex
sgr c. as argued by sunyaev , markevitch , & pavlinsky ( 1993 ) and cramphorn & sunyaev ( 2001 ) many other reflection sites could contribute to the diffuse emission in the galactic centre region and the galactic ridge .
x ray polarimetric studies would then allow one to reconstruct real 3-dimensional positions of the scattering clouds .
note however that ( in the first scattering approximation ) the degree of polarization is the same for the scattering angles of @xmath14 and @xmath23 , which correspond to the cloud shifted with respect to the primary source towards or away from the observer by the same distance . in order
to break this degeneracy one have to use detailed information on shape of the reflected continuum at low energies . for smaller scattering angles ( the scattering cloud is closer to the observer than the primary source ) the more distant side of the cloud is exposed to primary radiation and photoabsorption should suppress the reprocessed radiation at low energies .
more confusing can be the case when the emission of the primary source is already polarized .
for example this can be sgr a * jet x ray emission or emission from hot optically thin quasi - flat accretion disk ( sunyaev & titarchuk , 1985 ) .
the incident light is assumed to consist of unpolarized component with the intensity @xmath24 and the polarized component with the intensity @xmath25 .
the degree of polarization of the initial radiation is thus : @xmath26 .
the degree of polarization @xmath13 of the scattered continuum radiation ( in the one scattering approximation ) can then be easily calculated : @xmath27 where @xmath28 is the angle between the direction of polarization of the primary radiation and the perpendicular to the scattering plane .
this would make the determination of the cloud position with respect to the primary source more complicated .
however the continuum radiation will remain polarized in majority of cases and therefore the hypothesis of the sgr b2 as the `` reflection nebula '' can still be tested with the polarimetric observations .
polarization of the incident radiation will also affect shape of the continuum spectrum because it will modify the contributions of single and multiple scattered photons to the total spectra .
the equivalent width of the fluorescent lines will also change .
e.g. consider the single scattering approximation for the 90scattering .
for the completely polarized primary radiation the scattered continuum will be attenuated by a factor @xmath29 , while the intensity of the fluorescent line will not be affected .
thus the equivalent width of the line is : @xmath30 where @xmath31 is the equivalent width of the line in the case of unpolarized primary radiation .
therefore when the direction of the primary radiation polarization is in the scattering plane the contribution from the first scattering to the continuum spectrum is zero and the equivalent width of the iron line will be much larger than for unpolarized primary radiation . in reality presence of unpolarized component in the primary radiation and contribution from multiple scatterings
will limit the maximal value of the equivalent width .
finally we note that if the x ray flare from the sgr a * source was short compared to the light crossing time of the sgr b2 cloud ( see sunyaev & churazov 1998 ) then the time delay for photons undergoing two or more scatterings has to be taken into account when calculating the energy spectrum and the degree of polarization . in the first scattering approximation
the results are unchanged : the continuum is polarized in accordance with the scattering angle and the flux in the fluorescent lines is completely unpolarized .
we acknowledge helpful discussions with w. forman . this research was partially supported by the russian foundation for basic research ( projects 00 - 02 - 16681 and 00 - 15 - 96649 ) and by the program of the russian academy of sciences `` astronomy ( nonstationary astronomical objects ) '' chandrasekhar , s. 1950 , oxford , clarendon press , 1950 costa , e. , soffitta , p. , bellazzini , r. , brez , a. , lumb , n. , & spandre , g. 2001 , , 411 , 662 cramphorn , c. , sunyaev , r. 2001 , , submitted , astro - ph0108294 kaastra , j. s. & mewe , r. 1993 , , 97 , 443 koyama , k. , maeda , y. , sonobe , t. , takeshima , t. , tanaka , y. , & yamauchi , s. 1996 , , 48 , 249 murakami , h. , koyama , k. , & maeda , y. 2001a , , 558 , 687 murakami , h. , koyama , k. , tsujimoto , m. , maeda , y. , & sakano , m. 2001b , , 550 , 297 murakami , h. , koyama , k. , sakano , m. , tsujimoto , m. , & maeda , y. 2000 , , 534 , 283 sidoli , l. , mereghetti , s. , treves , a. , parmar , a. n. , turolla , r. , & favata , f. 2001 , , 372 , 651 sunyaev , r. a. & titarchuk , l. g. 1985 , , 143 , 374 sunyaev , r. a. , markevitch , m. , & pavlinsky , m. 1993 , , 407 , 606 sunyaev , r. a. & churazov , e. m. 1996 , astronomy letters , 22 , 648 sunyaev , r. & churazov , e. 1998 , , 297 , 1279 sunyaev , r. a. , uskov , d. b. , & churazov , e. m. 1999 , astronomy letters , 25 , 199 vainshtein , l. a. , sunyaev , r. a. , & churazov , e. m. 1998 , astronomy letters , 24 , 271
consider the narrow beam of unpolarized light in the scattering medium with the rayligh scattering phase function .
the light after @xmath18 scattering can be represented as a two component vector : @xmath32 where @xmath33 is the cosine of the angle with respect to the beam axis and @xmath34 and @xmath35 are the intensities of light with the directions of polarization perpendicular and parallel to the plane formed by the initial beam and the direction of the photon . from the symmetry of the problem it is clear that only two components are needed to fully describe the polarization .
the change of the polarization after additional one scattering is then governed by the simple transformation ( e.g. chandrasekhar , 1950 ) : @xmath36 the initial values of @xmath37 for a narrow beam of unpolarized light are obviousely : @xmath38 where @xmath39 is a dirac function . from ( [ ach ] )
it is clear that after any number of scatterings the intensities @xmath37 can be written in the form : @xmath40 e.g. for the radiation scattered once we obtain , by substituting ( [ am ] ) into ( [ ach ] ) : @xmath41 , @xmath42 .
i.e. @xmath43 ; @xmath44 ; @xmath45 . inserting the expression ( [ a2 ] ) into ( [ ach ] ) one
gets the recursive relation for @xmath46 , @xmath47 and @xmath48 : @xmath49 from ( [ arec ] ) it follows that @xmath50 .
the remaining two relations are then simplify to : @xmath51 using ( [ arec2 ] ) and known values of @xmath52 and @xmath53 we obtain the explicit expression : @xmath54 finally by substituting ( [ arec3 ] ) into ( [ a2 ] ) we can calculate the degree of polarization of photons experienced @xmath18 scatterings : @xmath55^{n-1}-1 ) } } , \label{afin}\end{aligned}\ ] ]
|
the sgr b2 giant molecular cloud is claimed to be an x
ray reflection nebula
the reprocessing site of a powerful flare of the sgr a * source , occurred few hundred years ago .
the shape of the x ray spectrum and the strength of the iron fluorescent line support this hypothesis .
we argue that the most clean test of the origin of x
rays from sgr b2 would be a detection of polarized emission from this source .
epsf # 1= # 1#2=.99 = .99 # 1#2#3=.31 = .31 = .31 [ firstpage ] polarization scattering ism : individual : sgr b galaxy : centre x - rays : general
|
the characterisation of the folding pathway of proteins is one of the fundamental problems in molecular biology and is under an increasing scientific attention due to the continuous advancements in experimental and theoretical biochemistry . after the work of anfinsen@xcite , who demonstrated that ribonuclease unfolds and refolds reversibly into its native ( active ) three - dimensional structure
, it has generally been accepted that the primary sequence usually contains sufficient information to direct the complete folding process .
what typically remains elusive to experimental and theoretical investigations is the pathway of this spontaneous process and the mechanisms that govern it .
a considerable progress in this direction is possible for proteins containing native disulfide bonds .
the formation of disulfide bonds during the folding process can be controlled experimentally through the use of an appropriate thioldisulfide redox couple and thiol quenching agent@xcite . by these means
the regeneration process can be halted , the intermediate species can be trapped , isolated and characterised .
historically , one of the most investigated proteins containing disulfide bonds has been the bovine pancreatic trypsin inhibitor ( bpti ) .
starting with the work of creighton @xcite a series of crucial studies have accumulated a wealth of evidence in favour of the existence of partially structured intermediates along the protein folding pathway . despite these efforts ,
the detailed characterisation of the intermediates turned out to be a delicate experimental matter and there is still not a universal agreement on whether intermediates contain native or non - native disulfide bonds@xcite and if there exists more than one pathway@xcite . in this context , the use of theoretical and computational tools @xcite has been extremely useful in complementing the experimental findings with a more precise characterisation of the folding pathway , albeit obtained for models that greatly simplify the complexity of the real system . in this paper
we propose a theoretical scheme to study the folding of proteins with disulfide bonds by suitably exploiting the ( known ) protein native structure . at a general level
our strategy falls in the class of approaches that build on the importance of the native state topology in steering the folding process@xcite , that is in bringing into contact pairs of amino - acid that are found in interaction in the native state . in the past few years
an increasing number of experimental and theoretical studies have confirmed the utility of these approaches in the characterisation of various aspects of protein folding processes @xcite . in the present work
we try to generalise this strategy by adding a suitable treatment of disulfide bonds accounting both for their strength as well as their capability to undergo intra - molecular reshuffling . in order to validate this approach
we investigate the folding process of the n - terminal core domain of hirudin hm2 from the blood - sucking leech hirudinaria manillensis @xcite .
hirudins are the most potent and specific inhibitors of thrombin ( a key enzyme in blood coagulation ) identified so far , and they are currently used as effective anticoagulants .
hirudin hm2 is composed of a compact n - terminal domain ( residues 1 - 47 ) stabilised by three disulfides ( cys6-cys14 , cys16-cys28 , cys22-cys37 ) and a highly flexible , negatively charged c - terminal tail .
structural studies conducted on several leech - derived disulfide - rich small proteins ( i.e. , hirudin , decorsin , and antistasin ) reveal that although these proteins display negligible sequence similarity and different function they share a common disulfide topology and 3d fold @xcite , suggesting that leeches use the same protein scaffold but different binding epitopes to affect hemostasis . notably , it has been found that these leech antihemostatic proteins possess a t - knot scaffold closely similar to that observed for other unrelated proteins , including b - transforming growth factors , wheat germ agglutinins and snake venom toxins @xcite . in this view , the results of our study may have relevant and more general implication on the elucidation of the folding pathway(s ) of other protein systems . the first experimental attempts to identify the folding pathway of hirudin date back to the studies of otto and seckler who argued that hirudin could be a viable alternative to bpti and showed that it was experimentally feasible to obtain and follow its unfolding / refolding processes@xcite . as for the case of bpti , also hirudin has been studied in different experiments @xcite leading to alternative formulations of its folding pathway @xcite . in particular when dissolved atmospheric oxygen was used as oxidising agent , the folding process appeared to occur first through the establishment of three non - native ( scrambled ) disulfides and later through their slow rearrangement into the native bonding pattern .
on the contrary , no evidence for the importance of these fully - oxidised disordered intermediates was found in the experiments of thannauser _ et al . _
carried out using oxidised dl - ditiothreitol ( dtt@xmath0 ) .
these alternative views prompted the present investigation of the hirudin pathway within a topology - based model . to ascertain that , despite its simplified nature , the model was suitable to characterize the main aspects of the folding process we have first tuned and validated it against the set of experimental measurements provided in ref .
@xcite . based on the success of this comparison
we have undertaken a detailed characterisation of the folding process by monitoring quantities inaccessible in previous experiments .
our study confirms the experimental evidence of ref .
@xcite suggesting that , under a certain set of experimental conditions , the rate - limiting step of the refolding process involves a transition from a species with two disulfides to ones with the three native disulfides@xcite .
a detailed analysis of the numerical dynamical trajectories further reveals the precise order of formation of the disulfide bonds .
the starting point of our analysis is the 1 - 47 fragment of hirudin@xcite resolved by nmr @xcite shown in fig .
[ fig : hirudin ] ( see methods and materials section ) .
the fragment under consideration contains three disulfide bonds between residues 6 - 14 , 16 - 28 , 22 - 37 and differs from the whole protein because it lacks the 18-residues long tail ; this tail plays an important role in the biological activity of the protein . however , it is highly mobile due to the virtual absence on any non - covalent contacts with the 1 - 47 globular fragment . for this reason we neglected the hirudin tail in our numerical characterisation of the folding process .
the effective energy scoring function that we adopted belongs to the class of topology - based hamiltonians @xcite .
the knowledge of the above mentioned native structure , @xmath1 , is exploited to construct an effective energy function that admits @xmath1 as the lowest energy state .
the simplest form of such energy function is as follows : @xmath2 where @xmath3 is the energy of a trial conformation @xmath4 and @xmath5 is its contact matrix , whose elements are 1 [ 0 ] if amino acids @xmath6 and @xmath7 are [ not ] in interaction . a standard criterion based on @xmath8 or @xmath9 distance of pairs of amino acids
is used to decide whether two amino acids are interacting . in the simple case of eqn .
( [ eqn : simple ] ) the energy minimum is attained when all native bonds are established .
such native interactions are weighted equally .
this is a simplification that can be justified when the effective amino acid interactions in the protein are of comparable strength and has the advantage of keeping the model transparent by avoiding the use of imperfect energy parametrisation @xcite .
it is important to notice that the equal weighting of native contacts does not imply that , in a thermalised ensemble , they are formed with equal probability .
in fact , it is the complex interplay of energy and structural entropy that dictates the most probable routes to the native state from an unfolded conformation as well as the presence of rate - limiting steps due to the establishment of crucial sets of native contacts@xcite . in the present study ,
the equal weighting of native contacts does not appear to be a good starting point since one has to account for the very energetic disulfide bridges that can occur between pairs of cys .
for this reason , in place of ( [ eqn : simple ] ) we adopt a scoring function consisting of two terms ( see also section methods and materials ) : @xmath10 @xmath11 enforces some general constraints on the peptide chain geometry and promotes the formation of native contacts between pairs of amino acids other than cys - cys .
these contacts are weighted in the same manner . the second term
, @xmath12 rewards the formation of disulfide bonds between pairs of cysteines .
it is important to stress that the formation of disulfide bonds is allowed among any pair of cys , not only the native ones . by doing
so we can investigate the extent to which species with one or more non - native disulfides are present and if they influence the dynamics toward the native state , as proposed in recent experiments .
the strength of the disulfide bonds relative to other non - covalent interactions is controlled by the parameter @xmath13 .
this parameter should not be regarded as a relative measure of the `` bare '' , i.e. in vacuum , disulfide strength . in fact , since our model does not treat the solvent explicitly , @xmath13 captures the effective strength of disulfide bonds in the presence of water and any other appropriate reducing / oxidising agent . for this reason the value of @xmath13 and
also that of the heat bath temperature , @xmath14 , have to be chosen in a suitable way so to reproduce as accurately as possible the conditions of a given experiment . in this study
we shall focus on the set of experiments carried out by thannhauser _
@xcite where hirudin was refolded in the presence of various concentrations of @xmath15 .
we shall first show that there exists a well - defined region in the @xmath16 parameter space where the rates of conversion between species with different numbers of disulfides match well those observed in experiments .
this is an important fact since it proves that , despite its simple form , the energy scoring function of eqn .
( [ eqn : ham ] ) can indeed be used to characterize the folding process obtaining the correct quantitative experimental picture .
based on such stringent validation of our strategy , we shall then monitor quantities that are inaccessible in current experiments and thus propose a vivid and detailed picture for the hirudin refolding process .
we want to stress the fact that the possibility to reproducing the results of thannhauser _ et al . _
@xcite for a suitable choice of the @xmath16 parameters support the hypothesis that our model is general enough to study any folding process involving the formation of disulfide bonds .
the experimental benchmark for our model is provided by a series of hirudin refolding experiments carried out by thannhauser _
@xcite under various concentrations of @xmath15 . by using several combined techniques
it was established that the refolding process occurs under the reaction shown in fig .
[ fig : model1 ] where @xmath17 , @xmath18 and @xmath19 denote respectively the species with 0 , 1 and 2 disulfides ( either native or non - native ) .
a certain species with three disulfides , denoted as @xmath20 was seen to convert with extreme rapidity to native hirudin and was , therefore , identified with a native arrangement of the three disulfide contacts .
the remainder of the ensemble of structures with three disulfides ( i.e. with at least two non - native disulfides ) is therefore denoted as @xmath21 .
the experimental characterisation of thannhauser _ et al _ provides the rates for the individual reactions of the above model for a few choices of the initial concentrations of @xmath15 and the reduced protein species .
our first goal was to see if , for suitably chosen values of @xmath14 and @xmath13 it was possible to reproduce such rates . , @xmath18 and @xmath19 , respectively .
the species denoted as @xmath20 contains the three native disulfides . the remainder of the ensemble of structures with three disulfides is therefore denoted as @xmath21.,width=278 ] several values of @xmath14 and @xmath13 were considered and , for each of them , we measured the rates of conversion between the @xmath17 , @xmath18 , @xmath19 , @xmath21 and @xmath20 species by using a monte carlo technique .
this was done by keeping track of how many transitions out of any given species occurred and how many of them ended up in each of the other species . to ensure that the obtained dynamics could be interpreted as a ( coarse grained ) viable dynamical trajectory @xcite only tiny local distortions of the peptide chain were attempted in the monte carlo moves . during such coarse - grained trajectories
the formation , breaking or reshuffling of disulfide bonds obeys a set of constraints dictated by the chemistry of the disulfide bonds and the thiol - disulfide coupling .
in particular no cys residue is allowed to participate to more than one disulfide ; the number of disulfide bonds can decrease or increase by at most one unit at each time step , respectively when an existing disulfide is broken or through the establishment of a new disulfide between two previously unbonded cys residues .
intramolecular rearrangements where one bond is broken in favour of a new one ( thus preserving the total number of disulfides ) are also allowed under the requirement that the new bond involves one of the cysteines of the broken disulfide .
a particular care is necessary to ensure that the monte carlo dynamics subject to these constraints does not violate detailed balance .
the difficulties arise from the fact that an elementary mc distortion of distinct starting configurations may result in structures that are compatible with a different number of allowed disulfide bonding patterns . to overcome this problem we have proceeded as follows : to a newly generated conformation @xmath4 we randomly associate one of the 15 possible pairing patterns of the three disulfides .
if the proposed bonding pattern of the new structure violates the previous criteria , the structure is rejected and time is incremented .
otherwise , the energy function is recalculated and the structure is rejected or accepted with the usual metropolis criterion ( see methods for further details ) . in this way ,
detailed balance is obviously satisfied , since the same number of bonding patterns is proposed for each structure .
the downside is that one encounters frequent metropolis rejections due to the `` blind '' proposal of bonding patterns .
for each of the explored values of @xmath14 and @xmath13 we have measured the correlation between the experimental rates and those obtained numerically . as a measure of the degree of correlation between these two quantities we used the non - parametric kendall analysis @xcite .
this statistical tool allows to establish if there is a relationship between two sets of data and how statistically significant it is .
being based on the comparative ranking of corresponding data in the two sets , the kendall analysis does not rely on any pre - assigned parametric dependence ( e.g. linear ) between the two quantities .
for this reason , it is regarded as a very robust measure of correlation and appears particularly appropriate in this context where the measured rates ( both experimental and numerical ) span several orders of magnitude @xcite .
our findings are summarised in fig .
[ fig : correl ] where we have shown the contour and density plot of the kendall correlation coefficient , @xmath22 , against the rates pertaining to the experimental conditions of fig .
5a in ref .
these effective first - order like rates where obtained starting from the best - fit experimental rates of table 2 in ref @xcite multiplied by the asymptotic concentration of dtt@xmath0 or dtt@xmath23 measured under the given experimental conditions ( see fig . 6 in the same reference ) .
a darker / lighter shadow in fig . [ fig : correl ] denotes the presence of a higher / lower degree of correlation .
it is apparent that the experimental data are well reproduced in the neighbourhood of @xmath24 , @xmath25 for which the highest correlation @xmath26 is observed . to ascertain the statistical significance of observing such correlation ,
we have computed the probability to observe , by pure chance a correlation larger than the observed one .
it turns out that this probability ( double - sided ) is equal to @xmath27 , which testifies the statistical significance of the observed correlation .
this establishes that there is a strong monotonic relation between the experimental and theoretical rates
. plane .
the white regions correspond to the minimum values of the correlations were observed ( @xmath28 ) . the highest correlation , @xmath29 ,
was observed in the dark area located around @xmath24 , @xmath25.,width=278 ] a scatter plot of the logarithm of experimental rates versus the numerically obtained ones is provided in fig .
[ fig : opt ] where the good interdependence of the data can be visually inspected .
if one were able to model the folding process with detailed and accurate energy functions one would expect an equality of the theoretical and experimental rates , a part from a time conversion factor . in our case ,
due to the simplicity of our model we do not observe such dependence but encounter , instead , another simple linear relationship between the logarithms of both sets of rates .
this is visible in the linear fit of figure [ fig : correl ] ; the corresponding correlation coefficient is @xmath30 .
its statistical significance ( two - sided ) over the set of 7 data points is 4% which is compatible with the significance of the more general ( and robust ) kendall correlation .
these values indicate that the correlation is highly significant from the statistical point of view .
a further validation can be carried out by comparing the asymptotic concentration of the various species observed in the experiment and in an equilibrated mc trajectory . in case of a perfect correlation between the experimental and numerical rates ,
this further check would be redundant . in this context ,
where the correlation is not perfect ( a significant discrepancy is seen for the transition between the @xmath21 and @xmath19 species ) this validation is useful to ascertain whether the differences in the rates result in significantly different equilibrium conditions .
the plot of fig .
[ fig : opt]b reveals that the asymptotic concentrations are in good accord and hence by setting @xmath31 and @xmath24 one can be confident that the mc trajectory is compatible both the dynamical and equilibrium properties of the experimental system .
( a ) , @xmath25 .
the experimental rates are expressed in min@xmath32 .
the points in the plot correspond to the following transitions a : @xmath33 , b : @xmath34 , c : @xmath35 , d : @xmath36 , e : @xmath37 , f : @xmath38 , g : @xmath39 .
( b ) scatter plot of the equilibrium fraction of the various species obtained in the monte carlo trajectory and in the experiment .
the experimental concentrations were extracted from fig .
5a of ref@xcite for @xmath40 s.,title="fig:",width=259 ] ( b ) , @xmath25 .
the experimental rates are expressed in min@xmath32 .
the points in the plot correspond to the following transitions a : @xmath33 , b : @xmath34 , c : @xmath35 , d : @xmath36 , e : @xmath37 , f : @xmath38 , g : @xmath39 .
( b ) scatter plot of the equilibrium fraction of the various species obtained in the monte carlo trajectory and in the experiment .
the experimental concentrations were extracted from fig .
5a of ref@xcite for @xmath40 s.,title="fig:",width=259 ] it is interesting to analyse the thermodynamic behaviour of the system at @xmath24 as a function of the heat bath temperature , @xmath14 . by using the standard ,
yet powerful method of histogram re - weighting ( see methods ) we have computed the average internal energy as a function of @xmath14 . the data , shown in fig .
[ fig : encvmu5](a ) indicate the presence of a point of inflection for temperatures close to 1 .
the presence of a transition at this temperature is further corroborated by the behaviour of the specific heat @xmath41 which displays a clear peak at @xmath42 , the folding temperature , and by the presence of two minima in the free energy profile in fig .
[ fig : encvmu5](b ) .
( b ) free energy profile , as a function of the internal energy , at the point @xmath24 , @xmath43.,title="fig:",width=259 ] ( b ) .
( b ) free energy profile , as a function of the internal energy , at the point @xmath24 , @xmath43.,title="fig:",width=259 ] this peak is associated to the folding transition in the system and its neatness suggests that the folding process has a two - state character @xcite .
this is indeed confirmed by an analysis of the free energy landscape at @xmath44 which exhibits two minima as a function of @xmath11 in correspondence of the unfolded and folded states ( data not shown ) .
the special point @xmath24 , @xmath25 , corresponding to the experimental conditions of ref .
@xcite appears therefore to be located slightly above the folding transition temperature ( for @xmath24 ) .
this is entirely consistent with the fact that the experimental conditions of ref .
@xcite , fig .
5a , do not particularly favour the formation of the native state which , indeed , involves only a few percent of the total equilibrium population .
we have portrayed in fig .
[ fig : freemu5 ] the free energy landscape for the different reduced species as a function of @xmath11 .
it can be seen that the free energy profiles have minima at different energy values according to the number of correctly formed disulfides .
the @xmath17 , @xmath18 , @xmath19 and @xmath21 species , in fact , display a minimum near @xmath45 which denotes an unfolded ensemble ( see fig . [
fig : encvmu5]b ) . on the contrary ,
the species with three native disulfides has a free - energy minimum for energies much closer to the native state energy . and
@xmath24 of the @xmath17 , @xmath18 , @xmath19 , @xmath21 and @xmath20 ensembles as a function of @xmath46.,width=278 ] the relative high values of free energy associated to the @xmath20 species reflect the particular experimental conditions reproduced here where the asymptotic fraction of species with native disulfides is low . clearly , by lowering the temperature one favours the formation of the native structure which is accompanied by an increase of the concentration with native bonding patterns for the cysteines .
this effect is visible in fig .
[ fig : ss]a where the average fraction of formed disulfide bonds is portrayed as a function of temperature .
it can be seen that above @xmath44 one has a significant formation of non - native bonds , that are superseded by native ones below @xmath44 .
an alternative picture to this one was put forward by chatrenet and chang based on a hirudin refolding experiment carried out in ref . @xcite . due to the much more oxidising conditions than the one considered in ref .
@xcite , chatrenet and chang @xcite observed that the folding of hirudin occurred through the formation of intermediates with three ( typically non - native ) formed disulfides . through a slow reshuffling process the disulfides would then rearrange in the native pairing from which the native conformation could be easily reached .
our findings indicate that this alternative scenario could be probably captured by our model by using larger values of the effective disulfide strength @xmath13 .
this is consistent with the different oxidising solvent conditions @xcite adopted by chatrenet and chang .
( a ) and ( b)@xmath47.,title="fig:",width=278 ] ( b ) and ( b)@xmath47.,title="fig:",width=278 ] to investigate this regime we explored the value of @xmath48 .
it is important to point out that the folding transition temperature ( identified through the location of the specific heat peak ) is almost insensitive to the disulfide strength in the range @xmath49 . by comparing the plots in fig .
[ fig : ss ] it is then possible to see that for the higher value of @xmath13 the total number of formed disulfide bonds is higher than for @xmath24 .
this result , accompanied by the fact that the total energy depends weakly on @xmath13 , confirms the intuition that , for large @xmath13 s the disulfides are established at early stages of the folding process when the rest of the protein is still unstructured .
this `` greedy '' disulfide formation not only impacts on the total number of formed disulfides but also on the relative fraction of correct ( native ones ) . as a result ,
a much higher fraction of wrong bonds is found at any temperature , as noticeable in fig .
[ fig : ss]b .
this picture suggests that the large @xmath13 regime of our model may be compatible with the scenario proposed by chatrenet and chang where the rate - limiting step corresponded to the disentanglement of non - native disulfides in intermediates states with three formed disulfide bonds .
however , for this alternative case the experimental rates are not available and we can not corroborate in a more quantitative way the parallel between the experimental conditions of chatrenet and chang and the `` large '' @xmath13 regime in our model .
so far we have focused on the overall thermodynamic characterisation of the refolding of hirudin , paying a particular attention to the validation of our model against experimental data .
having established that the detailed experimental reaction rates of ref .
@xcite can be well reproduced we can confidently use our model to investigate finer aspects of the refolding process .
we begin by examining the details of formation of the native arrangement of disulfides .
our interest is to find whether the formation of the @xmath20 species is aided by the establishment of some non - native disulfides that are later broken in favour of native bondings .
to do so , instead of subdividing the structures according to just the overall number of disulfides we indexed them with a pair of numbers , @xmath50 denoting the number of correct ( native ) disulfides , @xmath51 and wrong ( non - native ) ones , @xmath52 . in this way the fully reduced state , @xmath17 ,
is indicated as ( 0,0 ) while the @xmath20 state corresponds to ( 3,0 ) .
altogether there are nine possible states , as indicated in fig.[fig : model2 ] . we have generated several mc trajectories , each spanning about 20 million time steps ( see sections methods and materials ) at ( @xmath25 , @xmath24 ) and then recorded the probability of occurrence of all allowed transitions between the states of fig .
[ fig : model2 ] .
the typical dispersion on the measured rates over ten mc trajectories was typically less than 1% but augmented to about 10 % for the few very improbable transitions .
then we have considered all possible productive routes taking from @xmath17 to @xmath20 in a preassigned number of transitions .
the probability of occurrence for any such routes is obtained by multiplying the individual probabilities of any of the elementary steps . by considering productive routes involving an increasing number of transitions it can be established that the most probable route ( see fig .
[ fig : model2 ] ) is @xmath53 which is more than ten times as likely than the second ranking paths which is @xmath54 .
these findings seem robust against variations of the parameters in our model .
for example , even changing the temperature to 0.8 ( keping @xmath24 ) so that the formation of native species is strongly favoured ( see fig .
[ fig : ss]a ) it is found that the top ranking routes taking from the reduced state to the native one are the same as above .
interestingly , the relative weight ratio of the top productive routes is analogous to the one encountered for @xmath25 ; on the other hand the fact that the native formation is highly favoured at @xmath55 is reflected in an increase , by several orders of magnitude , of the weight of the productive routes .
we complete the present section by identifying the rate limiting steps of the folding process .
in a sequential reaction the rate limiting step is straightforwardly identified as the slowest one . in the present scheme
such simple analysis can not be carried out due to the presence of several alternative routes that can take to the native state . an objective and convenient way to identify the rate limiting step in such a situation is to identify the reaction whose rate change affects the most the production of the species of interest , in this case @xmath20 .
therefore , by using the measured rates for the transitions between the states of fig . [
fig : model2 ] we have integrated the associated master equations starting from a fully reduced population .
we have then changed by 10 % each of the rates and identified the time at which the concentration of @xmath20 crosses a pre - assigned threshold value . we found that , almost independently of the threshold value , the most sensitive step among all the allowed ones was @xmath56 which is the last step of the most probable route leading to @xmath20 . in principle
this may have not been the case , especially in the presence of equally important paths leading to @xmath20 .
the fact that the most probable routes include the rate limiting steps confirms the existence of a well defined succession of events taking to the native state . as mentioned before , based on experimental dynamical plots , thannhauser _ et al . _
@xcite had determined that the rate limiting step was the @xmath36 one . in their study
it was not possible to characterise by direct means whether the transition to the @xmath20 state occurred from a @xmath19 state comprising only native cysteine bonds , although this was reputed to be the most likely scenario .
our picture thus fully supports the experimental results concerning the @xmath36 rate - limiting step , but also adds novel insight in the process by indicating explicitly that the most crucial step involves a particular @xmath19 species , namely one with two native disulfides . in the remainder of this section
we shall further characterise the folding process by identifying which of the native disulfides are formed in the most probable routes ( or rate - limiting steps ) .
this is done by following a series of individual dynamical trajectories where the native conformation is reached starting from a reduced state .
the monte - carlo procedure allows for a detailed study of the folding pathways and for a systematic analysis of the specific order of formation of disulfide bonds . by setting @xmath24 the protein
is initially thermalised at an high temperature ( @xmath57 ) where , in our model , the disulfide bonds are completely reduced and then suddenly quenched to @xmath58 . for each kinetic trajectory the order of formation of the disulfide bonds was stored and a statistical analysis accomplished by comparing several runs .
in particular we have recorded the exact type of disulfide bridges forming the different species ( @xmath18,@xmath19 ) and from their relative concentration we have inferred the pathway .
the six cys residues of the fully reduced protein , 6 , 14 , 16 , 22 , 28 , 37 are equally likely to participate in forming the initial disulfide bond .
although in the early folding stages the contact @xmath59 appears quite frequently due to the short sequence distance between the amino - acids , after the molecule has equilibrated through a series of rapid internal disulfide interchange reactions , only 5 of the 15 possible @xmath18 states exist in significant quantities .
they are : ( 6 - 14 ) , ( 6 - 16 ) , ( 16 - 22 ) , ( 28 - 37 ) that are respectively present in the following equilibrium concentrations : @xmath60 @xmath61 @xmath62 and @xmath63 .
it is interesting to notice that the most common bond among the @xmath18 ensemble is a non - native one , ( 28 - 37 ) , while the only native bond that is present in significant quantities is ( 6 - 14 ) .
of the 45 possible @xmath19 states , seven occur in significant quantities .
two of them , ( 6 - 14;16 - 28 ) present in the @xmath64 of the cases and ( 6 - 14,22 - 37 ) present in the @xmath65 of the cases , are formed by native bonds .
other two , ( 6 - 14;16 - 22 ) and ( 6 - 14,28 - 37 ) present with concentration of @xmath66 and @xmath67 respectively , have a native and a non - native contact . whereas the last three , ( 6 - 16,28 - 37 ) with a frequency of @xmath68 , ( 14 - 22,28 - 37 ) with a frequency of @xmath60 and ( 16 - 22,28 - 37 ) with a frequency of @xmath60 , are formed by non - native disulfide bonds . the relatively high concentration of species with non - native states , which are not present in the most probable productive routes expected for the folding ( see the previous section ) , reflect the particular experimental conditions reproduced here ( induced by the choice of @xmath14 and @xmath13 ) in which the formation of the native state is not particularly favoured ( see fig . [
fig : opt]b ) .
the analysis of the folding pathway reported in the previous section and the statistical analysis of the dynamical productive pathway can be summarised in the following re - folding picture : the reduced proteins forms first the native contact @xmath69 and consequently either the states ( 6 - 14;16 - 28 ) or ( 6 - 14,28 - 37 ) almost with the same probability . the folding proceeds with the formation of the last disulfide bond .
this conversion is relatively slow in agreement with the finding of the previous section on the rate limiting step of the reaction .
the presence of conformation with three scrambled disulfide bonds turns out not to be statistically significant also from this kinetic analysis . a graphical representation of the most probable trajectory is shown in fig .
[ fig : pathway ] .
the fact that the majority of structures sampled in the quenching process involve non - native bonding patterns is not in contradiction with the findings of the previous section , where the most probable _ productive _ route was shown to be free of non - native bonds .
in fact , the high concentration of structures with non - native bonds does not automatically imply that they contribute significantly to the refolding `` flux '' towards the native state . on the contrary ,
species involving native bonds even if they do not accumulate to high equilibrium concentration , appear to take part to the most efficient routes leading to the native state .
the structure of the synthetic analogues of fragment 1 - 47 of hirudin hm2 in the free state was modelled on the nmr solution structure of natural fragment 1 - 47 @xcite , almost super - imposable to that of the corresponding segment ( 1 - 49 ) in intact hirudin hv1 variant@xcite ( pdb code : 5hir ) , showing @xmath70 sequence identity with hirudin hm2 , or to that of hv1 fragment 1 - 51 @xcite ( pdb code : 1hic ) . the best - representative model in the nmr ensemble
was selected using the program olderado @xcite available on - line at the site http://neon.chem.le.ac.uk . in our model
a generic conformation @xmath71 of the protein is modelled as a self - avoiding chain of connected @xmath8 atoms located at position @xmath72 , where @xmath6 is the amino - acid chain index ( ranging from 1 to 47 ) .
starting from the @xmath8 coordinates the peptide dihedral angles are calculated and hence , following standard geometrical rules @xcite we construct an effective @xmath73 centroid for all residues with the exception of the two end residues and the seven glycines ( indices : 10,18,23,25,34,40,42 ) .
the energy scoring function consists of two terms .
the first one incorporates a standard bias toward the formation of native non - disulfide bonds but also disfavours the formation of non - native bonds ( except for disulfide ones ) and penalises significant deviations from the native dihedral angles .
these details are known to improve the cooperatively of the modelled folding process @xcite .
the explicit expression of this term , evaluate on a trial structure @xmath4 is given by : @xmath74 \nonumber \\ & - 6 \left ( { \frac{r^n_{ij}}{r_{ij } } } \right)^{10 } \\ & + v_1 \ \sum^{'}_{i , j > i+2 } \left [ 1-\delta_{ij}(\gamma^{n } ) \right ] \left ( \frac{r^n_{ij}}{r_{ij } } \right)^{12 } \nonumber \\ & + v_2 \left [ \sum_{i=2,46 } ( \theta_i - \theta_i^{n})^{2 } + \sum_{i=3,46 } ( \phi_i - \phi_i^{n})^{2 } \right]\nonumber \\ & + v_{constraints}\end{aligned}\ ] ] where @xmath75 denotes the distance of the @xmath6th and @xmath7th @xmath8 atoms in the trial structure @xmath4 ; a superscript @xmath76 is used to denote analogous quantities pertaining to the native structure .
the contact map @xmath77 is computed by considering as threshold a distance of 8 @xmath78 between the @xmath8 atoms .
the prime in the summation indicates that no contribution is considered if both the amino - acids are cys .
finally , the angular term has been constructed using the standard dihedral angles , @xmath79 and @xmath80 @xcite .
the coefficients @xmath81,@xmath82 and @xmath83 are used to control the strength of interactions and are set equal to 1 , 5 and 1 energy units , respectively . the last term in the expression , @xmath84 ,
is used to enforce a series of knowledge - based constraints whose violation is penalised through an `` infinite '' energy penalty ( that is through a rejection of the violating conformation ) .
the constraints are as follows : ( 1 ) the distance between two consecutive @xmath8 atoms must remain in the interval 3.7 - 3.9 , ( 2 ) the distance between two non consecutive @xmath8 atoms must be greater than 4 , ( 3 ) the distance between any two @xmath9 must be greater than 2 @xmath78 and ( 4 ) the distance between any two @xmath8 and @xmath9 centroids must be greater than 2 .
the second term of the hamiltonian , @xmath12 rewards the formation of disulfides , irrespective of whether they are native or not . as explained in the next subsection , each proposed configuration , described in terms of @xmath85 s and @xmath73 s , comes with a set of three putative bonds among the six cysteines . of these putative bonds only those among residues
whose @xmath73 s are a separation smaller than 5 are considered to be effectively present and hence give a contribution equal to @xmath86 to the total energy .
as mentioned in the text , we used monte carlo dynamics for studying the folding process . at each monte - carlo step the current structure is distorted through local deformations @xcite based on two equally - probable moves : ( a ) _ single - bead move : _ a random @xmath87-carbon is chosen and is displaced randomly by at most 1 along each cartesian direction , ( b ) _ crankshaft move : _ two sites , @xmath6 and @xmath7 , with sequence separation at most 6 are chosen and all the sites between them are rotated around the axis joining @xmath6 and @xmath7 by an angle chosen randomly in the interval @xmath88 . as mentioned before , besides this structural rearrangement , at each monte carlo step we also associate a randomly - chosen pairing pattern for the six cysteines so that each of them is involved in a _ putative _ disulfide bond .
these bonds are termed putative because , to satisfy detailed balance , the pairing assignment is done blindly that is without inspecting whether a given pair of cysteines is at a distance compatible with the existence of a disulfide bond .
one then checks whether ( irrespective of the pairing distances ) the proposed bonding pattern is compatible with the undistorted configuration , i.e. if it respects the rules of section 2.1 about disulfide formation and thiol / disulfide coupling . if not the configuration is rejected , the monte carlo clock is advanced and a new distortion and bonding pattern is considered .
otherwise one proceeds as in ordinary metropolis schemes after having calculated the energy of the proposed configuration .
it is important to notice that this latter step involves the inspection of the distance of the putatively bonded cysteines to reward only those bonds that are geometrically feasible . the efficiency of the monte - carlo algorithm to study the thermodynamics was enhanced by the multiple markov - chain sampling scheme@xcite , a method that has proved quite effective in exploring the low temperature phase diagrams of proteins and interacting polymers .
all the runs have been performed by covering the temperature range of interest , @xmath89 $ ] , with 20 markov chains uniformly - spaced in temperature .
the data obtained in the multiple - markov - chain runs at different values of @xmath14 and @xmath13 were further processed through a generalised multiple histogram technique inspired by the work of ref .
this strategy allowed to reconstruct faithfully the density of states ( number of configurations ) in the multi - dimensional space of reaction coordinates constituted by @xmath11 and the number of correct ( native ) and wrong ( non - native ) disulfides : @xmath90 .
the data from the different runs were combined so to minimize the error propagation on the density of states .
the typical uncertainty of the reconstructed free - energy at the values of @xmath14 and @xmath13 considered here is about 0.5 energy unit ( this estimate follows from the analysis of free energy dispersion when half of the collected data is used ) . by these means it was possible to obtain the total energy and specific heat curves of fig .
[ fig : encvmu5 ] and the free energy profiles of fig .
[ fig : freemu5 ] .
we have proposed a theoretical framework to model and study the folding of proteins containing disulfide bonds .
the approach is based on the knowledge of the native state of a protein but contains an appropriate term to account for the possibility that native or non - native disulfide bonds can form .
the main advantage of the model proposed here is its simplicity which allows for a detailed description of all the folding pathways through the monitoring of the correct / incorrect contact formation .
the model has been validated by investigating the debated re - folding pathways of hirudin which has been object of several experimental studies .
it was shown that there exists a region in our two - dimensional parameter space where the rates of conversions between different oxidised species are in good agreement with experimental measurements @xcite .
starting from this successful comparison we have then attempted a detailed characterisation of the whole folding process . at a coarse - grained level our results
is consistent with the scenario described by thannhauser _
@xcite suggesting that the rate limiting step turns out to be @xmath91 . our approach , that allow for a precise identification of the formed contacts , shows clearly that the @xmath19 state appears to involve native intermediates , possibility that was reputed as the most probable situation in ref .
@xcite although experimentally it was impossible to show it directly .
the analysis of several folding trajectories also allowed the identification of the most probable folding route and the typical associated succession of formation of disulfide bridges .
furthermore , the thermodynamics of the system was elucidated by using statistical mechanical techniques to reconstruct the free energy profiles for the whole system and also for the different oxidised species . due to its simplicity ,
the proposed model , can not capture those aspects of the folding process that result from the delicate interplay of amino acid specific interactions .
the successful comparison of the theoretical predictions for hirudin with the experimental findings suggests that also , for disulfide containing proteins , suitable topology - based models can be profitably used to elucidate the folding pathways , even in the presence of non - native intermediates .
thus , the present approach , which is general and not specifically tailored for hirudin , ought to be straightforwardly applicable in other contexts providing a useful complement of experimental techniques in the characterisation of the folding process in the presence of disulfide bonds .
we thank f. cecconi , g.l . lattanzi and d. marenduzzo for stimulating discussions and suggestions . this work was supported by infm , fisr - miur 2001 and murst cofin 2001 .
clementi , c. , nymeyer , h. , and onuchic , j. topological and energetic factors : what determines the structural details of the transition state ensemble and en - route intermediates for protein folding ?
an investigation for small globular proteins .
jmb 298:937953 , 2000 .
cecconi , f. , micheletti , c. , carloni , p. , and maritan , a. molecular dynamics studies of hiv-1 protease : drug resistence and folding pathways .
proteins : structure function and genetics 43:365372 , 2001 .
filippis , v. d. , vindigni , a. , altichieri , l. , and fontana , a. core domain of hirudin from the leec hirudinaria manillensis : chemical synthesis , purification and characterization of a trp@xmath92 analog of fragment 1 - 47 . biochemistry 34:95529564 , 1995 .
micheletti , c. , seno , f. , banavar , j. r. , and maritan , a. learning effective amino acid interactions through iterative stochastic techniques .
proteins : structure function and genetics 42:422431 , 2001 .
folkers , p. , clore , g. , driscoll , p. , dodt , j. , koheler , s. , and gronenborn , a. solution structure of recombinant hirudin and the lys - > gln mutant : a nuclear magnetic resonance and hybrid distance geometry - dynamical simulated annealing study .
biochemistry 28:26012617 , 1989 .
szyperski , t. , guntert , p. , and et al .
, s. s. impact of protein protein contacts on the conformation of thrombin - bound hirudin studied by comparison with the thrombin - bound hirudin studied by comparison with the nuclear magnetic resonance solution structure of hirudin(1 - 51 ) .
jmb 228:12061211 , 1992 .
micheletti , c. , seno , f. , and maritan , a. recurrent oligomers in proteins - an optimal scheme reconciling accurate and concise backbone representations in automated folding and design studies .
proteins : structure function and genetics 40:662674 , 2000 .
|
a theoretical model for the folding of proteins containing disulfide bonds is introduced .
the model exploits the knowledge of the native state to favour the progressive establishment of native interactions . at variance with traditional approaches based on native topology ,
not all native bonds are treated in the same way ; in particular , a suitable energy term is introduced to account for the special strength of disulfide bonds ( irrespective of whether they are native or not ) as well as their ability to undergo intra - molecular reshuffling .
the model thus possesses the minimal ingredients necessary to investigated the much debated issue of whether the re - folding process occurs through partially structured intermediates with native or non - native disulfide bonds .
this strategy is applied to a context of particular interest , the re - folding process of hirudin , a thrombin - specific protease inhibitor , for which conflicting folding pathways have been proposed .
we show that the only two parameters in the model ( temperature and disulfide strength ) can be tuned to reproduce well a set of experimental transitions between species with different number of formed disulfide .
this model is then used to provide a characterisation of the folding process and a detailed description of the species involved in the rate - limiting step of hirudin refolding .
|
fig . 1 . the external periodic magnetic field as modeled by ( [ eq : field ] ) .
+ fig . 2 . the band structure for a spinless problem compressed in the ( longitudinal ) @xmath12-direction .
@xmath22 t , @xmath38 , @xmath24 , @xmath25 .
the states in the broad " bands ( upper quarter ) have a finite longitudinal mobility , while the longitudinal mobility of the states in the narrow " bands ( left and right quarters ) is infinitesimal .
the band structure on a bigger scale . only every fourth of the narrow " bands is shown .
the box in the middle is magnified for the spinless problem ( fig .
2 ) , and the problem with spin ( fig . 4 ) .
same as in fig . 2 but for the problem with spin .
fermi surface for @xmath39 a. u. only a part of the first brilloin zone is shown .
shaded areas are populated by electrons .
the curved lines of the fermi surface ( in the middle ) correspond to the states in the broad " bands , while the vertical lines on the extreme right and left represent the states in the narrow " bands .
|
we analyze the energy band structure of a two - dimensional electron gas in a periodic magnetic field of a longitudinal antiferromagnet by considering a simple exactly solvable model .
two types of states appear : with a finite and infinitesimal longitudinal mobility .
both types of states are present at a generic fermi surface .
the system exhibits a transition to an insulating regime with respect to the longitudinal current , if the electron density is sufficiently low .
+ pacs number(s ) : 75.50.rr , 75.70.ak = 8.75 in = -0.25 in = -0.75 in * electronic band structure in a periodic * * magnetic field * andrey krakovsky _ department of physics , new york university , new york , new york 10003 _ august 18 1995 the interest in the magnetoconductance properties of the two - dimensional electron gas in spatially
periodic lateral magnetic fields has been further stimulated by the recent experimental availability of such systems @xcite . in the work of carmona
et al .
@xcite spatial modulation of a magnetic field was produced by means of equidistantly located superconducting stripes where magnetic vortices were trapped by impurities resulting in periodic inhomogeneity of the external magnetic field , while in the work of of ye et al .
@xcite it was produced by deposition of ferromagnetic microstructures on top of the high - mobility two - dimensional ( 2d ) electron gas .
vast theoretical efforts on 2d electron gas in an inhomogeneous external magnetic field range from the theory of momentum - dependent tunneling through a magnetic barrier @xcite to properties of electronic states and transoprt in a weakly spatially modulated magnetic field [ 47 ] . in this short paper
we will be concerned with the one - electron energy band structure of the 2d electron gas under a periodic lateral magnetic field of an antiferromagnet , which is a limiting case of a strong periodic modulation .
we will show that two types of states appear : with a finite and infinitesimal longitudinal mobility .
both types of states are present at a generic fermi surface .
the system exhibits a transition to an insulating regime with respect to the longitudinal current if the electron density is sufficiently low .
the effect of a uniform magnetic field on energy bands produced by the periodic ( electric ) potential is well known @xcite .
the impact of the slightly inhomogeneous magnetic field on the landau levels of a free electron was considered by mller @xcite .
he showed that the energy bands exhibit a pronounced asymmetry in the lateral direction .
for a spatially modulated magnetic field a common theoretical model @xcite employs magnetic field perpendicular to the plane of the two - dimensional electron gas which has a carrier " field @xmath0 with a periodic modulation on top of it : @xmath1 in the work of peeters and vasilopoulos @xcite the effect of a periodic electric and weakly modulated magnetic field ( @xmath2 ) was considered .
they showed that the broadening of the landau levels is roughly proportional to the modulation amplitude @xmath3 .
the hofstadter - like " spectrum was obtained by wu and ulloa @xcite , and collective excitations were analyzed in @xcite by the same authors . in this work
we will deal with an electron gas confined to a plane in a perpendicular periodic magnetic field without a carrier " field .
in other words , in ( [ eq : period ] ) we take @xmath4 .
this corresponds to the extreme case of the other limit , @xmath5 .
such a periodic field will create an energy band structure of its own .
we see this type of arrangement experimentally realizable by bringing a two - dimensional electron gas in close contact with a mesoscopic longitudinal antiferromagnetic ( sandwich ) structure , without an external magnetic field .
when dealing with the one - electron spectrum it is useful to have some exactly solvable models ( potentials ) , as they elucidate the whole structure of the energy bands @xcite .
below we show that the energy bands can be obtained exactly in a simple way for a reasonably idealized periodic magnetic field .
we present full band picture for both electron with spin and spinless , and discuss the topology of the fermi surface .
the hamiltonian for a free spinless electron in a magnetic field is : @xmath6 in our case the electron is confined to a plane , and a periodic magnetic field of lateral antiferromagnet is superimposed .
the magnetic field can be modeled as ( see fig .
1 ) @xmath7 { \bf \hat{z}}\ , , \label{eq : field}\ ] ] for such a magnetic field the vector potential takes a form @xmath8 where @xmath9 we proceed with solution of eqs .
( 2)(6 ) in a standard way .
we look for the solution in a form @xmath10 thus , in the gauge ( [ eq : gauge ] ) the solution is a plane wave in the @xmath11-direction . for the @xmath12-dependent part of the wave function @xmath13 we arrive at @xmath14 \chi ( y ) = \left(e - \frac{k_x^2}{2 } \right)\chi ( y)\ , .
\label{eq : kronig}\ ] ] here atomic units are adopted ; @xmath15 is the energy up to an unimportant constant ; @xmath16 , the dimensionless magnetic length " is given by @xmath17 , @xmath18 is the bohr radius .
the eq .
( [ eq : kronig ] ) is precisely the schrdinger equation for the kronig - penney model , and can be easily solved exactly .
the resulting dispersion relation is given by @xmath19 where @xmath20 @xmath21 is the quasimomentum in the longitudinal direction .
the band structure for @xmath22 t , @xmath23 m , @xmath24 , @xmath25 is presented on figs . 2 and 3 .
it is compressed in the ( longitudinal ) @xmath12-direction . as was pointed out in @xcite ,
the pronounced asymmetry along the @xmath11-direction is the signature of the energy spectrum in the inhomogeneuos magnetic field . in the upper quarter of fig .
2 is the region where broad " bands are formed .
these bands have a finite width in the @xmath12-direction , and the particles occupying these states will have a finite mobility in the longitudinal direction .
the other set of narrow " bands occupies the left and right quarters of fig . 2 .
from the point of view of the kronig - penney model ( [ eq : kronig ] ) they correspond to the valence bands of the periodic potential .
these bands are infinitesimally narrow in the @xmath12-direction , and electrons populating them would have a vanishingly small longitudinal mobility . of course , in the transverse direction states in both types of bands would have some finite mobility . fig .
3 represents the same band structure on a bigger scale . the part of the spectrum shown on the fig .
2 corresponds to the area inside the box of fig .
3 .
dark areas of fig .
3 represent regions of broad " bands , while the parabolas represent narrow " bands . in order not to overcomplicate the picture
we show only every fourth of the latter .
as we will see below , the peculiarity of the energy spectrum in a periodic magnetic field will appear in the fact that at the fermi surface both types of states will appear . in this model
the problem of electrons with spin is equally easy to treat .
this amounts to simply adding the spin - dependent term to the left - hand side of eq .
( [ eq : kronig ] ) : @xmath26 \chi ( y ) = \left(e - \frac{k_x^2}{2 } \right)\chi ( y)\ , ; \label{eq : kronig.sp}\ ] ] which results in a slightly modified dispersion relation : @xmath27 with the same @xmath28 and @xmath29 as in eqs .
( [ eq : alpha ] ) , ( [ eq : beta ] ) . the detailed band structure for the same magnetic field as before
is shown on fig .
4 .
the main structure of the whole spectrum is still represented by the fig .
3 .
as compared to the spinless problem , the broad " bands are characterized by wider gaps in the density of states , while narrow " bands only slightly change their locations .
as we fill the spin up " and down " states up to the fermi level , the ground state exhibiting transverse oscillatory spin oscillations in the spirit of the ones discussed by chudnovsky @xcite may result .
possible fermi surface corresponding to cutting the energy manifold at @xmath30 a.u .
is presented on fig . 5 .
we show only a part of the first brilloin zone .
shaded areas are populated by electrons .
curved lines of the fermi surface ( in the middle ) correspond to states in the broad " bands . as discussed above
, these states have a finite mobility in both directions and will always contribute to the conductivity of the sample .
the vertical lines on the extreme right and left correspond to the sections of narrow " bands which are flat in the longitudinal direction . thus , these states will not contribute to the longitudinal current , while always contributing to the transverse one .
if the electron gas is dilute enough so that the fermi level drops below the @xmath31 level ( see fig .
3 ) , the sample will not conduct in the longitudinal direction at all as all the states at the fermi surface will have a vanishing mobility in the @xmath12-direction .
it is easy to estimate the electron density for transition to an insulating regime by counting the states with @xmath32 . in our range of parameters
we can totally neglect the width of the narrow " bands .
the transition density is given by @xmath33 ( all quantities are in atomic units ) .
the summations are over all @xmath34 for which the radicals remain positive . for our values of @xmath35 , @xmath36 , and @xmath3 , @xmath37 .
if any of the summations turns out to be restricted to the first narrow " band , one has to account for the bandwidth in the @xmath12-direction as well . in a realistic experimental situation
many - body effects will be present .
the simplest of them is screening .
screening will smear " the effective single - particle potential , which may result in suppressing smaller gaps predicted in the calculation .
however , these effects do not change the overall structure of the spectrum . in conclusion ,
we have presented a simple exactly solvable model of the electronic band structure in a spatially periodic magnetic field .
all the conclusions derived from the model are not restricted to this particular model , but illustrate the general structure of the energy bands of the two - dimensional electron gas in a periodic lateral magnetic field .
this type of system can be realized by imposing magnetic field of a longitudinal antiferromagnet on a high - mobility two - dimensional electron gas .
the band structure exhibits a pronounced asymmetry in the lateral direction and consists of the two types of bands .
the states in broad " bands will have a finite mobility in the longitudinal direction while the longitudinal mobility of the electrons occupying states in narrow " bands is infinitesimally small .
a generic fermi surface will contain both types of states .
if the electron density is sufficiently low , only the narrow " bands will be occupied resulting in vanishing of the longitudinal conductivity . the electron density for transition to the insulating regime has been estimated .
an interesting extension of this work is to account for screening in such a system .
the author would like to thank professors a. d. kent , p. m. levy , j. k. percus , and especially j. l. birman for useful discussions .
partial support of the physics department at new york university is gratefully acknowledged .
9 h. a. carmona , a. k. geim , a. nogaret , p. c. main , t. j. foster , m. henini , s. p. beaumont , and m. g. blamire , _ phys .
rev.lett . * 74 * _ , 3009 ( 1995 ) .
p. d. ye , d. weiss , r. r. gerhardts , m. seeger , k. von klitzing , k. eberl , and h. nickel , _ phys .
rev .
lett .
* 74 * _ , 3013 ( 1995 ) .
a. matulis , f. m. peeters , and p. vasilopoulos , _ phys .
rev .
lett .
* 72 * _ , 1518 ( 1994 ) .
j. e. mller , _ phys .
rev .
lett . * 68 * _ , 385 ( 1992 ) . f. m. peeters and p. vasilopoulos , _ phys .
rev.b , * 47 * _ , 1466 ( 1993 ) .
x. wu and s. e. ulloa , _ phys .
rev .
b , * 47 * _ , 10028 ( 1993 ) .
x. wu and s. e. ulloa , _ phys .
rev .
b , * 47 * _ , 7182 ( 1993 ) .
d. r. hofstadter , _ phys .
rev .
b , * 14 * _ , 2239 ( 1976 ) .
for the recent work see o. khn , v. fessatidis , h. l. cui , r. e. selbmann , and n. j. horing , _ phys .
rev .
b , * 47 * _ , 13019 ( 1993 ) ;
p. w. wiegmann and a. v. zabrovin _ phys .
rev .
lett . * 72 * _ , 1890 ( 1994 ) ; y. hatsugai , m. kohmoto , and y .- s .
wu _
phys .
rev .
lett . * 73 * _ , 1134 ( 1994 ) .
j. c. slater , _ phys .
rev .
* 87 * _ , 807 ( 1952 ) .
e. m. chudnovsky , _ europhys .
lett .
* 30 * _ , 175 ( 1995 ) .
|
the long standing efforts for building cherenkov detectors in sea depths started with the dumand project that measured the muon vertical intensity at depths ranging between 2 and 4 km with a prototype line at a depth of 4.8 km about 30 km off - shore the island of hawaii in nov .
1987 @xcite .
the project was canceled in 1995 , while in 1993 a first configuration of 36 phototubes ( pmts ) on 3 strings ( nt-36 ) was installed in lake baikal in siberia at the shallow depth of 1.1 km , 3.6 km off - shore @xcite .
the latest configuration of the experiment ( nt-200 + ) was put into operation in apr .
2005 , with an umbrella - like structure of 8 strings 72 m long , with pairs of up - looking and down - looking pmts with 37-cm photocatodes .
three external strings at 100 m from the center of nt200 have been added to increase the sensitivity at very high energies for cascades .
lake water compared to sea water offers the advantage of being a quieter environment and the frozen surface in winter can be used as a platform for deploying and replacing broken components .
nonetheless , the optical background can be more seasonal dependent than in sea water and the absorption length for blue light ( 480 nm ) is about 20 m , while in sea water it is typically between 50 - 60 m. both for sea water and lake water the scattering length , which mainly affects the angular resolution of neutrino telescopes , is between 15 - 70 m , but the angular distribution of scattered photons is very forward peaked ( @xmath0 ) so that the effective scattering length ( @xmath1 can be up to 200 - 300 m. the nestor collaboration begun surveys of an area close to the peloponnese coast at a mean depth of 4 km in the 90s .
they measured recently the muon flux with a hexagonal floor with 12 up - looking and down - looking pmts @xcite and it is studying the possibility to deploy towers of 12 of these floors .
[ fig1 ] summarizes the muon flux measurements of all these projects .
it is noticeable how the agreement of the measurement has improved with time . in the plot
, we also show the results for the first line of antares detector described below @xcite .
the antares collaboration , joining about 150 physicists , engineers , and sea experts from 7 european countries ( france , germany , italy , romania , russia , spain , the netherlands ) and 24 institutions , formed in 1996 .
marine properties of the antares site off - shore la seyne sur mer ( south france ) were studied between 1996 and 2000 in about 45 sea campaigns .
the properties of light propagation at the antares site , 2475 m below the sea level , are summarized in ref .
@xcite : the absorption length in the blue is about 60 m while the effective scattering length is about 250 m. the milestones of the detector construction are the deployment of the 45 km - long main electro - optical cable ( eoc ) in 2001 and of the junction box ( jb ) at its end in 2002 . between 2003 and 2005 various prototype lines and a mini - instrumentation line for environmental parameter measurements where deployed @xcite and in 2006 the first 2 lines of the detector . in 2007
further 8 lines where deployed and on may 30 , 2008 the whole detector made of a total of 12 lines was put into operation .
a scheme of the detector and of the line layout is shown in fig .
lines were connected during 5 submarine operations , one conducted by a submarine and all others by an unmanned remote operated vehicle .
antares is made of 12 lines held taught by buoys and anchored at the sea floor connected to the jb that distributes power and data from / to shore .
the instrumented part of the line starts at 100 m above sea level so that cherenkov light can be seen also from the region below .
lines are separated by 60 - 75 m from each other and each of the lines holds 25 floors called storeys .
each storey has 3 pmts ( hamamtsu 10 " @xcite ) looking downward at @xmath2 from the vertical housed inside pressure resistant glass spheres made of two halves closed by applying an under - pressure of 200 - 300 mbar .
the set - up including the pmt , the glass sphere , the silicon gel for optical coupling between the glass and the photocathode and the mu - metal cage for shielding the earth magnetic field is called optical module ( om ) @xcite .
storeys also include titanium containers housing the frontend electronics with a pair of asic chips per pmt used for signal processing and digitization that provide the time stamp and amplitude of the pmt signal .
each of the oms contains a pulsed led for calibration of the relative variations of pmt transit time and a system of led and laser optical beacons allows the relative time calibration of different oms .
an internal clock system distributes from shore the 20 mhz clock signal , that is synchronized by gps to the universal time with a precision of @xmath3 ns .
time calibrations allow a precision at the level of 0.5 ns ensuring the capability of achieving an angular resolution at the level of @xmath4 for muons above 10 tev @xcite .
antares is equipped with a positioning system that includes tiltmeters and compasses giving the orientation of storeys and an acoustic triangulation system of hydrophones and transceivers that reconstruct line shapes .
measurement of line displacements show that lines move coherently in the radial direction and that the achievable precision on the relative position is better than 10 cm .
this information can be stored in databases and used by the offline event reconstruction .
all data ( @xmath5 gb / s ) are sent to shore satisfying a l0 condition requiring hits above a threshold of 1/3 of a photoelectron ( pe ) .
a l1 is formed by a local coincidence of 2 l0 in a storey in a 20 ns window or a large pulse above a few pe @xcite .
this data flow rate is reduced by about a factor 1000 by a filter running on a pc farm on shore that looks for 5 l1 and a casual connection between l1 compatible with a light signal produced by a ultra - relativistic muon .
the measured muon trigger rate is about 3 hz in the 10 lines .
in addition dedicated triggers have been developed for gamma - ray bursts ( grbs ) and the galactic centre .
a continuous record of up to 100 s stores temporarily data in memory that can be saved on disk in case of an external trigger from a network of satellites .
the filter is necessary since sea water is an optical noisy environment due to bioluminescence and @xmath6 decay of @xmath7 producing electrons that annihilate and hence produce two photons above cherenkov threshold . in the 10 line configuration we observed that during 80% of the time the optical background rate is about 60 khz .
dedicated calibration runs are taken measuring the coincidence rate of @xmath7 in the same storey and allowing to monitor om efficiency at the level of 5% accuracy .
these efficiencies are used in the measurement of the rate of coincidences in the same storey versus depth as shown in fig .
[ fig3 ] ( on the left ) .
the smooth rate decrease is a measurement of muon energy losses in the detector and a demonstration that sea water is to a large extent a depth independent medium , unlike ice .
dark matter searches are also possible by looking for an excess of neutrinos from celestial bodies like the sun or the galactic centre .
neutrinos with energy below tev could be produced by the annihilation of supersymmetric dark matter particles like the neutralino , that would become gravitationally trapped in this bodies . in fig .
[ fig5 ] we show the number of muon neutrino events that could be measured in 3 yrs by antares in the direction of the sun as a function the neutralino mass .
each of the point corresponds to a supersymmetric model .
|
antares is the first undersea neutrino telescope .
it is in its complete configuration since may 2008 at about 2.5 km below the sea surface close to marseille .
data from 12 lines are being analyzed and are producing first results .
here we discuss first analysis results for 5 lines and 10 lines , and we also comment on the performance of the full detector . we show that the detector has capabilities for discriminating upgoing neutrino events from the much larger amount of downgoing atmospheric muons and that data and simulation are in good agreement .
we then discuss the physics reach of the detector for what concerns point - like source and dark matter searches .
|
the nonet of light scalar mesons formed by @xmath2 , @xmath3 , @xmath0 and @xmath4 is poorly understood @xcite . compared to expectations based on a standard quark antiquark picture
all nine states are rather light and their ordering is inverted . this can , however , naturally be explained assuming a tetraquark structure , which is also supported by certain decay channels , e.g. @xmath5 . here
we report about the status of an ongoing long - term project with the aim to study such tetraquark candidates using lattice qcd .
we mainly focus on the @xmath0 . in section [ sec001 ]
we summarize recently published results obtained with wilson twisted mass quarks , where diagrams with closed fermion loops ( also called `` singly disconnected diagrams '' ) have been neglected @xcite . in section [ sec455 ]
we discuss latest technical advances obtained with clover improved wilson quarks , in particular the inclusion of diagrams with closed fermion loops .
finally , in section [ sec769 ] we discuss first steps of a study of a possibly existing @xmath1 tetraquark , which has recently been predicted using a coupled system of covariant bethe - salpeter equations @xcite .
we use gauge link configurations with 2 + 1 + 1 dynamical quark flavors generated by the etm collaboration @xcite .
we consider several ensembles with lattice spacing @xmath6 .
the ensembles differ in the volume @xmath7 and the unphysically heavy @xmath8 quark mass corresponding to @xmath9 . for the computations presented in this section we have ignored diagrams with closed fermion loops , which are technically rather challenging ( cf .
section [ sec500 ] ) .
physical consequences are discussed in detail in @xcite .
@xmath0 has quantum numbers @xmath10 .
as usual in lattice qcd we extract the low lying spectrum in that sector by studying the asymptotic exponential behavior of correlation functions @xmath11 @xcite . @xmath12 and
@xmath13 denote suitable creation operators , i.e. operators generating the @xmath0 quantum numbers , when applied to the vacuum state . assuming that the experimentally measured @xmath0 with mass @xmath14 is a rather strongly bound four quark state , suitable creation operators to excite such a state are @xmath15 the first operator has the spin / color structure of a @xmath16 molecule .
the second resembles a bound diquark antidiquark pair , where spin coupling via @xmath17 corresponds to the lightest diquarks / antidiquarks ( cf .
e.g. @xcite ) .
further low lying states with @xmath0 quantum numbers , which need to be considered are the two - particle states @xmath18 and @xmath19 .
suitable creation operators to resolve these states are @xmath20 figure [ f001](left ) shows effective mass plots for a @xmath21 correlation matrix with a @xmath16 molecule operator ( [ eqn001 ] ) and a diquark - antidiquark operator ( [ eqn002 ] ) ( lattice volume @xmath22 , pion mass @xmath23 ) .
the corresponding two plateaus are around @xmath24 and , therefore , consistent both with the expectation for possibly existing @xmath0 tetraquark states and with two - particle @xmath18 and @xmath19 states , where both particles are at rest ( @xmath25 ; @xmath26 in our lattice setup ) . increasing this correlation matrix to @xmath27 by adding @xmath18 two - particle and @xmath19 two - particle operators ( eqs .
( [ eqn003 ] ) and ( [ eqn004 ] ) ) yields the effective mass results shown in figure [ f001](right ) .
two additional states are observed , whose plateaus are around @xmath28 . from these plots
one can conclude that the two low - lying states are the expected two - particle @xmath18 and @xmath19 states , while an additional stable @xmath0 tetraquark state does not exist .
the second and third excitation can also be interpreted as two - particle states , which have non - vanishing relative momentum .
a detailed discussion of these plots and corresponding arguments can be found in @xcite .
qualitatively identical results are found , when using other ensembles with light quark masses and spacetime volumes as mentioned in section [ sec388 ] .
recently we started to perform similar computations using gauge link configurations generated by the pacs - cs collaboration @xcite with 2 + 1 flavors of clover improved wilson sea quarks , a lattice spacing @xmath29 , a volume @xmath30 and a @xmath8 quark mass corresponding to @xmath31 . a significant advantage compared to wilson
twisted mass quarks is that parity and isospin are exact symmetries .
for example there is no pion and kaon mass splitting and @xmath32 and @xmath33 states are separated by quantum numbers ( these problems in the context of wilson twisted mass quarks and the @xmath0 are discussed in detail in @xcite ) .
a major improvement compared to our previous results presented in section [ sec377 ] is that this time diagrams with closed fermion loops are included , i.e. where strange quark propagators start and end at the same timeslice .
ignoring such diagrams introduces a systematic error , which is hard to quantify . with these new computations
this systematic error has been eliminated .
another important consequence of the inclusion of closed fermion loop is that quark - antiquark and four - quark trial states can have non - vanishing overlap .
this allows to study a larger correlation matrix containing not only the four - quark operators ( [ eqn001 ] ) to ( [ eqn004 ] ) , but also a quark - antiquark operator @xmath34 such a correlation matrix will enable us to make stronger statements about the structure of states from the @xmath0 sector .
the techniques we use to compute these diagrams with closed fermion loops are described in @xcite . at the moment
only the computation of a @xmath21 correlation matrix containing the @xmath35 operator ( eq . ( [ eqn645 ] ) ) and the four - quark @xmath16 molecule operator ( eq . ( [ eqn001 ] ) ) has been finished .
the effective masses corresponding to the diagonal element of the @xmath35 operator is plotted in figure [ f004 ] , left .
the plot on the right shows the two effective masses obtained by solving a generalized eigenvalue problem using the full @xmath21 correlation matrix .
all effective masses are around @xmath36 and , therefore , consistent with @xmath37 , with @xmath38 and also with @xmath39 .
effective masses as functions of the temporal separation . *
( left ) * @xmath35 operator . *
( right ) * @xmath21 correlation matrix ( operators : @xmath35 , @xmath16 molecule ) . , title="fig:",width=264 ] effective masses as functions of the temporal separation . * ( left ) * @xmath35 operator . *
( right ) * @xmath21 correlation matrix ( operators : @xmath35 , @xmath16 molecule ) .
, title="fig:",width=264 ] for statements regarding the structure of the observed states we need to include also the four - quark operators ( [ eqn002 ] ) to ( [ eqn004 ] ) , which is currently in progress .
this will allow us to address similar questions as in section [ sec377 ] , i.e. to answer , whether there is in addition to the two - particle @xmath18 and @xmath40 states also a third ( bound ) state near the mass of @xmath0 .
we also intend to study the volume dependence of the two - particle spectrum using `` lscher s method '' ( cf .
e.g. @xcite ) or improved techniques taking e.g. coupled channel scattering into account @xcite .
such computations are very challenging using lattice qcd , but first results have recently been published , e.g. for the @xmath41 and positive parity @xmath42 mesons @xcite .
recently a @xmath43 tetraquark has been predicted using a coupled system of covariant bethe - salpeter equations @xcite .
it is expected to be a mesonic molecule - like state formed by two @xmath44 mesons with a mass of @xmath45 . comparing this mass to @xmath46
indicates that the binding energy could be quite large , @xmath47 .
the uncertainty quoted in @xcite is , however , of the same order of magnitude .
investing such a @xmath43 tetraquark using lattice methods could provide an independent confirmation of its existence .
we use the same setup as discussed for the @xmath0 meson in section [ sec455 ] . in a first attempt we consider a single creation operator of molecule type , which models a bound @xmath48 state , @xmath49 diagrams with closed fermion loops
are ignored . in figure [ f005 ]
we compare the corresponding effective mass ( blue points ) with two times the @xmath44 effective mass ( red points ) .
we do not observe any indication of a four quark @xmath43 state , which is lighter than @xmath50 . on the other hand
the effective mass corresponding to the creation operator ( [ eqn650 ] ) still seems to slightly decrease at large temporal separations and does not exhibit a clear plateau .
this signals a trial state @xmath51 , which has a poor ground state overlap .
the ground state could either be two unbound @xmath44 ( then the plateau is expected at @xmath50 ) or still a @xmath43 bound state , however , with a structure quite different from ( [ eqn650 ] ) .
effective mass as functions of the temporal separation for the @xmath48 molecule operator ( blue points ) and the @xmath44 meson multiplied by a factor @xmath52 ( red points).,width=264 ] currently we explore the latter alternative . while the creation operator ( [ eqn650 ] ) places the two @xmath44 mesons essentially on top of each other , they could be rather far separated , e.g. by a distance up to @xmath53 ( for a recent study of this distance for heavy - heavy - light - light tetraquarks cf .
@xcite ) . to this end
we employ corresponding more general molecule type @xmath43 creation operators .
computations are in progress .
we acknowledge helpful discussion with christian fischer .
m.w . acknowledges support by the emmy noether programme of the dfg ( german research foundation ) , grant wa 3000/1 - 1 .
m.g . acknowledges support by the marie - curie european training network itn strongnet grant pitn - ga-2009 - 238353 .
is funded by the irish research council and is grateful for the hospitality at the university of cyprus and cyprus institute , where part of this work was carried out .
. acknowledges support by strongnet . this work was supported in part by the helmholtz international center for fair within the framework of the loewe program launched by the state of hesse and by the dfg and the nsfc through funds provided to the sino - german crc 110 `` symmetries and the emergence of structure in qcd '' .
99 j. o. daldrop _ et al . _
[ etm collaboration ] , pos lattice * 2012 * , 161 ( 2012 ) [ arxiv:1211.5002 [ hep - lat ] ] . c. alexandrou _ et al . _
[ etm collaboration ] , jhep * 1304 * , 137 ( 2013 ) [ arxiv:1212.1418 ] .
m. wagner _
[ etm collaboration ] , pos confinementx * * , 108 ( 2012 ) [ arxiv:1212.1648 [ hep - lat ] ] . m. wagner _
[ etm collaboration ] , arxiv:1302.3389 [ hep - lat ] .
m. wagner _
_ , arxiv:1309.0850 [ hep - lat ] .
w. heupel , g. eichmann and c. s. fischer , phys .
b * 718 * , 545 ( 2012 ) [ arxiv:1206.5129 [ hep - ph ] ] .
r. baron _ et al .
_ [ etm collaboration ] , jhep * 1006 * , 111 ( 2010 ) [ arxiv:1004.5284 [ hep - lat ] ] .
r. baron _ et al .
_ [ etm collaboration ] , comput .
. commun . * 182 * , 299 ( 2011 ) [ arxiv:1005.2042 [ hep - lat ] ] . j. weber _ et al .
_ , arxiv:1310.1760 [ hep - lat ] .
c. alexandrou , p. .de forcrand and b. lucini , phys .
lett . * 97 * , 222002 ( 2006 ) [ hep - lat/0609004 ] .
m. wagner and c. wiese [ etm collaboration ] , jhep * 1107 * , 016 ( 2011 ) [ arxiv:1104.4921 [ hep - lat ] ] .
s. aoki _ et al . _
[ pacs - cs collaboration ] , phys .
d * 79 * , 034503 ( 2009 ) [ arxiv:0807.1661 [ hep - lat ] ] .
m. lscher , commun .
phys . * 105 * , 153 ( 1986 ) .
m. lscher , nucl .
b * 354 * , 531 ( 1991 ) .
m. lscher , nucl .
b * 364 * , 237 ( 1991 ) . c. b. lang _ et al .
_ , phys . rev .
d * 86 * , 054508 ( 2012 ) [ arxiv:1207.3204 [ hep - lat ] ] .
d. mohler , s. prelovsek and r. m. woloshyn , phys .
d * 87 * , no . 3 , 034501 ( 2013 ) [ arxiv:1208.4059 [ hep - lat ] ] .
v. bernard _ et al .
_ , jhep * 1101 * , 019 ( 2011 ) [ arxiv:1010.6018 [ hep - lat ] ]
. m. doring _ et al .
_ , eur . phys .
j. a * 48 * , 114 ( 2012 ) [ arxiv:1205.4838 [ hep - lat ] ] . m. doring , plenary talk at `` 31st international symposium on lattice field theory '' ( 2013 ) .
|
we review the status of an ongoing long - term lattice investigation of the spectrum and structure of tetraquark candidates .
we focus on the light scalar meson @xmath0 .
first steps regarding the study of a possibly existing @xmath1 tetraquark are also outlined .
|
dwarf novae ( dne ) are a subclass of cataclysmic variables ( cvs ) , comprised of a low - mass , main sequence secondary star and a white dwarf ( wd ) primary .
they are characterized by their quasi - periodic outburst episodes that are typically 2 - 6 mag in amplitude .
the roche lobe - filling secondary loses gas through the inner langrangian point carrying substantial angular momentum that leads to the formation of a disk around the accreting white dwarf . a thermal - viscous instability known as
the disk instability model ( osaki 2005 and references therein ) causes the accretion disk to transition from a cool , quiescent , optically thin state to a stable state where the disk is much hotter , more luminous , optically thick and approaching a steady state .
the disk is heated due to the release of gravitational potential energy as the accreted material spirals through the optically thick disk toward the wd surface .
the energy released can heat the disk to temperatures on the order of 100,000 k. this heating of the disk accounts for the increase in luminosity during a dwarf nova ( dn ) outburst , and should dominate the system s luminosity in outburst .
when the outburst concludes , the disk should , theoretically , be relatively empty of material in its hotter inner regions and optically thin . at this time , the disk luminosity should be a negligible contributor to the system s luminosity .
this period of low mass - transfer is known as quiescence . in dwarf novae
, it has long been noted that the systems generally undergo a transition from emission line - dominated spectra in the optical and uv ( in quiescence ) to absorption line spectra as the continuum gradually increases in strength up to outburst maximum with the development of broad absorption troughs and in the uv , p cygni profile structure in the resonance lines . in the optical ,
emission line cores remain but have narrower widths than their counterparts in quiescence ( warner 1995 and refererences therein ) .
the spectra in outburst and quiescence exhibit a mix of h and he lines as well as metals , unlike the dichotomy between the da ( h - rich ) and db ( he - rich ) composition sequences of the isolated white dwarfs . two types of dwarf nova outbursts have been recognized , type a , which have a fast optical rise , delay progessively at shorter wavelengths and describe a wide loop in the two color diagram as the system progresses from quiescence to outburst , then back to quiescence , and type b , in which the optical rise is slower and the cycle describes a narrow loop in the two color diagram with only a small or zero delay between the uv and optical ( smak 1984a , b ) on the decline from outburst , systems of both types show fluxes that fall simultaneously at all wavelengths but with the uv declining fastest near the end of the decline .
pringle & verbunt ( 1984 ) and verbunt ( 1987 ) found that the uv flux distributions evolve similarly for a given kind of outburst .
early attempts to theoretically interpret the spectral energy distributions of dwarf novae in outburst relied on comparisons of the the ratios of continuum fluxes in emission - line free regions with model predictions .
the uv flux distributions of dne in outburst have been mostly fitted with @xmath0 flux distributions where @xmath1 ( verbunt 1987 ) .
moderately good agreement was achieved for systems expected to approximate steady state disks . the largest uncertainty in the disk model fitting is the sensitivity to the wd mass ( verbunt 1987 ) .
the outburst spectra have overall flux distributions that resemble in both the optical and uv , the flux distribution of a b2 - 3 v - iii star ( t@xmath2k ) .
la dous ( 1991 ) noted in her archival study of 32 dwarf novae in outburst that the uv continuum flux distributions were virtually identical regardless of inclination angle , except for the systems oy car , z cha , ab dra , ww cet , cm del , rx and , and vw vul ( with increasingly bluer distributions but all redder than the bulk of objects ) and uz ser that is much bluer than the other objects .
la dous ( 1991 ) used the published flux plots from an earlier generation of accretion disk models to conclude that the inner disk radius , inclination angle and mass transfer rate at a fixed wd mass of 1 @xmath3 are the most critical factors in modeling the observed energy distribution .
in particular , the appearence of the uv spectra at outburst is strongly correlated with the orbital inclination . in lower inclination systems
the outburst spectra appear in pure absorption while for higher inclinations ( @xmath4 degrees ) the spectra are continuous and finally for @xmath5 degrees , strong emission lines are present ( la dous 1991 ) .
accretion rates have been obtained from absolute magnitudes @xmath6 ( requiring distances ) for various subgroups of cvs with due account for the sensitivity to system parameters such as the inclination angle ( smak , 1989 , 1994 ) .
generally , during dwarf nova maximum , outburst accretion rates have been estimated in the range @xmath7 to @xmath8 while during the rise to outburst among dwarf novae , accretion rates in the range @xmath9 to @xmath10 have been derived ( warner 1995 ) .
however , relatively few individual accretion rates have been published for dwarf novae in outburst based upon detailed accretion disk model fitting .
there is a comprehensive atlas of optical spectra of dwarf novae in outburst ( morales - rueda and marsh 2002 ) but none for uv spectra in which the data has been modeled and parameters derived .
some examples of early derivations of accretion rates in outburst from disk model fitting are for vw hyi , z cha , oy car and ss cygni . for vw hyi
, an accretion rate @xmath11 was obtained by polidan , mauche and wade ( 1990 ) , and polidan and holberg(1984 ) . for z cha ,
horne and cook ( 1995 ) obtained @xmath12 and for oy car , rutten et al .
( 1992 ) obtained @xmath13 .
for ss cygni at maximum , kiplinger ( 1979 ) found excellent agreement between a model accretion disk with @xmath14 and the observed flux distribution from the ir through the fuv regions .
if @xmath15 is not constant throughout the disk , then steady state disk model fitting in outburst will yield discrepancies due to a non - steady state radial temperature distribution , t(r ) , in the disk .
of course , there is a strong sensitivity to the white dwarf mass which is poorly known for the majority of systems .
orbital inclinations are likewise uncertain and distances remain for the most part rough estimates except for less than two dozen systems that have reliable trigonometric parallaxes .
thus , even for a given inclination angle and distance , the accretion disk fits to be described in section 3 yield inferred white dwarf masses uncertain by at least three - tenths of a solar mass .
while our spectral coverage is restricted and excludes the regions where the boundary layer and the white dwarf probably contribute to the fuv flux , the international ultraviolet explorer ( iue ) spectra do cover the range of wavelengths where accretion disks during outburst are emitting a large fraction of their light . in view of all these caveats
, our disk model fits should be regarded only as first approximations to the real accretion rates during outburst . moreover
, these spectra nicely complement the compilation of optical spectra of dwarf novae in outburst described by morales - rueda and marsh ( 2002 ) .
all the spectral data were obtained from the multimission archive at space telescope ( mast ) iue archive are in a high activity state , very near or at outburst .
we restricted our selection to those systems with swp spectra , with resolution of 5 and a spectral range of 1170 to 2000 .
all spectra were taken through the large aperture at low dispersion .
the massa & fitzpatrick ( 2000 ) flux calibration - correction algorithm was applied to all the iue data used ; please see their paper for a description of the corrections it makes to the data . when more than one spectrum with adequate signal - to - noise ratio was available , the two best spectra were analyzed .
an observing log of the observations is given in table 1 , where we list : ( 1 ) the system name , ( 2 ) data i d , ( 3 ) date of observation , ( 4 ) universal time of observation , ( 5 ) exposure time , and ( 6 ) activity state .
transition refers to an intermediate state between outburst and quiescence , and an asterix in the last column indicates that an outburst state was determined without a ground - based optical light curve at the time of the iue observation .
the activity state of the spectra was determined by examining the aavso light curves for each system as well as the flux level of the iue spectrum that typically made it obvious the spectrum was obtained in outburst .
unfortunately , we were unable to cross - reference this aavso data with data from vsnet during the preparation of this paper due to their archives being off - line . in the case of those systems
not covered by the aavso , their activity state was assessed based upon either mean photometric magnitudes taken from the ritter & kolb(2003 ) catalogue or from iue fine error sensor ( fes ) measurements at the time of the iue observation
. the fes counts , when available , were converted to optical magnitudes to help ascertain the brightness state at the time of the iue observation , and used as consistency checks on our model fitting since the model - predicted optical magnitude from a best - fitting model should always be fainter than observed optical magnitude of the system .
this is becuase of the contribution in the optical of the secondary .
in addition , the presence of p - cygni profiles , absorption lines , and comparison with spectral data and flux levels for the systems during other activity states was used to ascertain the state of the system .
the reddening of the systems was determined based upon all estimates listed in the literature .
the three principal sources of reddening were the compilations of la dous ( 1991 ) , verbunt ( 1987 ) and bruch & engel ( 1999 ) .
if there was a range of values found , then the value we assume in our analyses follows the range shown in column ( 7 ) of table 2 , and the ranges and assumed values are separated by a semicolon . if only one value was listed in the literature , then we adopted that value , and if there was none listed , we took the reddening to be zero .
the spectra were then de - reddened with the iuerdaf routine unred .
as a constraint on the goodness of a particular model fit , we adopted a distance from either a trigonometric parallax , the warner ( 1995 ) and harrison et al .
( 2004 ) relations or from the scale factor of the best - fitting wd model to hst spectra of dwarf novae during quiescence .
the adopted distance was used as a basis for comparison with distances yielded by the accretion disk model fitting .
the most direct means of knowing the distance is to have trigonometric parallaxes such as those measured by thorstensen ( 2003 ) for 14 dwarf novae . but dwarf novae and nova - like variables also offer the advantage of estimating distances from correlations between their absolute magnitudes at maximum brightness and their orbital periods .
distance estimates were obtained from either the absolute magnitude at outburst @xmath16 versus orbital period @xmath17 relation of warner ( 1995 ) or from a more recent relationship by harrison et al .
( 2004 ) based upon their recent hst fgs parallaxes of dwarf novae .
the warner ( 1995 ) relation is @xmath18 and the harrison et al.(2004 ) relation is @xmath19 at the outset of the modeling process , we applied both of these relationships to each system .
then , we conducted an exhaustive search of the literature for previous distance estimates .
if the literature search revealed other estimates , then we adopted some reasonable mean based upon the different methods used to obtain each distance estimate and the distance computed from the two calibrated @xmath20 relations .
if no other distance estimates existed , then we simply adopted a reasonable mean of the two @xmath20 relations .
if a trigonometric parallax was available , then we adopted it for the distance . in the case of two systems ( u gem and yz cnc ) ,
there was more than one parallax .
for u gem , all were identical .
for yz cnc , they were slightly different , so we took a mean of the parallax values . in the case of su uma ,
the parallax obtained was unreliable , so we combined it in a mean with other distances estimates and the @xmath20 relations .
the adopted distances used as constraints in the synthetic spectral fitting procedure are shown in table 2 .
we adopted model accretion disks from the optically thick disk model grid of wade & hubeny ( 1998 ) . in these accretion disk models ,
the innermost disk radius , r@xmath21 , is fixed at a fractional white dwarf radius of @xmath22 .
the outermost disk radius , r@xmath23 , was chosen so that t@xmath24 is near 10,000k since disk annuli beyond this point , which are cooler zones with larger radii , would provide only a very small contribution to the mid and far uv disk flux , particularly the swp fuv bandpass .
the mass transfer rate is assumed to be the same for all radii .
thus , the run of disk temperature with radius is taken to be : @xmath25 where @xmath26 and @xmath27 limb darkening of the disk is fully taken into account in the manner described by diaz et al .
( 1996 ) involving the eddington - barbier relation , the increase of kinetic temperature with depth in the disk , and the wavelength and temperature dependence of the planck function .
the boundary layer contribution to the model flux is not included .
however , the boundary layer is expected to contribute primarily in the extreme ultraviolet below the lyman limit . after masking emission lines , artifacts and poor quality data in the spectra
, we determined separately for each system , the best - fitting accretion disk model using iuefit , a @xmath28 minimization routine . in virtually every dwarf nova system in our iue sample , the accretion disk by itself provided very good agreement with the outburst spectra .
however , if the accretion disk fit to the outburst spectrum was not entirely satisfactory over the entire wavelength range despite the disk providing nearly 100% of the flux , then an attempt was made to slightly improve the fit by adding the flux contribution of a hot white dwarf .
for this minority of cases , wd model spectra with solar abundances were created for high gravity stellar atmospheres using tlusty ( hubeny 1988 ) and synspec ( hubeny & lanz 1995 ) .
we took a range of gravities in the fitting from log @xmath29 in steps of 0.5 .
for the white dwarf radii , we use the mass - radius relation from the evolutionary model grid of matt wood ( private communication ) .
these models have realistic non - degenerate h - rich envelopes , use the lamb and van horn equation of state and include the effect of non - zero temperature on the wd radius .
we took the best - fitting accretion disk model and combined it with a hot wd model , using a @xmath28 minimization routine called diskfit .
the best - fitting composite accretion disk plus white dwarf was determined based upon the minimum @xmath28 value achieved and consistency of the scale factor - derived distance with the adopted distance for each system . the scale factor , @xmath30 , normalized to a kiloparsec and solar radius , can be related to the white dwarf radius r through : @xmath31 , where @xmath32 , and @xmath33 is the distance to the source .
an illustration of the accuracy we can obtain in a formal error analysis with confidence contours are given in winter & sion ( 2003 ) .
in that paper we presented accretion rates of em cygni , cz ori and ww ceti in quiescence , in which @xmath34 and @xmath35 confidence contours are given . in table 2
, we list the final composite fitting results for all systems , by column , as follows : ( 1 ) a reference number , ( 2 ) system name , ( 3 ) system type ( and sub - type , if applicable ) , ( 4 ) @xmath36 , ( 5 ) e(b - v ) , ( 6 ) adopted distance , ( 7 ) inclination , ( 8) mass accretion rate @xmath15 , ( 9 ) percent flux contribution of the wd , and ( 10 ) percent flux contribution of the disk . the best fitting accretion disk model for each system
is shown in the accompanying figures 1 - 8 , listed in order of increasing orbital period as in table 2 .
the thick solid line is the best fitting accretion disk .
the accretion disk is the overwhelmingly dominant contributor in the fuv .
if in rare cases a hot white dwarf flux contribution was added in an attempt to slightly improve the fits , then the thick solid line represents their combination , the dashed curve represents the accretion disk alone , and the dotted curve represents the white dwarf . also included are a reference number and name given in table 2 to clearly identify each system . to discuss a particular system , the figures are labelled figure n.nn , where n is the figure number and nn is the reference number of a particular dwarf nova .
there are several distances listed in column ( 6 ) of table 2 that are followed by either a @xmath37 , or a @xmath37 in parentheses .
if the distance is followed by a @xmath37 , that distance is taken from a parallax ( or parallaxes in the case of u gem ) .
if the distance is followed by a @xmath37 in parentheses , then it is an adopted distance that incorporates a parallax ( see the explanation in section 3.1 ) .
all parallaxes are taken from either harrison et al .
( 2004 ) , or thorstensen ( 2003 ) .
a colon in column 7 denotes an uncertain result .
although the eclipsing dwarf novae oy car and z cha are part of the iue archival sample , these systems are not included in the statistics of this study and their derived accretion rates can not be taken seriously .
this is because the accretion disk in these systems is self - eclipsed ( i.e. the outer disk obscures the inner disk ) .
hence , any derived parameters ( for example , the accretion rate , percentage of disk flux contribution ) are systematically affected . in table 2
it is seen that virtually all of the accretion rates derived from the fuv flux distribution with steady state , optically thick disk models fall within the range @xmath38 to @xmath39 for the distances , white dwarf masses , orbital inclinations , and interstellar reddening values we have adopted .
this is no great surprise although the range of accretion rates that we derive is somewhat broader than the range derived by warner ( 1995 ) whose accretion rates were based upon the absolute magnitudes of the accretion disk corrected for inclination and limb darkening effects .
it should be acknowledged however that some of the distances for the 44 dwarf novae in our study involved the use of the warner ( 1995 ) relation .
hence , the rough agreement with the range of accretion rates found by warner ( 1995 ) from his optical study is biased to some degree in favor of such an agreement
. moreover , as is well known , in all of the outburst spectra , almost without exception , ( 1238 , 1242 ) , ( 1548 , 1550 ) and ( 1393 , 1402 ) , the hallmark resonance doublets are seen in prominent blue - shifted absorption and originate in the cv wind . in many of the hotter disk models , one sees synthetic si iv absorption but the disk temperatures are not high enough for significant non - wind , or .
the strongest metallic absorption seen in the spectra , aside from the wind resonance doublets , is due to a blend of multiplet 2 of ( 1300 ) , and ( 1302 ) . in virtually all of the model accretion disk fits ,
the observed absorption centered at 1300 is reasonably well represented by the synthetic disk absorption at the solar abundance of silicon .
a difficulty of fitting the width of the lyman @xmath40 profiles in most of the outburst spectra is the presence of which complicates discerning the longward wing of lyman @xmath40 .
fortunately , the shortward wing is usually quite apparent and one can assume , as a first approximation , mirror symmetry between the two wings . nonetheless , in some cases
the looks suspiciously like a pseudo - absorption feature created by the geocoronal emission reversal seen in the core of lyman @xmath40 .
several individual spectra are noteworthy for reasons other than just the goodness of their fit with steady state , optically thick accretion disk energy distributions . in figure 1.02 , the outburst spectrum of the su uma system sw uma along with the best - fitting accretion disk model .
there are profiles ( e.g. 1550 , 1260 ) which appear to have emission flanking absorption on both the shortward and longward sides of the absorption .
similar profile shapes were seen during the early stages of the july 2001 superoutburst of wz sge ( sion et al.2003 ) .
the emission seen near could well be due to the ( 1526 , 1533 ) doublet , which would match the same structure seen at ( 1260 ) .
if this is the case , then is barely seen and may be very weak relative to nitrogen . in figures 1.04 , 3.16 , 4.18 , 6.33 , 7.35 , 7.39 , 7.40 , 8.41 , 8.42 , 8.43 , and 8.44 ,
the outburst spectra of t leo , wx hyi , yz cnc , rx and , ah her , z cam , em cyg , ru peg , sy cnc , dx and , and bv cen , respectively , reveal a deviation from a steady state flux distribution longward of 1700 .
in every case , the disk model flux is lower than the observed flux .
the deviation tends to be more pronounced among z cam - type systems .
a anonymous referee suggests this may be due to the secondary star or possibly a shortfall of flux due to the temperature of the outermost disk boundary being set at 10,000k . in our sample , except for t leo
, this deviation is seen primarily in systems above the period gap .
as stated earlier , one difficulty of fitting the width of the lyman @xmath40 profiles in this and other outburst spectra is the presence of ( 1238 , 1242 ) which is some cases may actually be a pseudo - feature created by the geocoronal emission reversal seen in the core of lyman @xmath40 . in figure 2.09 ,
tv crv has an unusually strong ( 1393 , 1402 ) absorption suggesting possible suprasolar si abundance or a higher ionization fraction in the outflowing wind material .
our analyses represent the first detailed look at all of the dwarf novae observed with the iue swp in outburst in a systematic way . we have applied a uniform analysis to the largest spectroscopic sample of fuv spectra of dwarf novae , obtained during the same brightness state with the same telescope , and the same instrumental setup , taking into the account the best available distance information .
accordingly , this large body of accretion rates will serve as a basis of comparison with accretion rates derived by other independent methods in other wavelength regions .
moreover , the fuv - derived accretion rates of individual systems can be compared with fuv - derived accretion rates of other systems in the same dwarf nova subclass or across subclasses .
a number of implications and conclusions arise from this analysis .
overall , the steady state , optically thick disk model fits to the outburst spectra shown in figures 1 through 8 are quite acceptable and in some cases excellent .
all of the accretion rates derived from the fuv fall within the range @xmath38 to @xmath39 for the distances , white dwarf masses , orbital inclinations , and reddening values we have adopted .
therefore , the main uncertainties in the derived accretion rates , or in constraining the outburst accretion rates , arise from the uncertainties in the published values of the white dwarf masses , the orbital inclinations and the reliability of the distances .
on the other hand , there are several systems with outburst spectra in the iue archive that merit attention because they differ significantly from the fuv energy distribution of steady state accretion disks . aside from the eclipsing systems z cha and oy car , which reside below the period gap and
are seen nearly edge - on with absorbing curtains , all of the systems that depart from a steady state disk distribution do so longward of 1600 and all of these systems lie above the period gap except for t leo .
the u gem - type and z cam - type systems are wx hyi , yz cnc , rx and , ah her , z cam , em cyg , ru peg , sy cnc , dx and , and bv cen .
if one takes our fuv - derived accretion rates in outburst at face value , then in principle they can be used in combination with accretion rates derived for dwarf nova quiescence , and the timescales of outburst and quiescence to estimate the time - averaged accretion rate of the system during its lifetime as a dwarf nova .
that is , when average accretion rates are known during quiescence and during outburst , then a true time - averaged accretion rate or mean accretion rate , @xmath41 , may be derived .
if @xmath42 is an average accretion rate in outburst , @xmath43 is an average accretion rate in quiescence , @xmath44 is the average duration of outburst and @xmath45 is the average duration of quiescence , then for a dwarf nova duty cycle @xmath46 the `` duty - cycle '' averaged @xmath47 is given by @xmath48 t_{tot}^{-1}\ ] ] now it appears that all dwarf novae both above and below the gap have @xmath49 to @xmath50 in outburst .
suppose we assume the outburst times and quiescence times as well as the wd masses of all dwarf novae are the same .
then the only possible causes for a difference in accretion heating of a wd in a dn above the gap compared with a wd below the gap are : ( 1 ) a difference in wd ages due to evolution age ; ( 2 ) a different amount of disk accretion during the outburst ( i.e. more mass from the disk accreted above the gap versus the mass fraction of the disk accreted below the gap ) or ; ( 3 ) a significantly higher accretion rate during quiescence for dn above the gap than below the gap .
looking at the above equation , if @xmath43 is sufficiently high and the quiescence long enough ( as in many long period dwarf novae ) , then the second term in equation ( 4 ) can compete and possibly overwhelm the first term if the outburst duration is short enough , if @xmath42 is lower , say , less than @xmath50 , or if the accretion rate during quiescence is higher than a nominal @xmath51 that we assume in this paper .
while the above time - averaged accretion rate are gotten directly from model - derived accretion rates , the time - averaged accretion rates derived for selected dwarf novae by patterson ( 1984 ) were obtained for the most part from a mean optical brightness level in aavso light curves on the assumption that the accretion disk is the sole source of the system light . by converting the mean magnitude to a flux and combining it with an estimated distance ,
a luminosity was obtained .
the accretion rate providing that luminosity was derived for a given assumed or known white dwarf mass .
it is of interest to carry out a quantitative comparison of the two approaches to estimating @xmath41 . for illustration
, we take sample values for the outburst timescale and the recurrence timescale of a typical dwarf nova .
we assume for simplicity that the recurrence time of dwarf nova outbursts is constant at , say 21 days , and the outburst timescale is constant at 7 days . for these parameters ,
if the outburst accretion rate is @xmath52 then , by ( 1 ) , @xmath53 .
if the outburst accretion rate is @xmath54 and the quiescent rate is an assumed to be @xmath51 , then @xmath55 .
suppose we take average values of @xmath44 and @xmath56 for two well - studied systems such as u gem and ss cygni . for u gem , from table 2
, we take @xmath42 to be @xmath57 , @xmath44 of 14 days ( wide outburst ) , @xmath56 of 100 days and @xmath43 of @xmath51 . hence for these parameters u gem would have @xmath58 .
considering ss cygni with the same quiescence accretion rate , @xmath44 of 14 days , @xmath56 of 40 days and @xmath42 of @xmath10 from table 2 , we find that @xmath59 .
patterson ( 1984 ) estimated @xmath60 for u gem and @xmath61 for ss cygni .
we note that the distances adopted by patterson ( 1984 ) , 78 pc and 95 pc for u gem and ss cygni respectively , are closer than the fgs parallax distances we used for the two systems .
comparing patterson s values with the time - averaged accretion rates estimated using table 2 , our value of @xmath41 for u gem is 10 times higher than estimated in patterson ( 1984 ) while for ss cygni , our @xmath41 is 1.4 times higher than the value estimated by patterson ( 1984 ) .
until the accretion rates during dwarf nova quiescence are more accurately known , it is premature to compute the time - averaged accretion rates in this direct manner .
moreover , individual accretion rates will not be secure for dwarf novae until there is further progress in determining better white dwarf masses , orbital inclinations and distances .
we are grateful to an anonymous referee for helpful comments .
this work was supported by nsf grant ast05 - 07514 , nasa grant nng04ge78 g and by summer undergraduate research support from the delaware space grant consortium .
some or all of the data presented in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
support for mast for non - hst data is provided by the nasa office of space science via grant nag5 - 7584 and by other grants and contracts .
wz sge & swp03507 & 1978 - 12 - 01 & 19:21:00 & 24 & peak outburst
+ sw uma & swp27871 & 1986 - 03 - 08 & 04:32:00 & 210 & outburst + wx cet & swp36511 & 1989 - 06 - 15 & 10:41:00 & 3300 & late decline + t leo & swp33642 & 1988 - 05 - 25 & 20:53:00 & 300 & outburst + v1159 ori & swp56889 & 1996 - 03 - 03 & 17:01:00 & 5400 & rise to outburst + v436 cen & swp54246 & 1995 - 03 - 28 & 10:09:00 & 2400 & outburst @xmath62 + bc uma & swp50668 & 1994 - 05 - 01 & 21:09:00 & 3000 & peak outburst + ek tra & swp09705 & 1980 - 08 - 05 & 22:20:00 & 1140 & outburst @xmath62 + tv crv & swp41842 & 1991 - 06 - 14 & 14:25:00 & 1800 & outburst @xmath62 + oy car & swp25838 & 1985 - 05 - 02 & 17:34:00 & 720 & rise to outburst + vy aqr & swp21720 & 1983 - 12 - 08 & 03:22:00 & 600 & outburst + er uma & swp40947 & 1991 - 02 - 27 & 22:20:00 & 4500 & outburst @xmath62 + ir gem & swp38524 & 1990 - 04 - 05 & 02:03:00 & 1068 & outburst + ay lyr & swp09342 & 1980 - 06 - 22 & 06:46:00 & 3600 & late outburst + vz pyx & swp44129 & 1992 - 03 - 07 & 05:36:00 & 2400 & outburst @xmath62 + vw hyi & swp40045 & 1990 - 11 - 04 & 10:45:00 & 150 & peak outburst + z cha & swp30670 & 1987 - 03 - 31 & 17:03:00 & 600 & peak outburst + wx hyi & swp23952 & 1984 - 09 - 13 & 19:22:00 & 3600 & outburst @xmath62 + su uma & swp35261 & 1989 - 01 - 05 & 08:58:00 & 1800 & peak outburst + yz cnc & swp15560 & 1981 - 11 - 24 & 10:07:00 & 540 & outburst + tu men & swp10665 & 1980 - 11 - 23 & 17:58:00 & 1800 & outburst @xmath62 + ab dra & swp07413 & 1979 - 12 - 15 & 18:51:00 & 4500 & rise to outburst + cy lyr & swp21030 & 1983 - 09 - 12 & 22:09:00 & 3600 & outburst + kt per & swp07382 & 1979 - 12 - 13 & 01:19:00 & 2700 & early decline + ar and & swp15445 & 1981 - 11 - 07 & 12:34:00 & 2100 & outburst + cn ori & swp15950 & 1982 - 01 - 04 & 08:36:00 & 2400 & outburst + x leo & swp15985 & 1982 - 01 - 06 & 14:03:00 & 6240 & outburst + vw vul & swp18875 & 1982 - 12 - 23 & 19:08:00 & 7200 & transition @xmath62 + uz ser & swp15078 & 1981 - 09 - 22 & 18:41:00 & 2400 & outburst + u gem & swp10327 & 1980 - 10 - 10 & 13:28:00 & 353 & peak outburst + tw vir & swp10951 & 1981 - 01 - 03 & 06:47:00 & 3600 & outburst @xmath62 + ss aur & swp18610 & 1982 - 11 - 20 & 17:33:00 & 1560 & rise to outburst + hx peg & swp37459 & 1989 - 10 - 25 & 19:31:00 & 3600 & outburst @xmath62 + rx and & swp17678 & 1982 - 08 - 13 & 22:59:00 & 600 & peak outburst + cz ori & swp16039 & 1982 - 01 - 14 & 03:18:00 & 2700 & outburst + ah her & swp17662 & 1982 - 08 - 11 & 22:12:00 & 1200 & outburst + tz per & swp17643 & 1982 - 08 - 09 & 19:37:00 & 2580 & outburst + tt crt & swp47805 & 1993 - 06 - 04 & 19:54:00 & 3300 & outburst + ss cyg & swp47746 & 1993 - 05 - 27 & 07:57:00 & 180 & outburst + z cam & swp21721 & 1983 - 12 - 08 & 05:41:00 & 240 & early decline + em cyg & swp07297 & 1979 - 12 - 02 & 22:45:00 & 1500 & outburst + ru peg & swp15079 & 1981 - 09 - 22 & 20:43:00 & 840 & decline + sy cnc & swp07313 & 1979 - 12 - 05 & 00:40:00 & 1200 & mid decline + dx and & swp37687 & 1989 - 11 - 26 & 15:25:00 & 2700 & rise to outburst + ww cet & swp10664 & 1980 - 11 - 23 & 16:21:00 & 1800 & transition @xmath62 + bv cen & swp24867 & 1985 - 01 - 08 & 20:48:00 & 3600 & transition @xmath62 + 01 & wz sge & su - wz & 0.056687846 & 0.00 & 43.3@xmath37 & 78&@xmath63 & @xmath64&@xmath6599.5 + 02 & sw uma & su(dq ? ) & 0.056815 & 0.00 & 110 & 60&@xmath66 & 0 & 100 + 03 & wx cet & su - wz & 0.05827 & 0.00 ? & 187 & 41&@xmath67 & 0 & 100 + 04 & t leo & su & 0.05882 & 0.00 & 101@xmath37 & 60&@xmath68 & 0 & 100 + 05 & v1159 ori&su - er & 0.06217801 & 0.00 ? & 213 & 60&@xmath67 : & 0 & 100 + 06 & v436 cen & su & 0.062501 & 0.05 ; 0.00 & 320 & 41&@xmath69 & 1 & 99 + 07 & bc uma & su - wz & 0.06261 & 0.00 ? & 285 & 75&@xmath66 & 0 & 100 + 08 & ek tra & su & 0.06288 & 0.05;0.00 & 200 & 18&@xmath70 : & @xmath711 & @xmath6599 + 09 & tv crv & su & 0.0629 & 0.00 & 268 & 60&@xmath67 & 0 & 100 + 10 & oy car & su & 0.062917 & 0.00 & 85 & 75&@xmath72 & @xmath7111 & @xmath6589 + 11 & vy aqr & su - wz & 0.06309 & 0.00 & 97@xmath37 & 41&@xmath67 & 0 & 100 + 12 & er uma & su - er & 0.06366 & 0.00 ? & 460 & 60&@xmath73 : & @xmath711 & @xmath6599 + 13 & ir gem & su & 0.0684 & 0.00 & 250 & 60&@xmath68 & 0 & 100 + 14 & ay lyr & su & 0.0737 & 0.00 & 380 & 41&@xmath66 & 0 & 100 + 15 & vz pyx & su ( ip ? ) & 0.07332 & 0.00 & 280 & 60&@xmath74 & 5 & 95 + 16 & vw hyi & su & 0.074271 & 0.00 & 65 & 60&@xmath66 & 0 & 100 + 17 & z cha & su & 0.074499 & 0.00 & 97 & 81&@xmath75 : & @xmath7119 & @xmath6581 + 18 & wx hyi & su & 0.0748125 & 0.00 & 280 & 60&@xmath67 & 1 & 99 + 19 & su uma & su & 0.07635 & 0.00 & 270@xmath37 & 41&@xmath76 & 2 & 98 + 20 & yz cnc & su & 0.0868 & 0.00 & 265@xmath37 & 41&@xmath77 & @xmath712 & @xmath6598 + 21 & tu men & su & 0.1172 & 0.08 & 320 & 41&@xmath68 & 0 & 100 + 22 & ab dra & zc & 0.15198 & 0.10 & 400 & 81&@xmath67 & 8 & 92 + 23 & cy lyr & ug & 0.1591 & 0.18 & 400 & 60&@xmath68 & 0 & 100 + 24 & kt per & zc & 0.16265777 & 0.15 - 0.54 ; 0.2 & 245 & 60&@xmath69 & @xmath710.5&@xmath6599.5 + 25 & ar and & ug ( su ? ) & 0.16302 & 0.02 & 270 & 60&@xmath68 : & 0 & 100 + 26 & cn ori & ug & 0.163199 & 0.00 & 295 & 60&@xmath73 & 2 & 98 + 27 & x leo & ug & 0.1646 & 0.00 & 350 & 41&@xmath78 & 12 & 88 + 28 & vw vul & zc & 0.16870 & 0.15 & 650 & 41&@xmath68 & 19 & 81 + 29 & uz ser & ug ( zc ? ) & 0.173 & 0.35 & 300 & 18&@xmath79 & 1 & 99 + 30 & ww cet & ug ( zc ? ) & 0.1758 & 0.00 & 190 & 60&@xmath67 : & @xmath711 & @xmath6599
+ 31 & u gem & ug & 0.17690619 & 0.00 - 0.05 ; 0 & 96@xmath37 & 75&@xmath80 & 1 & 99 + 32 & tw vir & ug & 0.18267 & 0.00 & 500 & 60&@xmath74 : & @xmath711 & @xmath6599 + 33 & ss aur & ug & 0.1828 & 0.10 & 201@xmath37 & 41&@xmath81 : & @xmath710.5&@xmath6599.5 + 34 & hx peg & & 0.2008 & 0.00 & 550@xmath82 & 41&@xmath74 : & @xmath83&@xmath6591.5 + 35 & rx and & zc & 0.209893 & 0 - 0.06 ; 0 & 200 & 41&@xmath84 & 1.5 & 98.5 + 36 & cz ori & ug & 0.214667 & 0.00 & 260 & 18&@xmath85 : & @xmath7114 & @xmath6586 + 37 & ah her & zc & 0.258116 & 0.3 & 660@xmath37 & 41&@xmath73 : & @xmath710.5&@xmath6599.5 + 38 & tz per & zc & 0.262906 & 0.27 & 275 & 41&@xmath72 & 5 & 95 + 39 & tt crt & ug & 0.2683522 & 0.00 & 500 & 60&@xmath86 : & @xmath713 & @xmath6597 + 40 & ss cyg & ug & 0.27513 & 0.04 ; 0 & 166@xmath37 & 18&@xmath66 & @xmath710.5&@xmath6599.5 + 41 & z cam & zc & 0.2898406 & 0.0 - 0.06 ; 0 & 112@xmath37 & 60&@xmath70 & 1 & 99 + 42 & em cyg & zc & 0.290909 & 0.05 ; 0 & 350 & 75&@xmath81 & @xmath710.5&@xmath6599.5 + 43 & ru peg & ug & 0.3746 & 0.00 & 282@xmath37 & 41&@xmath67 & 0 & 100 + 44 & sy cnc & zc & 0.38 & 0.00 & 450 & 18&@xmath68 & 0 & 100 + 45 & dx and & ug & 0.440502 & 0.00 & 500 & 60&@xmath68 & 7 & 93 + 46 & bv cen & ug & 0.610108 & 0.05 - 0.36 ; 0.1 & 500 & 60&@xmath66 & @xmath710.5&@xmath6599.5 +
|
we present a synthetic spectral analysis of nearly the entire far ultraviolet international ultraviolet explorer ( iue ) archive of spectra of dwarf novae in or near outburst .
the study includes 46 systems of all dwarf nova subtypes both above and below the period gap .
the spectra were uniformly analyzed using synthetic spectral codes for optically thick accretion disks and stellar photospheres along with the best - available distance measurements or estimates .
we present newly estimated accretion rates and discuss the implications of our study for disk accretion physics and cv evolution .
|
in the strong interaction , one of the most stringent constraints on possible violation of parity ( p ) and time - reversal ( t ) symmetry comes from the measurement of the electric dipole moment ( edm ) for neutron ( nedm ) and proton ( pedm ) @xmath13 .
the current upper bound is given by |d_n| < 6.310 ^ -26e for neutron@xcite , and |d_p| < 5.4 10 ^ -24e for proton@xcite , which are estimated from the results for edm of mercury atom @xmath14 given by @xmath15 @xcite on the other hand , qcd , which is regarded as the fundamental theory of the strong interaction , allows a gauge invariant , renormalizable cp odd operator in the lagrangian , called the @xmath0 term : & & id^4xg_(x)g_(x ) [ l_qcd ] , g_(x ) = _ g_(x ) in euclidean space - time with @xmath16 the field strength of gluon .
some model estimations@xcite yield |d_n| ~o(10 ^ -15~10 ^ -16 ) e , and this leads to a bound @xmath17 . hence @xmath0 must be small or
even must vanish in qcd .
a smallness of @xmath0 in the qcd sector , however , is not protected in the presence of the electroweak sector of the standard model , where the quark mass matrix , arising from yukawa couplings to the higgs field , may be written as |^r_i(x ) m_ij^l_j(x ) + |^l_i(x ) m^_ij^r_j(x ) , where @xmath18 and @xmath19 represent left and right handed quark fields with flavor indices @xmath20 .
diagonalizing the mass matrix and making it real , the parameter @xmath0 becomes = _ qcd + m , where @xmath21 is the original @xmath0 parameter in qcd .
therefore , either @xmath21 and @xmath22 are individually small , or the two contributions cancel out to the degree that the experimental upper bound on nedm is satisfied . in either of the two cases , it seems necessary to explain why nature chooses such a small value for @xmath0 ; this is the `` strong cp problem '' .
one of the most attractive explanations proposed so far is the peccei - quinn mechanism@xcite .
unfortunately , the axion , a new particle predicted by this mechanism , has not been experimentally observed yet .
the present model estimations of the hadronic contribution to nedm , @xmath23 , are maybe enough to convince us of the the smallness of @xmath0@xcite .
however , a theoretically reliable and accurate estimation for nedm will be required to determine the value of @xmath0 , if non - zero value of nedm is observed in future experiments .
lattice qcd calculations seem ideal for this requirement . indeed more than 15 years ago the first attempt was made to estimate nedm in a quenched lattice qcd simulation@xcite by calculating the spin dependent energy difference of the nucleon in the presence of the uniform static electric field . in this simulation , the @xmath0
was converted to the phase of the quark mass term by the chiral rotation , and a non - zero value of nedm was obtained for non - zero @xmath0 .
unfortunately this non - zero value of nedm turned out@xcite to be a lattice artifact due to the explicit chiral symmetry breaking of the wilson fermion action employed in the simulation ; if the @xmath0 term appears in the quark mass , an additional disconnected contribution must be included in the quenched calculation of ref.@xcite .
after this attempt , there has been no further lattice investigation on this problem for a long time .
recently , a new lattice strategy @xcite for the extraction of nedm has been proposed , and a lattice related investigation on nedm has been made@xcite . also , a preliminary lattice qcd result of nedm estimated from the nucleon electromagnetic three - point function has been reported at the lattice 2004 conference@xcite .
the value of nedm divided by @xmath0 is consistent with zero within statistical errors in this calculation , however . in all attempts , an expansion in @xmath0
have been used to avoid the complex action due to the @xmath0 term . since chiral symmetry seems important in the calculation of nedm@xcite , it may be preferable to use lattice fermion formulations having good chiral symmetry , such as the domain - wall fermion@xcite or the overlap fermion@xcite . in these formulations ,
the bosonic definitions of the topological charge agree well with the fermionic definitions .
in fact the domain - wall fermion was employed in ref.@xcite . in this paper
we propose a new method to calculate nedm in lattice qcd .
we consider the strategy of extracting nedm from the cp - odd part of the electromagnetic form factor of the nucleon , as in the recent lattice calculation @xcite as well as in some past model calculations @xcite .
we first examine the @xmath0 dependence of the three - point ( two nucleons and the electromagnetic current ) function by expanding up to first order in @xmath0 under the assumption that @xmath0 has fairly small value .
an important finding is that the parity - odd part of the three - point function contains a contribution coming from the parity - even form factors multiplied by the parity - odd part of the nucleon two - point function .
this contribution has to be subtracted from the three - point function in order to properly extract the cp - odd form factor relevant for nedm .
this point is not considered in ref.@xcite .
we present the formula to extract the relevant form factor in terms of the nucleon two- and three - point functions , which is valid not only for lattice simulations but also for general cases as long as chiral symmetry of quarks is well formulated in the calculations .
we test our formula in a lattice qcd simulation using the domain - wall fermion in the quenched approximation . a renormalization group ( rg ) improved gauge action
is employed based on our experience that the residual chiral symmetry breaking due to a finite fifth dimensional extent is smaller than in the case for the plaquette gauge action at a similar lattice spacing@xcite . since our primary purpose is to check the feasibility of the lattice qcd calculation of nedm based on our formula , we focus on accumulating the statistics to obtain a sufficient sampling of topological charge on a @xmath6 lattice at @xmath5 gev with the quark mass of @xmath24 , corresponding to @xmath25 .
our numerical results reveal that the contribution to the three - point function from the parity - even form factors multiplied by the parity - odd part of the nucleon two - point function is really significant .
we demonstrate the correctness of our formula by showing that two independent ways to extract the relevant form factor give consistent results only if the contribution from the parity - even form factors is properly subtracted . this paper is organized as follows . in sec .
[ sec : edm_formula ] we derive the formula to extract the relevant form factor in terms of the two- and three - point functions to first order in the expansion in @xmath0 .
section [ sec : lattice_formulation ] contains simulation parameters and technical details . in sec .
[ sec : result ] we present the simulation results for the cp - odd form factors together with the cp - even ones .
it is demonstrated that we can not obtain the correct value for the cp - odd form factor without subtracting the contribution from the parity - even form factors .
we conclude our investigation in sec .
[ sec : conclusion ] .
we consider the electromagnetic form factors of the nucleon , defined by n ( p , s)j_^emn(,s^ ) & = & |u(p , s)u(,s^ ) , where @xmath26 is the momentum transfer , @xmath27 is the on - shell nucleon state with momentum @xmath28 , energy @xmath29 and helicity @xmath30 . the electromagnetic current for quarks , @xmath31 ,
is given by j_^em = _ f e_f|_f _ _ f , where @xmath32 is the quark field with flavor @xmath33 and the electric charge @xmath34 .
in the small momentum transfer limit , @xmath35 , the above form factor is described by the following effective interaction : |n ( x ) n(x ) , where @xmath36 is the u(1 ) electromagnetic field .
we can define the electric dipole moment @xmath37 as _ n= _ q^20= . so far we do not specify cp properties of the system and therefore @xmath0 dependences are omitted in the above expressions . in the lattice calculation we must extract the form factors from the three - point correlation functions such as g_nj_n^(q , t , ) & & n(p , t ) j_^em(q,)|n(,0 ) , where @xmath38 and @xmath39 are the interpolating fields of the nucleon at time @xmath40 , which contain annihilation and creation operators with momentum @xmath28 . here
@xmath41 is the @xmath0 vacuum , which may be defined as = e^i q_5 0 , q_5 = d^3 x k_0 ( , t ) , using the conserved ( but gauge non - invariant ) chiral current @xmath42 . throughout this paper
, we do not use the explicit representation of @xmath43 . inserting the complete set of states between the interpolating fields and the current , we obtain g_nj_n^(q , t , ) & = & e^-e_n^(t-)e^-e_n^^t & & _ s , s^n n^(,s)n^(,s)j_^emn^(,s^)n^(,s^)|n+ , where @xmath44 , @xmath45 , and the dots represent exponentially suppressed contributions , which are assumed to be negligible .
hereafter we represent explicitly superscript of @xmath0 on @xmath0 vacuum effects .
since the interpolating fields @xmath46 and @xmath47 can be expanded as n(,0 ) & = & z_n^_s a_n^(,s ) u_n^(p , s ) + , |n(,0 ) & = & ( z_n^)^ * _ s a_n^^(,s ) | u_n^(p , s)+ , [ eq : nfieldexpand ] we have n n^(,s)&= & z_n^u_n^(p , s ) , n^(,s^)|n&= & ( z_n^)^ * |u_n^(,s^ ) , since n^(,s)&= & a_n^^(,s ) + n^(,s)&= & a_n^(,s ) . in eq .
( [ eq : nfieldexpand ] ) the dots represent contributions from all other possible states including the negative energy state of the nucleon , the negative parity states or other excited states . except a special case considered in the appendix
, we will not consider these states in this paper .
the spinors @xmath48 and @xmath49 are on - shell nucleon wave functions which satisfy the dirac equation : ( ip + m_n^e^-i f_n()_5 ) u_n^(,s)&= & |u_n^(,s)(ip + m_n^e^-i f_n()_5 ) = 0 . since cp is broken in the @xmath0 vacuum , a cp non - invariant phase factor @xmath50 can appear in the mass term .
this means that @xmath51 must be odd in @xmath0 , while @xmath4 and @xmath52 are even in @xmath0 .
therefore we have m_n^ & = & m_n + o(^2 ) , z_n^= z_n + o(^2 ) f_n ( ) & = & f_n^1 + o(^3 ) for small @xmath0 . the bi - spinor projection satisfies _
s u_n^(,s)|u_n^(,s ) & = & [ eq:2-pt_in_theta ] with the normalization that @xmath53 . the form factor , which we wish to extract from the three - point function , is written as n^(,s)j_^emn^(,s^)&= & |u_n^(,s ) w_^(q ) u_n^(,s^ ) it satisfies |u_n^(,s)q_w_^(q ) u_n^(,s^ ) = 0 from the current conservation @xmath54 .
taking account of the parity properties , the form factors in general should have the form that w_^(q ) & = & g(^2 ) w_^even(q ) + i h(^2 ) w_^odd(q ) , where we take @xmath55 for their normalizations and w_^even(q)&= & _ f_1(q^2)+q _ _ , + w_^odd(q)&= & q___5 + f_a(q^2)(q_q-_q^2)_5 with @xmath56/2 $ ] . here
@xmath57 are the electromagnetic form factors and @xmath58 is called `` anapole form factor '' . as mentioned in sec .
[ subsec : edm_def ] , @xmath59 gives the nucleon electric dipole moment .
for small @xmath0 , we can expand @xmath60 and @xmath61 . using all the necessary expressions and expanding in terms of @xmath0
, we obtain g_nj_n^(q , t , ) & = & z_n^2 e^-e_n(t-)e^-e_n^t & & + o(^2 ) . [ eq:3pt_ff ] on the other hand
, @xmath62 can be evaluated by the path integral as g_nj_n^(q , t , ) & = & n(,t ) j_^em(q , ) |n(,0)e^iq & = & g_nj_n ( q , t , ) + ig_nj_n^q ( q , t , ) + o(^2 ) [ eq:3pt_pi ] with g_nj_n ( q , t , ) & = & n(,t ) j_^em(q , ) | n(,0 ) , + g_nj_n^q ( q , t , ) & = & n(,t ) j_^em(q , ) |n(,0 ) q , where @xmath46 , @xmath47 , and @xmath63 are c - number operators in the path - integral , albeit using the same notation as the quantum operators , and @xmath64 is the topological charge , defined by q = d^4xg_(x)g_(x ) in the continuum theory . the symbol @xmath65 denotes the path - integral average of an operator @xmath66 in qcd with @xmath67 , which is given by = | a _
e^s_g(a ) + |d(a ) o , where @xmath68 and @xmath69 are the qcd gauge and quark actions , respectively . by comparing eqs.([eq:3pt_ff ] ) and ( [ eq:3pt_pi ] ) we obtain g_nj_n ( q , t , ) & = & z_n^2 e^-e_n(t-)e^-e_n^t w_^even(q ) , [ eq:3pt_even ] + g_nj_n^q ( q , t , ) & = & z_n^2 e^-e_n(t-)e^-e_n^t .[eq:3pt_odd ] these formulae are one of the main results of this paper . they show that the parity - even form factor @xmath70 multiplied with the parity - odd term @xmath71 of the two - point function has to be removed from the three - point function @xmath72 at order @xmath0 , in order to obtain @xmath73 .
we will see in sec .
[ sec : result ] how significant this subtraction is .
in order to determine @xmath71 we consider the two - point function of the nucleon given by g_nn^(p , t ) & & n(,t ) |n(,0)&= & z_n^^2 e^-e_n^ t [ eq:2ptoperator ] + & = & n(,t ) |n(,0 ) e^iq . [ eq:2ptpathintegral ] by comparing the operator expression ( [ eq:2ptoperator ] ) with the path integral expression ( [ eq:2ptpathintegral ] ) at each order of @xmath0 , we obtain g_nn(,t ) & & n(,t ) |n(,0)= z_n^2 e^-e_n t at @xmath74 and g_nn^q(,t ) & & n(,t ) |n(,0 ) q= z_n^2 e^-e_n t _ 5 at @xmath75 .
these formulae tell us that @xmath71 can be determined numerically from the nucleon propagators through the appropriate insertion of @xmath64 .
for the quark field we employ the domain - wall fermion action@xcite : s_f & = & -_x , s , x,s|(x , s)d_dw(x , s;x,s)(x,s ) + _ x m_f|q(x)q(x ) , + d_dw(x , s;xs ) & = & d^4(x , x)_s , s + _ x , xd^5(s , s ) + ( m-5)_x , x_s , s , + d^4(x , x ) & = & _ , + d^5(s , s ) & = & \ { cc ( 1-_5)_2,s & ( s=1 ) + ( 1-_5)_s+1,s + ( 1+_5)_s-1,s & ( 1<s < n_s ) + ( 1+_5)_n_s-1,s & ( s = n_s ) , . where @xmath76 are four dimensional space - time coordinates , @xmath77 are coordinates in the fifth dimension of length @xmath78 , and @xmath79 is the domain - wall height .
here @xmath80 are four dimensional physical quark fields given by q(x ) & = & ( 1-_5)(x,1 ) + ( 1+_5)(x , n_s ) , + bare quark mass .
this action has exact chiral symmetry at @xmath81 , and therefore it is expected that the fermion fields respond properly to the topological charge of the gauge fields . for the gauge field
we employ the following improved action : s_g = \{c_0_plaqu_plaq + c_1_rectu_rect } , [ eq : gauge_action ] where the first term represents the plaquette and the second the @xmath82 rectangle loop . in this paper
we take @xmath83 and @xmath84 , proposed by iwasaki using the renormalization group analysis@xcite . the action with this parameter choice
is called the rg improved gauge action . in continuum
qcd , the topological charge density is defined by q(x ) = _ g_(x)g_(x ) and the topological charge is given by the integration of @xmath85 over space - time : q = d^4xq(x ) . in our lattice calculation
we define the topological charge by the improved bosonic form@xcite , which is given by q_imp & = & _ x[top_def ] + q_plaq & = & _ ( f^p_f^p _ ) , q_rect = _ ( f^r_f^r _ ) where @xmath86 and @xmath87 are field strengths constructed from the plaquette clover and the rectangular clovers as depicted in fig .
[ fig : clover ] .
their explicit expressions are given by f^p_(x)&=&_x , , , + f^r_(x ) & = & _ x,,\{&+ & } . in order to remove @xmath88 discretization errors in @xmath89 at the tree level , we take @xmath90 and @xmath91@xcite . to compute nucleon three - point functions , we construct the amplitudes for the six diagrams in fig.[diag_u ] , each of which consists of two quark propagators and one double quark propagator with an insertion of the electromagnetic current @xmath92 : j_^em(x ) = e^u v^u_(x ) + e^d v_^d(x ) + e^s v_^s(x ) with @xmath93 the conserved vector current of flavor @xmath94 defined by v^i_(x ) & = & _ s j^i_(x , s ) , + j^i_(x , s ) & = & .
since this current is conserved , the renormalization factor satisfies @xmath95 . to insert @xmath96
, we use the ordinary source method .
we do not consider the disconnected diagrams depicted in fig .
[ discdiag ] , since this contribution vanishes in the su(3 ) flavor symmetric limit due to the fact that @xmath97 . in future investigations , however , it is necessary to check if this contribution is indeed small or not for @xmath98 .
we generate quenched gauge configurations on a @xmath99 lattice with the rg improved gauge action at @xmath100 .
gauge configurations separated by 200 sweeps are used for measurements after 2000 sweeps for thermalization , where one sweep consists of one heat - bath update followed by four over - relaxation steps .
the quark propagator is calculated using the domain - wall fermion with @xmath101 and @xmath102 , where @xmath78 is a size of the fifth dimension and @xmath79 is the domain - wall height .
these parameters are the same as those employed in ref.@xcite , in which the lattice spacing was determined to be @xmath103gev from the @xmath104 meson mass .
the residual quark mass @xmath105 , which represents the explicit chiral symmetry breaking effect at finite @xmath78 , is found to be @xmath106mev .
we take the bare quark mass @xmath107 corresponding to the ratio of pseudoscalar to vector meson masses @xmath108@xcite . to reduce the effects of unwanted excited hadron states , we calculate the quark propagator employing an exponentially smeared source , given by @xmath109 with @xmath110 and @xmath111 , where @xmath112 is the distance from the origin at @xmath113 in the lattice unit .
we take the periodic boundary condition for both gauge and fermion fields in order to minimize possible violation of the equivalence between the bosonic and the fermionic definitions of topological charge . in the calculation of three - point functions ,
the electromagnetic current is inserted at the fixed time slice @xmath114 .
the nucleon sink is varied as a function of @xmath40 with the nucleon source fixed at @xmath115 .
we move the final state nucleon with one of the three non - zero spatial momenta @xmath116 , while the initial state nucleon is placed at rest @xmath117 .
we accumulate 730 configurations for measurements .
errors are estimated by the single elimination jackknife procedure for all measured quantities .
in order to reduce ultra - violet fluctuations of gauge configurations in the calculation of the topological charge , we employ the cooling method using the action in eq.([eq : gauge_action ] ) with @xmath118 and @xmath119 . in fig .
[ cool20 ] we show both @xmath120 and @xmath89 as a function of the cooling steps , where @xmath120 is the naive plaquette definition of the topological charge and @xmath89 is the improved one in eq.([top_def ] ) .
it is clearly observed that @xmath89 approach integer values much faster than @xmath120 as the cooling step increases .
the values after 20 cooling steps are also closer to integer values .
the deviation from integer is within a few percent for @xmath89 after 20 cooling steps . as the topological charge ,
we employ @xmath89 measured after 20 cooling steps on each configuration . in fig .
[ conf_topo ] we show the time history of the topological charge as a function of the gauge configuration separated by 200 sweeps , where we do not recognize any long autocorrelation .
in addition , the histogram of the topological charge in fig . [ hist ] exhibits nearly a gaussian distribution .
these observations suggest that the sampling of the topological charge is sufficiently good in our calculation . from the analysis in sec .
[ sec : formula_2pt ] , the @xmath121 contribution of the nucleon propagator should behave as = \ { cc z_n^2 e^-m_nt+ & ( |p|=0 ) + z_n(p)^2e^-e_nt + & ( |p|=1 ) .
, where @xmath122 and the dots represent the contribution from the excited states . here @xmath123 is a shorthand notation of @xmath124 with @xmath125 .
the nucleon propagators from the smeared source and the point sink with @xmath126 and @xmath127 are plotted in fig .
[ npp ] , and the corresponding effective masses are given in fig.[nem ] .
choosing @xmath128 for the fitting range , we obtain e_na & = & \ { cc 0.7114(43 ) & ( |p|=0 ) + 0.8068(77 ) & ( |p|=1 ) . , z_n & = & \ { cc 2162(87 ) & ( |p|=0 ) + 1453(104 ) & ( |p|=1 ) .
note that @xmath129 at @xmath130 agrees with the relativistic dispersion relation @xmath131 within the error . in sec .
[ sec : formula_2pt ] we have shown that the parity - odd part of the nucleon propagator should have the following form at the order of @xmath75 : & = & |z_n|^2f_n^1 e^-m_nt,[eq : odd_part0 ] + & = & |z_n(p)|^2f_n^1 e^-e_nt .
[ eq : odd_part ] in fig.[npp_q_gm5 ] we plot the parity - odd part with the insertion of @xmath64 , while the parity - even part without @xmath64 , @xmath133 $ ] , is given in fig.[npp_noq_gm5 ] for comparison .
the latter vanishes at all time slices as expected . on the other hand ,
the former is non - zero , showing an exponential fall - off at large @xmath40 .
after averaging over the forward and backward propagators in time , we fit the party - odd part at @xmath134 , by the form @xmath135 with @xmath136 fixed to the value obtained from the previous fit of the @xmath121 contribution .
the fit , represented by the solid line in fig .
[ npp_q_gm5 ] , is reasonably good , showing that the decay rate of the parity - odd part is consistent with that of the parity - even part for both @xmath126 and 1 .
note however that the small deviation indicates that the present statistics are still insufficient at large @xmath40 for @xmath127 .
another confirmation of consistency is found in the effective mass plot of fig .
[ neff_q_gm5 ] , where the parity - odd part of the nucleon propagator with @xmath64 and the parity - even part without @xmath64 are compared .
with the aid of @xmath136 and @xmath52 obtained previously , @xmath137 is converted to @xmath71 as f^1_n = \
{ cc -0.247(17 ) & ( |p|=0 ) + -0.243(20 ) & ( |p|=1 ) . .
despite the deviation at large @xmath40 for @xmath127 , the two estimates of @xmath71 agree within the errors . in our analysis
we will use @xmath71 obtained from @xmath126 .
we note that the exponential fall - off of @xmath138 is very sensitive to the distribution of the topological charge .
one could use this quantity to judge whether the sampling of the topological charge is sufficiently good : if the exponential fall - off of @xmath138 does not agree with that of @xmath139 , the sampling is not satisfactory . in order to remove the exponential fall - off due to the nucleon propagation in the three - point functions , we divide them by the two - point function @xmath140 with the smeared source at @xmath115 and the point sink at @xmath40 : ^e_(q;t , ) & = & r(q;,t ) , [ 3pt_ov_nne ] + ^o_(q;t , ) & = & r(q;,t)[3pt_ov_nno ] with r(q;,t ) = ^1/2 , where the factor @xmath141 is to remove the interpolating field dependent normalization factors . here
we introduce a new two - point function @xmath142 that has the point source at @xmath115 and the point sink at @xmath40 , and @xmath143 represents that the source point is started at @xmath144 . in fig .
[ corr ] we plot the @xmath40 dependence of @xmath145 with @xmath114 and @xmath146 .
as we expect , the correction factor @xmath141 is independent of @xmath40 in the large @xmath40 region .
the form factors are extracted from the ratios defined in eqs.([3pt_ov_nne ] ) and ( [ 3pt_ov_nno ] ) by applying appropriate projections of gamma matrices . from the @xmath121 contribution @xmath147 of eq.([3pt_ov_nne ] ) , @xmath148 and @xmath149 can be extracted as & = & ( f_1(q^2 ) + f_2(q^2 ) ) , [ elec_mom ] + & = & _ ijk(f_1(q^2 ) + f_2(q^2 ) ) , [ magne_mom ] where we use eq.([eq:3pt_even ] ) .
the @xmath148 and @xmath149 are related to the electric ( @xmath150 ) and the magnetic ( @xmath151 ) form factors as g_e(q^2 ) & = & f_1(q^2 ) + f_2(q^2 ) [ ge_f12 ] , + g_m(q^2 ) & = & f_1(q^2 ) + f_2(q^2)[gm_f12 ] .
these quantities become the electric charge and the magnetic moment at @xmath152 .
therefore @xmath153 and @xmath154@xcite for the neutron(proton ) .
we can extract the form factor @xmath155 , which is relevant to the electric dipole moment , in two different ways from @xmath156 of eq.([3pt_ov_nno ] ) : & = & f_3(q^2 ) & + & f^1_n , [ f3_proj1 ] + & = & -q_if_3(q^2)&+ & f^1_n[f3_proj2 ] , where we use eq.([eq:3pt_odd ] ) .
it will be clear that the consistency between the two different extractions of @xmath155 provides a crucial test for the validity of actual calculations .
note that the anapole form factors which appears in @xmath157 is canceled out after applying the trace of spinor with these projections .
we first consider the parity - even part of the form factor . in figs .
[ gm ] and [ ge ] we show the magnetic and electric form factors obtained from eqs.([elec_mom ] ) and ( [ magne_mom ] ) .
we observe good plateaux except for the electric form factor of the neutron .
employing a constant fit with @xmath158 for @xmath151 and @xmath159 for @xmath150 we obtain g_m(q^2 ) & = & \ { cl -0.591(37 ) & + 0.952(60 ) & .
+ g_e(q^2 ) & = & 0.502(33 ) at @xmath160 gev@xmath161 . here
the value for @xmath162 is not quoted since it is difficult to find a plateau for the fit .
note also that the value of @xmath150 slightly depends on the fitting range , as could be seen from the figure .
these values are compared with the previous lattice result@xcite obtained at a similar nucleon mass and lattice spacing using the nonperturbatively @xmath163 improved wilson quark and the plaquette gauge actions : @xmath164 for the neutron , and @xmath165 and @xmath166 for the proton .
experimentally the proton electromagnetic form factors at small @xmath167 have been measured as@xcite , g_m^p(exp)(q^2=0.58gev^2 ) = 0.8480.011 , g_e^p(exp)(q^2=0.58gev^2 ) = 0.2940.008 .
recent experimental measurements for neutron form factors cover a wide range of @xmath167 from very small @xmath168 gev@xmath161 to more than @xmath169 gev@xmath161 with high precision@xcite .
for example , the neutron magnetic form factor at @xmath170 gev@xmath161 is measured as@xcite g_m^n(exp)(q^2=0.6 gev^2 ) = -0.568 0.007 ( stat.)0.015 ( syst . ) , and the electric form factor at @xmath171 gev@xmath161 is@xcite g_e^n(exp)(q^2=0.5 gev^2 ) = 0.04630.0062 ( stat.)0.0034 ( syst . ) .
the fact that our results are not so much different from these experimental values may suggest that systematic errors involved in our calculations , such as the large quark mass , the small lattice size , the non - zero lattice spacing and the quenching , are not so large on these quantities .
the form factors @xmath148 and @xmath149 are obtained from @xmath151 and @xmath150 using eqs.([ge_f12 ] ) and ( [ gm_f12 ] ) .
we show the time dependence of @xmath148 and @xmath149 in figs .
[ f1 ] and [ f2 ] .
we observe plateaus at @xmath172 for @xmath173 and @xmath174 but not for @xmath175 , which behaves badly as @xmath162 does .
we think that this behavior is caused by the contribution from other excited states , as pointed out in ref.@xcite .
therefore we have tried to include the contributions of the parity - odd nucleon state @xmath176 in our analysis , using the formulae given in appendix [ app1 ] .
although larger errors prevent us from extracting the reliable values for @xmath162 and @xmath175 , we find a much better plateau for @xmath162 as shown in fig .
[ ge_nodd ] and for @xmath175 in fig .
[ f12_nodd ] . applying a constant fit to @xmath177 and @xmath173 with @xmath178
, we obtain f_1^p(q^2 ) = 0.515(37 ) , f_2^n(q^2 ) = -0.560(40 ) , f_2^p(q^2 ) = 0.399(37 ) , which are compared with @xmath179 , @xmath180 from ref.@xcite .
we now consider the main target of our calculation : the parity - odd part of the form factors .
substituting @xmath181 , @xmath182 , @xmath183 and @xmath71 into eqs.([f3_proj1 ] ) and ( [ f3_proj2 ] ) , we extract the form factor @xmath155 in two different ways .
we first plot @xmath184 $ ] and @xmath185 $ ] in figs .
[ f3_split ] for the nucleon and in [ f3_split2 ] for the proton , together with the second terms in the right hand side of eqs.([f3_proj1 ] ) and ( [ f3_proj2 ] ) , composed of @xmath148 , @xmath149 and @xmath71 . comparing the two extractions , we find that @xmath184 $ ] and @xmath185 $ ] shown by open symbols do not agree with each other .
in particular , for the case of the proton , @xmath184 $ ] is very large , while @xmath185 $ ] is almost zero .
we note that @xmath184 $ ] and @xmath185 $ ] are almost zero for the neutron is consistent with the previous result@xcite , in which the contribution of the mixing term was not considered .
figure [ f3 ] shows the result after subtracting the second terms from @xmath184 $ ] ( circles ) and @xmath185 $ ] ( triangles ) for the neutron ( top panel ) and the proton ( bottom panel ) .
it is gratifying to observe that the two ways of projections give values of @xmath1 consistent with each other .
in particular , the significant difference between the two projections for the proton almost vanishes after removing the second terms . making a constant fit to the result of eq.([f3_proj2 ] ) over @xmath186 , we obtain f_3(q^20.58 ^ 2 ) = \ { rc -0.024(5 ) e & ( ) + 0.021(6 ) e & ( ) .with @xmath103gev@xcite .
note that the opposite sign between neutron and proton is consistent with the estimate of ref.@xcite , where the baryon edm is proportional to its magnetic moment .
it should be remembered that our values are obtained at a non - zero @xmath167 . with the same caution in mind ,
our result may be compared with other estimates of @xmath187 : @xmath188fm@xcite , @xmath189fm@xcite and @xmath190fm@xcite .
in this paper we have carried out a feasibility study for the lattice qcd calculation of the neutron electric dipole moment in the presence of the @xmath0 term .
we took the strategy to extract the nucleon edm from the nucleon electromagnetic form factor at small @xmath0 .
we explained how one could extract the parity - odd part of the form factors , which becomes edm at zero momentum transfer , from the relevant correlation functions for small @xmath0 .
we have pointed out that the contribution of the parity - even form factors multiplied by the parity - odd part of the nucleon two - point function has to be subtracted from the parity odd three - point function .
we have derived the formula to carry out this subtraction , which is one of the main results . in the second half
, we applied our formula to an actual lattice qcd calculation , employing the domain - wall quark action with the rg improved gauge action in quenched qcd .
we used a @xmath6 lattice at @xmath103gev with the quark mass of @xmath24 , which corresponds to @xmath191 . in order to obtain a sufficient sampling of topological charges
, we accumulated 730 configurations , which enabled us to extract the parity - odd part both in the nucleon two- and three - point functions .
we have shown that two different @xmath9 matrix projections yield values for @xmath1 consistent with each other , if the subtractions mentioned above are made .
our lattice calculation gives @xmath192 for neutron , @xmath193 for proton at @xmath194 . in order to obtain the physical value of nedm
, we have to perform various extrapolations : @xmath152 extrapolation , the chiral extrapolation ( @xmath195 ) and the continuum extrapolation ( @xmath196 ) .
in addition , we have to remove the quenching error in nedm , since nedm vanishes at zero quark mass in qcd but it does not in quenched qcd . while it would be possible to remove each of these systematic errors step by step using the method proposed in this paper , we are now looking for better alternatives with which we can remove them more easily .
s. a. would like to thank prof .
a. kronfeld for useful discussion , which leads to the simpler derivation of eq.([eq:2-pt_in_theta ] ) .
we thank prof .
m. fukugita for his valuable comments on the manuscript .
this work is supported in part by grants - in - aid of the ministry of education ( nos .
12640253 , 13640259 , 13640260 , 14046202 , 15204015 , 15540251 , 15540279 , 15740134 , 15740165 , 16028201 , 16540228 , 17540249 ) .
the numerical simulations have been carried out on the parallel computer cp - pacs .
if we include the electromagnetic transistion from @xmath197 to @xmath176 , eqs.([elec_mom ] ) and ( [ magne_mom ] ) are modified as & = & ( f_1(q^2 ) + f_2(q^2))&+ & _ + -(t , ) , [ fodd1 ] + & = & _ ijk(f_1(q^2 ) + f_2(q^2))&+ & _ ijk_+-(t,),[fodd2 ] where = , _ + -(t , ) = e^-(e_n_--e_n)(t- ) and @xmath198 is the parity - odd nucleon mass and @xmath199 and @xmath200 are the @xmath197 to @xmath176 transition form factors with @xmath201 .
@xmath202 is the parity - odd nucleon amplitude which can be determined from the @xmath176 propagator . in order to obtain these form factors we have to consider additional projection formula as follows : & = & ( f_1(q^2 ) + f_2(q^2 ) ) & + & _
+ -(t,),[fodd3 ] + & = & f_1(q^2 ) - f_2(q^2)&+ & _ + -(t,),[fodd4 ] + & = & _ ijf_1(q^2 ) + f_2(q^2)&- & _ + -(t,).[fodd5 ] employing four equations out of the above five formula we can determine @xmath203 and @xmath204 . the form factors @xmath148 and @xmath149 plotted in figs . [ ge_nodd ] and [ f12_nodd ] .
are obtained with the use of eqs.([fodd3 ] ) , ( [ fodd4 ] ) and ( [ fodd5])(@xmath205 and @xmath206 ) . in fig .
[ nem_odd ] we show the effective mass plot for the parity - odd nucleon . as for @xmath207
we have subtracted the @xmath197 contribution from @xmath208 $ ] with @xmath209 , in order to obtain the energy @xmath210 for the parity - odd state .
the global fit with @xmath211 for @xmath198 and @xmath212 for @xmath210 gives m_n_-a=1.017(8 ) , e_n_-a=1.056(18 ) , where @xmath213 is close to the value expected from the relativistic dispersion relation @xmath214 .
90 p. g. harris et al . , phys .
82 * , 904 ( 1999 ) . v. f. dmitriev , r. a. senkov , phys .
. lett . * 91 * , 212303 ( 2003 ) .
m. v. romalis , w. c. griffith , j. p. jacobs , e. n. fortson , phys .
lett . * 86 * , 2505 ( 2001 ) .
r. j. crewther , p. di vecchia , g. veneziano , e. witten , phys . lett . * b88 * , 123 ( 1979 ) ; erratum , _ ibid .
_ * b91 * , 487 ( 1980 ) .
p. di vecchia , acta phys .
austriaca suppl .
* 22 * , 341 ( 1980 ) .
r. d. peccei , h. r. quinn , phys .
* 38 * , 1440 ( 1977 ) ; r. d. peccei , adv .
high energy phys . 3 , 503 - 551 ( 1989 ) . m. pospelov , a. ritz , nucl
* b558 * , 243 ( 1999 ) ; nucl . phys . * b573 * , 177 ( 2000 ) ; chuan - tsung chan , e. m. henley , t. meissner , hep - ph/9905317 ; m. pospelov , a. ritz , phys .
* 27 * , 2526 ( 1999 ) ; phys . rev . *
d63 * , 073015 ( 2001 ) .
b. borasoy , phys . rev . *
d61 * , 114017 ( 2000 ) .
s. aoki , a. gocksch , phys .
lett . * 63 * , 1125 ( 1989 ) ; erratum , _ ibid . _ * 65 * , 1172 ( 1990 ) .
s. aoki , a. gocksch , a. v. manohar , s. r. sharpe , phys .
* 65 * , 1092 ( 1990 ) .
d. guadagnoli , v. lubicz , g. martinelli , s. simula , jhep * 0304*,019 ( 2003 ) .
p. faccioli , d. guadagnoli , s. simula , phys .
rev . * d70 * , 074017 ( 2004 ) .
f. berruto , t. blum , k. orginos , a. soni , nucl .
140 , 411 ( 2005 ) .
s. aoki , t. hatsuda , phys . rev . *
d45 * , 2427 ( 1992 ) .
d. b. kaplan , phys .
* b288 * , 342 ( 1992 ) ; y. shamir , nucl . phys . * b406 * , 90 ( 1993 ) ; v. furman , y. shamir , nucl . phys . *
b439 * , 54 ( 1995 ) .
r. narayanan , h. neuberger , nucl . phys . *
b443 * , 305 ( 1995 ) ; h. neuberger , phys . lett .
* b417 * , 141 ( 1998 ) ; h. neuberger , phys .
b427 * , 353 ( 1998 ) ; h. neuberger , phys . rev . *
d57 * , 5417 ( 1998 ) .
cp - pacs collaboration , a. ali khan @xmath215 @xmath216 . , phys . rev .
* d63 * , 114504 ( 2001 ) .
y. iwasaki , nucl .
b258 * , 148 ( 1985 ) .
cp - pacs collaboration , a. ali khan @xmath215 @xmath216 . , phys . rev . * d64 * , 114501 ( 2001 ) .
p. weisz , nucl .
b212 * , 1 ( 1983 ) ; p. weisz , r. wohlert , nucl . phys . *
b236 * , 397 ( 1984 ) ; erratum , _ ibid . _
* b247 * , 544 ( 1984 ) ; m. lscher , p. weisz , commun .
phys . * 97 * , 59 ( 1985 ) ; erratum , _ ibid . _
* 98 * , 433 ( 1985 ) .
cp - pacs collaboration , a. ali khan @xmath215 @xmath216 . , phys . rev .
* d64 * , 114506 ( 2001 ) .
s. eidelman _
et al . _ ,
b592 * , 1 ( 2004 ) .
m. gckeler @xmath215 @xmath216 .
, phys . rev . *
d71 * , 034508 ( 2005 ) .
l. e. price , @xmath215 @xmath216 .
* d4 * , 45 ( 1971 ) .
gao , int .
* e12 * , 1 ( 2003 ) . w. xu , @xmath215 @xmath216 . ,
phys . rev . * c67 * , r012201 ( 2003 ) .
h.zhu , @xmath215 @xmath216 .
87 * , 081801 ( 2001 ) .
a. tang , w. wilcox , r. lewis , phys . rev . *
d68 * , 094503 ( 2003 ) .
em current is connected with @xmath217 quark line , and in @xmath218 with @xmath219 quark line.,title="fig:",height=226 ] em current is connected with @xmath217 quark line , and in @xmath218 with @xmath219 quark line.,title="fig:",height=226 ] + em current is connected with @xmath217 quark line , and in @xmath218 with @xmath219 quark line.,title="fig:",height=226 ] em current is connected with @xmath217 quark line , and in @xmath218 with @xmath219 quark line.,title="fig:",height=226 ] + em current is connected with @xmath217 quark line , and in @xmath218 with @xmath219 quark line.,title="fig:",height=226 ] em current is connected with @xmath217 quark line , and in @xmath218 with @xmath219 quark line.,title="fig:",height=226 ] $ ] , @xmath121 contribution of the nucleon propagator with @xmath220 projection for @xmath126 ( top ) and @xmath127 ( bottom ) .
solid lines denote the fitting results .
the parity - odd state dominates the backward propagation in @xmath40 denoted by triangles .
, title="fig:",width=377 ] $ ] , @xmath121 contribution of the nucleon propagator with @xmath220 projection for @xmath126 ( top ) and @xmath127 ( bottom ) .
solid lines denote the fitting results .
the parity - odd state dominates the backward propagation in @xmath40 denoted by triangles .
, title="fig:",width=377 ] ( top ) and @xmath127 ( bottom ) .
solid(dotted ) lines denote the central values(errors ) of the global fit.,title="fig:",width=377 ] ( top ) and @xmath127 ( bottom ) .
solid(dotted ) lines denote the central values(errors ) of the global fit.,title="fig:",width=377 ] $ ] , parity odd part of the nucleon propagator with @xmath64 for eq.([eq : odd_part0 ] ) @xmath126 ( top ) and eq.([eq : odd_part ] ) @xmath127 ( bottom ) .
the sign of propagator is negative .
solid lines denote the fitting results.,title="fig:",width=377 ] $ ] , parity odd part of the nucleon propagator with @xmath64 for eq.([eq : odd_part0 ] ) @xmath126 ( top ) and eq.([eq : odd_part ] ) @xmath127 ( bottom ) .
the sign of propagator is negative .
solid lines denote the fitting results.,title="fig:",width=377 ] $ ] , parity odd part of the nucleon propagator without @xmath64 for eq.([eq : odd_part0 ] ) @xmath126 ( top ) and eq.([eq : odd_part ] ) @xmath127 ( bottom).,title="fig:",width=377 ] $ ] , parity odd part of the nucleon propagator without @xmath64 for eq.([eq : odd_part0 ] ) @xmath126 ( top ) and eq.([eq : odd_part ] ) @xmath127 ( bottom).,title="fig:",width=377 ] ( filled ) for eq.eq.([eq:odd_part0 ] ) @xmath221 ( top ) and eq.([eq : odd_part ] ) @xmath127 ( bottom ) after averaged with backward propagator in time direction .
the parity even part without @xmath64 ( open ) , which are already shown in fig .
[ nem ] , are also plotted for comparison.,title="fig:",width=377 ] ( filled ) for eq.eq.([eq:odd_part0 ] ) @xmath221 ( top ) and eq.([eq : odd_part ] ) @xmath127 ( bottom ) after averaged with backward propagator in time direction .
the parity even part without @xmath64 ( open ) , which are already shown in fig .
[ nem ] , are also plotted for comparison.,title="fig:",width=377 ] for neutron ( top ) and proton ( bottom ) .
horizontal lines represent the constant fit.,title="fig:",width=377 ] for neutron ( top ) and proton ( bottom ) .
horizontal lines represent the constant fit.,title="fig:",width=377 ] for neutron ( top ) and proton ( bottom ) .
horizontal lines represent the constant fit.,title="fig:",width=377 ] for neutron ( top ) and proton ( bottom ) .
horizontal lines represent the constant fit.,title="fig:",width=377 ] $ ] ( open ) and the mixing term ( filled ) in eq.([f3_proj1 ] ) .
bottom : @xmath185 $ ] ( open ) and the mixing term ( filled ) in eq.([f3_proj2 ] ) .
, title="fig:",width=377 ] $ ] ( open ) and the mixing term ( filled ) in eq.([f3_proj1 ] ) .
bottom : @xmath185 $ ] ( open ) and the mixing term ( filled ) in eq.([f3_proj2 ] ) .
, title="fig:",width=377 ] for neutron ( top ) and proton ( bottom ) .
circles represent the result with @xmath222 projection of eq.([f3_proj1 ] ) , while triangles represent the one with @xmath223 projection averaged over @xmath224 in eq.([f3_proj2]).,title="fig:",width=377 ] for neutron ( top ) and proton ( bottom ) .
circles represent the result with @xmath222 projection of eq.([f3_proj1 ] ) , while triangles represent the one with @xmath223 projection averaged over @xmath224 in eq.([f3_proj2]).,title="fig:",width=377 ] ( top ) and @xmath127 ( bottom ) .
horizontal lines denote the central value ( solid ) and the errors ( dotted ) of the global fit.,title="fig:",width=377 ] ( top ) and @xmath127 ( bottom ) .
horizontal lines denote the central value ( solid ) and the errors ( dotted ) of the global fit.,title="fig:",width=377 ]
|
we carry out a feasibility study for the lattice qcd calculation of the neutron electric dipole moment ( nedm ) in the presence of the @xmath0 term .
we develop the strategy to obtain the nucleon edm from the cp - odd electromagnetic form factor @xmath1 at small @xmath0 , in which nedm is given by @xmath2 where @xmath3 is the momentum transfer and @xmath4 is the nucleon mass .
we first derive a formula which relates @xmath1 , a matrix element of the electromagnetic current between nucleon states , with vacuum expectation values of nucleons and/or the current . in the expansion of @xmath0 , the parity - odd part of the nucleon - current - nucleon three - point function contains contributions not only from the parity - odd form factors but also from the parity - even form factors multiplied by the parity - odd part of the nucleon two - point function , and
therefore the latter contribution must be subtracted to extract @xmath1 .
we then perform an explicit lattice calculation employing the domain - wall quark action with the rg improved gauge action in quenched qcd at @xmath5 gev on a @xmath6 lattice . at the quark mass @xmath7 ,
corresponding to @xmath8 , we accumulate 730 configurations , which allow us to extract the parity - odd part in both two- and three - point functions . employing two different dirac @xmath9 matrix projections ,
we show that a consistent value for @xmath1 can not be obtained without the subtraction described above .
we obtain @xmath10 @xmath110.024(5 ) @xmath12fm for the neutron and @xmath10 0.021(6 ) @xmath12fm for the proton .
|
heavy - fermion ( hf ) superconductivity has been a matter of interest with respect to an intimate interplay between magnetism and superconductivity since the discovery of hf superconductivity in cecu@xmath15si@xmath15 @xcite . from the extensive experimental and theoretical works on cecu@xmath15si@xmath15 , it has been revealed that the unconventional superconducting ( sc ) phase takes place at the border of an antiferromagnetic ( afm ) phase and even coexists with it when the superconductivity occurs very close to the afm phase .
cecu@xmath15si@xmath15 was the only cerium - based hf superconductor at ambient pressure , before a new family of hf compounds ceirin@xmath0 and cecoin@xmath0 forming in two - dimensional ( 2d ) tetragonal structures was discovered to show superconductivity below @xmath1 = 0.4 k and 2.3 k , respectively @xcite .
remarkably , the value of @xmath16 k for the latter is a record of high @xmath1 to date among previous examples .
meanwhile , the hf antiferromagnet cerhin@xmath0 , which crystallizes in the same crystal structure , was discovered to show superconductivity below @xmath1 = 2.2 k at pressures exceeding a critical pressure of 1.63 gpa @xcite . in cemin@xmath0 ( m = ir , co , and rh ) ,
the unconventional superconductivity with line - node gap is suggested from the power - law temperature ( @xmath2 ) dependence of specific heat @xcite , nuclear relaxation rate @xmath3 @xcite , penetration depth @xcite , thermal conductivity @xcite , and so on . systematic investigation of this new family of hf superconductors allows us to unravel a relationship between the occurrence of unconventional superconductivity and possible spin fluctuations in the normal state .
in fact , a number of experiments point to non - fermi - liquid ( nfl ) behavior above @xmath1 in these compounds . for instance , in cecoin@xmath0 , the physical quantities such as the specific heat divided by temperature @xmath17 , the pauli susceptibility @xmath18 and @xmath19 reveal a significant increase upon cooling without entering any conventional fermi - liquid regime where these quantities stay constant @xcite .
the @xmath2 dependence of resistivity does not follow a conventional @xmath20 dependence but indicate approximately a @xmath2-linear dependence @xcite . in the magnetic - field ( @xmath13 ) induced normal state , @xmath17 continues to increase upon cooling down to 50 mk .
these nfl behaviors are considered to originate from magnetic fluctuations due to the proximity to an afm quantum critical point ( qcp ) . in cecoin@xmath0
, therefore , the unconventional superconductivity arises from the nfl normal state .
the afm spin fluctuations in ceirin@xmath0 and cerhin@xmath0 have been investigated by inelastic neutron diffraction @xcite and nuclear quadrupole resonance ( nqr ) experiments @xcite . in both the compounds , anisotropic afm spin fluctuations develop in the normal state .
the @xmath2 variation of @xmath3 in ceirin@xmath0 pointed to the presence of anisotropic afm spin fluctuations near the qcp @xcite .
systematic study on the family of cemin@xmath0 is promising in unraveling possible role of spin fluctuations for the emergence of superconductivity . in this paper
, we report results of in - nqr at zero field ( @xmath13 = 0 ) and co - nmr under a magnetic field @xmath21 t on single crystals of cecoin@xmath0 , focusing on the characteristics of spin fluctuations in the normal state . in the normal state at @xmath22 , @xmath3 varies nearly as @xmath23 over a wide @xmath2 range .
this unique behavior is consistent with the anisotropic afm spin - fluctuation s ( sfs ) model in case of the proximity to an afm qcp .
the terminology of the anisotropic afm spin fluctuations means here that a magnetic correlation length @xmath24 within the tetragonal plane develops more dominantly than @xmath25 does along the @xmath26-axis .
it is suggested that the value of @xmath16 k for cecoin@xmath0 becomes larger than @xmath1 = 0.4 k for ceirin@xmath0 because of approaching the afm qcp .
the afm spin fluctuations in cecoin@xmath0 are suppressed by the magnetic field , reinforcing that cecoin@xmath0 is closely located to the qcp .
high - quality single crystals of cecoin@xmath0 and lacoin@xmath0 were grown by in self - flux method as described elsewhere @xcite . for the nmr measurement in cecoin@xmath0
, we used several pieces of single crystals stacked with a shape of flat plate which undergo a sc transition at @xmath27 2.3 k. cecoin@xmath0 has two inequivalent in sites per a unit cell .
the @xmath4in - nqr @xmath28 measurement at @xmath22 was performed at the in(1 ) site located on the top and the bottom faces of the tetragonal unit cell at the nqr frequency @xmath29 8.16 mhz @xcite .
the co - nmr @xmath28 was measured under @xmath13 = 1.1 t. the quadrupole frequency of @xmath30co for cecoin@xmath0 is estimated as @xmath31 0.23 mhz from the co - nmr spectra , so it is experimentally difficult to detect co - nqr signals at @xmath13 = 0 . note that a principal axis of the electric field gradient is parallel to the tetragonal @xmath26-axis at the in(1 ) and co sites . in order to deduce the @xmath32-derived contribution of magnetic fluctuations , the in - nqr and co
- nmr @xmath28 measurements have been made on lacoin@xmath0 at @xmath22 and 0.9 t , respectively .
these measurements were carried out by a conventional phase - coherent laboratory - built pulsed nqr / nmr spectrometer .
figure 1 indicates the @xmath2 dependence of @xmath3 along with the data on lacoin@xmath0 at @xmath22 for the lowest quadrupole transition ( @xmath33 ) .
the @xmath2 dependence of @xmath3 is consistent with the result reported previously @xcite . in the sc state
, @xmath3 decreases rapidly without any coherence peak just below @xmath16 k. in contrast to the suggestion by kohori _
et al . _
@xcite , we remark that a simple behavior of @xmath34 is not valid , presumably affected by the close proximity to the afm qcp .
actually , a possibility has raised that the anomalous power - law @xmath2 behavior in london penetration depth at low @xmath2 well below @xmath1 is due to an extension of quantum criticality into the sc state @xcite . in this context
, a deviation from @xmath35 allows us to address what happens in system located very close to afm qcp when it becomes superconducting . in order to clarify this ,
further experiments under pressure are now in progress .
dependence of @xmath3 of @xmath4 in in cecoin@xmath0 along with the data in lacoin@xmath0 at @xmath22 .
the inset indicates the @xmath2 dependence of @xmath3 of @xmath30co at @xmath21 t in cecoin@xmath0 along with the data at @xmath36 t in lacoin@xmath0 . ] here , we focus on the characteristics of spin fluctuations in the normal state .
@xmath3 in lacoin@xmath0 is proportional to @xmath2 as seen in fig . 1 .
in cecoin@xmath0 , however , @xmath3 is faster than that in lacoin@xmath0 .
it is remarkable that @xmath37 varies nearly as @xmath23 in a wide @xmath2 range , where @xmath37 corresponds to the @xmath32 contribution of relaxation by subtracting a value of @xmath3 in lacoin@xmath0 at each temperature .
this novel power - law behavior in @xmath37 points to the presence of strong afm spin fluctuations . above @xmath38
150 k , @xmath3 is almost independent of temperature , exhibiting that the system is in a localized regime .
it is known in hf systems that @xmath39 is scaled to the quasielastic linewidth in neutron - scattering spectrum , leading to a tentative estimation of the bandwidth of hf state .
kohori _ et al .
_ claimed that @xmath6-like dependence would arise from the spin fluctuations at the afm qcp for isotropic three dimensional ( 3d ) systems . the sfs theory for strongly correlated electron systems near an afm instability revealed that @xmath40 and @xmath41 for 3d and 2d systems , respectively by assuming an expansion of dynamical susceptibility as @xmath42 . here
, the staggered susceptibility at the afm propagation vector @xmath43 , @xmath44 , follows the curie - weiss law above the nel temperature , namely , @xmath45 .
note @xmath46 k at the afm qcp . at the afm qcp or @xmath47 k for 3d systems
, @xmath3 is expected to be proportional to @xmath48 at high @xmath2 and varies as @xmath23 for @xmath49 smaller than about @xmath50 @xcite . here , @xmath49 is the reduced temperature defined by @xmath51 .
it is hence unexpected for isotopic 3d systems to vary nearly as @xmath6 as observed in cecoin@xmath52 over such the wide @xmath2 range .
we argue , therefore , this novel future in terms of the anisotropic afm - sfs model due to the layered structure .
let us now consider a case where a magnetic correlation length ( @xmath25 ) along the @xmath26-axis for afm spin fluctuations remains in a much shorter range than @xmath24 within the basal plane .
this is then modeled by assuming @xmath53 instead of @xmath54 . in this model ,
@xmath55 is derived @xcite .
recently , another calculation by assuming @xmath56 with being @xmath57 an anisotropy parameter has shown that physical properties are interpolated smoothly between 2 and 3 dimensions @xcite
. this result may be consistent with the claim that the observed @xmath2 dependence of @xmath6 in a wide @xmath2 range is due to anisotropic afm sfs , since this @xmath2 variation is intermediate behavior between those for 2d and 3d systems .
dependence of @xmath58 at @xmath22 in cecoin@xmath0 .
here @xmath59 .
the solid line indicates the relation @xmath60 with @xmath12 k in terms of the anisotropic afm spin - fluctuations ( sfs ) model ( see the text ) .
the dotted and broken line correspond to the respective relations in terms of the isotropic 2d- and 3d - afm - sfs models . ] the @xmath2 dependence of @xmath58 is shown in fig . 2 .
the data are consistent with the anisotropic sfs model with @xmath9 k as indicated by the solid line in the figure .
neither the 3d nor 2d sfs models are applicable as indicated by the dotted or broken lines .
the value of @xmath61 k close to zero demonstrates that cecoin@xmath0 is closely located to the afm qcp .
this anisotropic sfs model was already shown to interpret well the @xmath28 results of ceirin@xmath0 with @xmath62 = 8 k that is significantly larger than that for cecoin@xmath0 @xcite .
also the recent inelastic neutron experiment revealed that @xmath24 is larger than @xmath25 when approaching @xmath63 k in the hf antiferromagnet cerhin@xmath0 at ambient pressure @xcite .
these magnetic correlations in cemin@xmath0 are thus commonly characterized by the anisotropic afm spin fluctuations .
we now discuss the relation between the value of @xmath1 and what extent is closer to the afm qcp in cecoin@xmath0 and ceirin@xmath0 .
this closeness to the afm qcp is evaluated from the value of @xmath62 .
it is noteworthy that the value of @xmath1 becomes larger as approaching the afm qcp , since @xmath1 increases as 0.4 k and 2.3 k as @xmath62 decreases as 8 k and 0.6 k for ceirin@xmath0 and cecoin@xmath0 .
this suggests an intimate interrelation between the afm spin fluctuations and the onset of @xmath64-wave superconductivity in these compounds .
et al . _ reported that the value of @xmath1 goes up as the lattice parameter @xmath65 increases @xcite .
in this context , the increase in @xmath65 might enhance the 2d character of afm spin fluctuations leading to the enhancement in @xmath1 . over a long decade
, it is currently proposed that the value of @xmath1 could be enhanced as the dimensionality in electronic structure is reduced from 3d to 2d regime @xcite . consistently with this suggestion , the pressure - induced sc transition temperature in cein@xmath66 , that forms in the 3d cubic structure , is one order smaller than @xmath67 2.2 k in cerhin@xmath0 that forms in the layered structure .
note that , in cein@xmath66 , a @xmath68 = const behavior is observed over a wide @xmath2 range above @xmath1 in a pressure range where the afm order is suppressed and the superconductivity appears @xcite .
it is , therefore , indicated that afm spin fluctuations are absent in this pressure range in the normal state of cein@xmath66 .
by contrast , the afm spin fluctuations in cerhin@xmath0 continue to develop down to @xmath1 , exhibiting a character close to the afm qcp @xcite .
the record high value of @xmath1 in cecoin@xmath0 at @xmath69 = 0 and in cerhin@xmath0 under pressure to date is likely due to the existence of strong afm spin fluctuations that develop in the vicinity of qcp .
next , we move on to the magnetic field effect on the afm spin fluctuations at the close vicinity of qcp in cecoin@xmath0 .
since an energy scale is anticipated to be very reduced around a qcp , it is expected that the criticality at the afm qcp is intimately influenced by applying small magnetic field . as a matter of fact , ybrh@xmath15si@xmath15 and ceru@xmath15si@xmath15 , which are anticipated to be located near a qcp ,
show the strong field dependence of sfs @xcite .
a recent theory also predicts that a strong magnetic field dependence should be observed at the qcp @xcite .
vs @xmath2 plot at @xmath21 t along with the @xmath58 vs @xmath2 plot at @xmath22 t. the solid lines indicate the relations @xmath70 at @xmath71 1.1 t and @xmath72 at @xmath13 = 0 t based on the anisotropic spin - fluctuation s model ( see the text ) . ] in order to address possible @xmath13-induced change in afm - sfs characteristics , we show the @xmath2 dependences of @xmath19 in cecoin@xmath0 and lacoin@xmath0 , in the inset of fig . 1 , that were measured at @xmath13 = 1.1 t parallel to the @xmath26-axis via the co - nmr experiments .
figure 3 indicates the @xmath58 vs @xmath2 plot at @xmath13 = 1.1 t along with the @xmath58 vs @xmath2 plot at @xmath13 = 0 t. the respective data at @xmath71 0 t and 1.1 t are consistent with the anisotropic sfs model with @xmath73 0.6 and 2.5 k as indicated by the solid lines .
interestingly , the application of magnetic field makes the value of @xmath62 increase from 0.6 k to 2.5 k , making it away from the proximity to the afm qcp .
this result reinforces that the cecoin@xmath0 is closely located to the afm qcp .
in conclusion , the measurements of the nuclear spin - lattice relaxation rate @xmath3 by means of the @xmath4in - nqr and @xmath30co - nmr have revealed that the magnetic nature in cecoin@xmath0 is characterized by the strong afm spin fluctuations in the vicinity of qcp . at zero magnetic field ,
the anomalous @xmath2 dependence of @xmath37 is well explained by the anisotropic spin - fluctuations model that predicts the relation @xmath74 with @xmath61 k. noting that this model was applied to ceirin@xmath0 with @xmath75 k , we have suggested that @xmath1 rises in cecoin@xmath0 because of approach to the afm qcp where @xmath76 k. the application of magnetic field makes the value of @xmath62 increase from 0.6 k to 2.5 k , making it away from the qcp .
this result also suggests that cecoin@xmath0 is closely located to the qcp .
we would like to thank k. ishida for valuable comments and discussion .
this work was supported by the coe research grant from mext of japan ( grant no .
10ce2004 ) .
one of the authors ( y.k . ) was supported by the jsps .
00 f. steglich , j. aarts , c. d. bredl , w. lieke , d. meschede , w. franz , and h. schafer : phys .
* 43 * ( 1979 ) 1892 . c. petrovic , p.g .
pagliuso , m.f .
hundley , r. movshovich , j.l .
sarrao , j.d .
thompson , z. fisk , and p. monthoux : j. phys . : condens .
matter bf 13 ( 2001 ) l337 . c. petrovic , r. movshovich , m. jaime , p.g .
pagliuso , m.f .
hundley , j.l .
sarrao , z. fisk , and j.d .
thompson : europhys .
* 84 * ( 2001 ) 354 .
h. hegger , c. petrovic , e.g. moshopoulou , m.f .
hundley , j.l .
sarrao , z. fisk , and j.d .
thompson : phys .
* 84 * ( 2000 ) 4986 .
fisher , f. bouquet , n.e .
phillips , m.f .
hundley , p.g .
pagliuso , j.l .
sarrao , z. fisk , and j.d .
thompson : phys .
b * 65 * ( 2002 ) 224509 .
t. mito , s. kawasaki , g .- q .
zheng , y. kawasaki , k. ishida , y. kitaoka , d. aoki , y. haga , and y. nuki : phys .
b * 63 * ( 2001 ) r220507 .
y. kohori , y. yamato , y. iwamoto , t. kohara , e.d .
bauer , m.b .
maple , and j.l .
sarrao : phys .
b * 64 * ( 2001 ) 134526 .
n.j . curro , b. simovic , p.c .
hammel , p.g .
pagliuso , j.l .
sarrao , j.d .
thompson , and g.b .
martins : phys .
b * 64 * ( 2001 ) 180514(r ) .
zheng , k. tanabe , t. mito , s. kawasaki , y. kitaoka , d. aoki , y. haga , and y. nuki : phys .
* 86 * ( 2001 ) 4664 .
w. higemoto , a. koda , r. kadono , y. kawasaki , y. haga , d. aoki , r. settai , h. shishido , and y. nuki : j. phys .
* 71 * ( 2001 ) 1023 .
ormeno , a. sibley , c.e .
gough , s. sebastian , and i.r .
fishier : phys .
* 88 * ( 2002 ) 47005 .
s. zcan , d.m .
broun , b. morgan , r.k.w .
hasellwimmer , j.r .
waldram , j.l .
sarrao , s. kamal , c.p .
bidinosti , and p.j .
turner , cond - mat/0206069 .
k. izawa , h. yamaguchi , y. matsuda , h. shishido , r. settai , and y. nuki : phys . rev
. lett .
* 87 * ( 2001 ) 57002 .
sidorov , m. nicklas , p.g .
pagliuso , j.l .
sarrao , y. bang , a.v .
balatsky , and j.d .
thompson : phys .
* 89 * ( 2002 ) 157004 .
w. bao , g. aeppli , j.w .
lynn , p.g .
pagliuso , j.l .
sarrao , m.f .
hundley , j.d .
thompson , and z. fisk : phys .
b * 65 * ( 2002 ) 100505(r ) .
h. shishido , r. settai , d. aoki , s. ikeda , h. nakawaki , n. nakamura , t. iizuka , y. inada , k. sugiyama , t. takeuchi , k. kindo , t.c .
kobayashi , y. haga , h. harima , y. aoki , t. namiki , h. sato , and y. nuki : j. phys .
* 71 * ( 2002 ) 162 .
a. ishigaki and t. moriya : j. phys .
jpn . * 65 * ( 1996 ) 3402 . c. lacroix , a.solontsov , and r. ballou : phys .
b * 54 * ( 1996 ) 15178 .
h. kondo : j. phys .
* 71 * ( 2002 ) 3011 .
pagliuso , r. movshovich , a.d .
bianchi , m. nicklas , n.o .
moreno , j.d .
thompson , m.f .
hundley , j.l .
sarrao , and z. fisk : physica b * 312 - 313 * ( 2002 ) 129 .
p. monthoux and g.g .
lonzarich : phys .
b * 59 * ( 1999 ) 14598 .
s. nakamura , t. moriya , and k. ueda : j. phys .
* 65 * ( 1996 ) 4026 .
s. kawasaki , t. mito , y. kawasaki , g .- q .
zheng , y. kitaoka , h. shishido , s. araki , r. settai , and y. nuki : phys .
b * 66 * ( 2002 ) 054521 .
y. kohori , t. kohara , y. yamato , g. tomka , and p.c .
riedi : physica b * 281 - 282 * ( 2000 ) 12 . k. ishida , y. kawasaki , y. kitaoka , k. asayama , h. nakamura , and j. flouquet : phys . rev .
b * 57 * ( 1998 ) r11054 .
k. ishida , k. okamoto , y. kawasaki , y. kitaoka , o. trovarelli , c. geibel , and f. steglich : phys .
lett . * 89 * ( 2002 ) 107202 .
j. rossat - mignod , l.p .
regnault , j.l .
jacoud , c. vettier , p. lejay , j. flouquet , e. walker , d. jaccard , and a. amato : j. of magn .
magn . mater .
* 76 - 77 * ( 1988 ) 376 .
q. si , s. rabello , k. ingersent , and j.l smith : nature * 413 * ( 2001 ) 804 .
|
we report in - nqr and co - nmr experiments of cecoin@xmath0 that undergoes a superconducting transition with a record high @xmath1 = 2.3 k to date among heavy - fermion superconductors . at zero magnetic field ,
an anomalous temperature ( @xmath2 ) dependence of nuclear spin - lattice relaxation rate @xmath3 of @xmath4 in is explained by the relation @xmath5 based on the anisotropic spin - fluctuations model in case of the proximity to an antiferromagnetic ( afm ) quantum critical point ( qcp ) .
the novel behavior of @xmath6 over a wide @xmath2 range of @xmath7 k arises because the staggered susceptibility almost follows the curie law @xmath8 with @xmath9 k and hence @xmath10 for @xmath11 .
we highlight that the behavior @xmath6 is due to the proximity to the anisotropic afm qcp relevant with its layered structure , and is not associated with the afm qcp for isotropic 3d systems .
we have also found that the afm spin fluctuations in cecoin@xmath0 are suppressed by small magnetic field so that @xmath12 k at @xmath13=0 increases to @xmath14 k at @xmath13 = 1.1 t , reinforcing that cecoin@xmath0 is closely located at the qcp .
|
scalar fields play a central role in modern cosmology in driving inflation of the early universe and describing dark energy of the present epoch.@xmath0 observations predict that our universe contains roughly one - third of dark matter and baryons and two- thirds of dark energy.@xmath1 within the standard friedmann - robertson - walker ( frw ) models it was shown that there exist scaling solutions that are the unique late - time attractors whenever they exist.@xmath2 here , in this paper , we wish to look for scaling solutions in higher - dimensional cosmological models . in kaluza - klein theories as well as in brane world scenarios ,
our 4d universe is believed to be embedded in a higher - dimensional manifold .
one of these models is the ricci - flat 5d cosmological solutions presented by liu and wesson.@xmath3 this model is 5d ricci - flat , implying that it is empty viewed from 5d .
however , as is known from the induced matter theory,@xmath4 4d einstein equations with matter could be recovered from 5d equations in apparent vacuum .
this approach is guaranteed by campbell s theorem that any solution of the einstein equations in n - dimensions can be locally embedded in a ricci - flat manifold of ( n+1)-dimensions.@xmath5 in section ii , we suppose the 4d induced matter be composed of a perfect fluid and a scalar field . in section iii , we use phase - plane analysis to study the evolutions of the model . in section
iv we study the scaling solution .
section v is a short discussion .
an exact 5d cosmological solution was given firstly by liu and mashhoon in ref .
11 and restudied by liu and wesson in ref . 7 .
this solution reads@xmath6 -dy^{2 } , \nonumber \\
b&=&\frac{1}{\mu}\frac{\partial a}{\partial t}\equiv \frac{\dot{a}}{\mu } \label{5dmatric } \\
a^{2}&=&(\mu ^{2}+k)y^{2}+2\nu y+\frac{\nu ^{2}+k}{\mu ^{2}+k } , \nonumber\end{aligned}\]]where @xmath7 , @xmath8 and @xmath9 are two integration functions of @xmath10 , @xmath11 is the 3-d curvature index @xmath12 and @xmath13 is a constant .
this solution satisfies the 5d vacuum equation @xmath14 with the three invariants being@xmath15so @xmath13 determines the curvature of the 5d manifold .
the 5d line element in ( [ 5dmatric ] ) contains the 4d one,@xmath16% . \label{c}\]]using this 4d metric
we can calculate the 4d einstein tensor by @xmath17 @xmath18 .
its non - vanishing components are@xmath19generally speaking , the einstein tensor in ( [ d ] ) can give a 4d effective or induced energy - momentum tensor @xmath20 via @xmath21 with @xmath22 .
in previous works,@xmath23 this energy - momentum tensor was modelled by a perfect fluid with density @xmath24 and pressure @xmath25 , plus a variable cosmological term @xmath26 .
this @xmath26 could be served to describe dark energy . in this paper
, we let the energy - momentum tensor consist of two parts : @xmath27 , \nonumber\end{aligned}\]]where @xmath28 represents a perfect fluid and @xmath29 represents a scalar field .
then , from the 4d conservation laws @xmath30 and @xmath31 , we obtain@xmath32 the hubble parameter @xmath33 and the deceleration parameter @xmath34 should be defined in terms of the proper time as.@xmath35 @xmath36using this we can recover the corresponding friedmann equations as follows:@xmath37where@xmath38
from ( [ 5dmatric ] ) we see that in general we have @xmath39 , so @xmath40 is not the proper time .
this leads to the only difference between our results ( [ f])-([i ] ) and those of the standard frw models .
similar as in ref . 5 , we define @xmath41 and @xmath42 as @xmath43 for an exponential potential @xmath44 and a spatially flat universe @xmath45 , we find that the evolution equation for @xmath41 and @xmath42 are of the same form as in ref
. 5 , @xmath46 , \nonumber \\ y^{^{\prime } } & = & -\lambda \sqrt{\frac{3}{2}}x y+\frac{3}{2}y[2x^{2}+\gamma _ { m}(1-x^{2}-y^{2 } ) ] , \label{xandy}\end{aligned}\]]where a prime denotes a derivative with respect to the logarithm of the scale factor , @xmath47 . define @xmath48then ( [ xandy ] ) gives @xmath49 and ( [ j ] ) gives @xmath50 .
meanwhile , from ( [ i ] ) , the effective equation of state for the scalar field gives@xmath51 from the analysis of ref . 5 we know that there are five fixed points ( critical points ) corresponding to @xmath52 and @xmath53 in the plane - autonomous system , including two stable nodes.@xmath54 the two stable nodes represent two possible late - time attractor solutions and , therefore , are of particular physical interest . we list them in the following .
late - time attractor solution dominated by the scalar field alone .
this solution corresponds to the stable node@xmath55for which we have @xmath56 , then we get@xmath57so the universe is dominated by the scalar field alone and expands with a power - law .
it can be shown that if , furthermore , @xmath58 , the universe is accelerating . ( 2 ) . scaling solution .
it corresponds to the stable node@xmath59for which we have@xmath60so this represents a scaling solution where the dark energy density of the scalar field is proportional to that of the baryons and dark matter .
it is known that in the standard spatially - flat frw models the scaling solution ( [ p ] ) corresponds to a power - law expansion . here
let us check how the universe expands in the 5d model ( [ 5dmatric ] ) .
firstly , by substituting the value of @xmath41 of ( [ p ] ) into its definition in ( [ j ] ) and using ( [ decelerate ] ) , we find that the scalar field @xmath61 can be integrated , giving@xmath62thus we obtain the exponential potential as@xmath63meanwhile , substituting the value of @xmath42 of ( [ p ] ) into its definition in ( [ j ] ) , and then using ( [ v ] ) , we get a constraint between @xmath64 and @xmath65 as@xmath66 consider the 5d metric ( [ 5dmatric ] ) .
the form @xmath67 is invariant under an arbitrary coordinate transformation @xmath68 .
this would enable us to choose the coordinate @xmath10 such that although @xmath40 does not equal to the proper time @xmath69 globally , it can tend to the proper time as @xmath70 .
be aware that in the 4d standard frw models the scaling solution ( [ p ] ) corresponds to the expansion @xmath71 and @xmath72 .
thus we obtain the late - time approximation of the 5d scaling solution ( [ p ] ) as:@xmath73for @xmath74 , the 5d solution ( [ 5dmatric ] ) reads@xmath75using ( [ t ] ) in ( [ u ] ) , we find@xmath76
in this paper we have used the phase - plane analysis to study the stability of evolution of a 5d cosmological model . from eqs.([t])-([w ] ) we can draw three conclusions : ( 1 ) .
the scaling solution ( [ p ] ) , as a late - time attractor of the evolution of the universe , is the same in both the 4d frw models and the 5d induced matter theory .
( 2 ) . in the 5d model ( [ t])-([w ] ) , the expansion rate @xmath64 and @xmath33 of ( [ t ] ) seems not rely on the value of @xmath42 . in other words , our 4d universe ,
no matter it is situated on which hypersurface of @xmath42 in the 5d manifold , looks similar and expands with almost the same rate , in late times .
the constant @xmath13 in ( [ u ] ) represents the curvature of the 5d manifold ( see eq.([b ] ) ) .
it is reasonable to believe that @xmath77 may correspond to three different topologies for the 5d manifold ( 5dmatric ) .
however , from ( [ t])-([w ] ) , we find that @xmath13 does not affect @xmath64 and @xmath33 so much in late - times .
( 4 ) . in eq . ( [ t ] ) ,
if @xmath78 , we have @xmath79 and @xmath80 , and our model approaches to the milne model which has no event or particle horizon , and so it is of particular interest .
it was shown@xmath23 that in the early - time of the universe the 5d solution ( [ 5dmatric ] ) deviates from the standard frw models greatly .
for instance it may have a big bounce rather than a big bang and before the bounce the universe is contracting .
meanwhile , the value of the 5d curvature @xmath13 is sensitive for having or not having a bounce .
however , in the late time , we have @xmath81 and the coordinate time @xmath10 gives back to the cosmic time @xmath69 , so the bounce model approaches to the frw model . as a late - time attractor , the scaling solution seems not show up noticeable deviations from the standard frw models . further study along this line
is needed to distinguish the two theories .
this work was supported by national natural science foundation ( 10273004 ) and national basic research program ( 2003cb716300 ) of p. r. china .
99 a. h. guth , _ phys .
_ * d23 , * 347 ( 1981 ) ; a. d. linde , _ phys .
_ * b108 , * 389 ( 1982 ) ; a. d. linde , _ phys .
_ * b129 , * 77 ( 1983 ) .
k. coble , s. dodelson and j. a. frieman , _ phys .
_ * d55 , * 1851 ( 1997 ) , astro - ph/9608122 ; r. r. caldwell and p. j. steinhardt , _ phys .
_ * d57 , * 6057 ( 1998 ) , astro - ph/9710062 ; i. zlatev , l. m. wang and p. j. steinhardt , _ phys .
_ _ lett . _ * 82 , * 896 ( 1999 ) , astro - ph/9807002 ; p. j. steinhardt , l. m. wang and i. zlatev , _ phys . rev . _
* d59 , * 123504 ( 1999 ) , astro - ph/9812313 .
r. j. van den hoogen , a. a. coley and d. wands , _ class .
* 16 , * 1843 ( 1999 ) , gr - qc/9901014 ; i. p. c. heard and d. wands , _ class .
* 19 , * 5435 ( 2002 ) , gr - qc/0206085 ; z. k. guo , y. s. piao , r. g. cai and y. z. zhang , _ phys .
_ * b576 , * 12 ( 2003 ) , hep - th/0306245 ; z. k. guo , y. s. piao and y. z. zhang , _ phys . lett . _
* b568 , * 1 ( 2003 ) , hep - th/0304048
. h. y. liu , _ phys .
_ * b560 , * 149 ( 2003 ) , hep - th/0206198 ; b. l. wang , h. y. liu and l. x. xu , _ mod .
_ * a19 , * 449 ( 2004 ) , ( gr - qc/0304093 ) ; x. l. xin , h. y. liu and b. l. wang , _ chin .
lett . _ * 20 * , 995 ( 2003 ) , ( gr - qc/0304049 ) .
|
a five - dimensional ricci - flat cosmological solution is studied by assuming that the induced 4d matter contains two components : the usual fluid for dark matter as well as baryons and a scalar field with an exponential potential for dark energy . with use of the phase - plane analysis
it is shown that there exist two late - time attractors one of which corresponds to a universe dominated by the scalar field alone and the other is a scaling solution in which the energy density of the scalar field remains proportional to that of the dark matter .
it is furthermore shown that for this 5d scaling solution the universe expands with the same rate as in the 4d frw models and not relies on which 4d hypersurface the universe is located in the 5d manifold .
|
one of the most general and successful concepts in physics is the separation of a physical system into a simpler , controlled subsystem that is interacting with a perturbing subsystem .
a specific example of this method is given by a basic impurity problem , consisting of a noninteracting homogeneous medium and one particle disturbing it .
in the case of a noninteracting fermi gas , this is called fermi - polaron problem@xcite .
this theoretical model can help to map out the phase diagram of a strongly population - imbalanced fermi gas@xcite , where the quasiparticle energy and effective mass serve as input parameters for landau - pomeranchuk hamiltonians@xcite helping to quantify zero temperature phase separation and the ground state energy of different phases .
moreover , the @xmath2 fermi - polaron system was shown to undergo a transition of its own , featuring as possible ground states@xcite a polaronic spin-1/2 quasiparticle and the composite spin-0 molecule , consisting of the impurity and a single bath atom . however
, this transition does not generalize easily to population - imbalanced fermi gases as it might be preempted by phase separation .
two of the most common approaches to the fermi - polaron problem are variational anstze and the use of an approximate diagrammatic technique@xcite .
these methods are able to calculate key quasiparticle properties , e.g. , effective mass or polaron residue and helped to map the transition between polaronic and molecular states .
et al . _ extended@xcite these ideas to the case of a mass - imbalanced fermi - polaron problem
and depicted the ground state phase diagram with respect to polaronic , molecular , trimer@xcite and tetramer@xcite states , where a trimer ( tetramer ) is the bound state of two ( three ) bath particles with the impurity .
concerning other techniques , functional renormalization group@xcite and fixed - node diffusion monte carlo@xcite have been successfully applied .
experimental works include refs . .
another approach was presented by prokofev and svistunov in 2008 , diagrammatic monte carlo@xcite ( diagmc ) .
its key ingredient is the sampling of feynman diagram integrals by a set of ergodic updates linking all topologies and internal variables . by reducing the diagrammatic space
, they managed to reach sufficiently high expansion orders allowing them to extract energies and effective masses in very good agreement with other techniques despite the fermionic sign problem .
recently , the method was used@xcite for the extraction of polaron quasiparticle residues and two - dimensional geometries . up to now , these diagmc implementations have only been applied to the special case of equal masses of impurity and bath atoms .
this is important as such a system can be created by different spin states of a homogeneous atomic gas .
however , also mixtures like @xmath0li@xmath1k are experimentally realizable in ultracold atom systems . in our work , we extend diagmc to the case of arbitrary mass - imbalance and present the dependence of polaron energy and residue on the imbalance ratio . determining the polaronic spectral function for mass - imbalance helps understanding the stability of quasiparticles close to the limit of a heavy impurity .
it features the repulsive polaron@xcite , an excited state with finite lifetime .
we show that two - particle - hole wave - functions remain essentially exact in three dimensions and demonstrate the implications for the mass - imbalanced phase diagram .
we also present results for tan s contact parameter for a mass - balanced polaron system .
this paper is structured as follows : section [ sec : sec2 ] presents the basic fermi - polaron model summarizing the diagrammatic ingredients for imbalanced masses . in sec .
[ sec : sec3 ] , some changes of the diagrammatic routine are proposed in order to increase performance , while sec . [
sec : sec4 ] introduces the enhancements of bold diagmc we use in our code .
section [ sec : sec5 ] will explain a newly developed regrouping technique helping to increase extrapolation speed of the series .
finally , sec .
[ sec : sec6 ] exhibits our results .
a brief conclusion is given in sec .
[ sec : sec7 ] , while the appendices give a detailed derivation of the mass - imbalanced @xmath3 matrix and explain our extrapolation procedure .
the system referred to as a fermi - polaron problem consists of two main ingredients : a noninteracting fermi bath and an impurity resonantly interacting with it .
this can be realized at ultracold temperatures because s - wave scattering between any bath particles is forbidden due to pauli s principle and p - wave scattering is energetically suppressed .
the simplest hamiltonian compatible with these requirements is @xcite @xmath4 @xmath5 labels a field operator for a particle in state @xmath6 and with momentum @xmath7 , @xmath8 is the bare coupling constant and @xmath9 incorporates the energy dispersion . for convenience
, we label the impurity by @xmath10 and bath particles by @xmath11 , although this does not necessarily refer to pure spin states .
@xmath12 is the bath particle fermi momentum , @xmath13 its fermi energy .
we are working in units assuring @xmath14 .
for the diagrammatic series , we employ the imaginary time representation and establish the following particle - hole convention : bath lines with momentum @xmath15 are denoted particles possessing positive time while lines with @xmath16 are called holes and propagate with negative time .
the positive time direction is defined to be from left to right .
the inter - particle interaction can be modeled by a pseudo - potential in case of a zero - range interaction .
it is important , though , that this potential assures the correct two - particle scattering length@xcite @xmath17 .
we introduce the @xmath3 matrix @xmath18 in conventional form@xcite and tabulate it prior to the monte carlo run .
this is tantamount to replacing the bare coupling @xmath8 by the experimentally accessible scattering length @xmath17 .
we refer to appendix [ sec : appa ] for details .
the green s functions are given by @xmath19 where @xmath20 is a tuning parameter used for convergence reasons .
reduced mass @xmath21 and total mass @xmath22 are introduced canonically @xmath23
the first - order diagram is used for normalization purposes .
the ( local ) appearance of this diagram topology as part of the whole diagram will be used to identify reducible diagrams in partially bold diagmc in sec .
[ sec : sec4 ] . ]
this ( unphysical ) second - order diagram connects the first - order fake diagram with higher - order diagrams . ]
the set of updates we use is different from the approach of prokofev - svistunov@xcite . rather than linking different orders by worm diagrams , we prefer to implement these transitions by direct updates .
we propose the following update pairs : * first - to - fake and fake - to - first , * change - fake ( self - inverse ) , * insert - mushroom and remove - mushroom , * insert-@xmath3 matrix and remove-@xmath3 matrix , * reconnect ( self - inverse ) .
a fake diagram is used for normalization purposes and is graphically identical to the first - order@xcite diagram , but with analytically easy weights , cf . fig .
[ fig : fake ] .
its internal variables are updated by the update change - fake .
the updates first - to - fake and fake - to - first connect this fake diagram with the lowest - order diagram we sample , presented in fig .
[ fig : order2 ] . note that this diagram is unphysical if no self - consistent bold scheme@xcite is used . the first - order diagram is not included in our simulation because its contribution experiences a @xmath24-behavior for @xmath25 , thus forcing the program to spend a lot of time on small times .
it is straightforward to include the first - order self - energy by a numerical tabulation in @xmath26 space@xcite .
there are four updates linking different orders : insert - mushroom , insert-@xmath3 matrix and their inverse updates remove - mushroom and remove-@xmath3 matrix .
insert-@xmath3 matrix chooses any @xmath3 matrix of the current diagram and splits it into two linked @xmath3 matrices .
the resulting ( unphysical ) diagram has good overlap with the previous configuration if @xmath27 is artificially attributed as weight of the underlying impurity propagator .
here , @xmath28 denotes the impurity green s function evaluated by plugging the first - order self - energy contribution into dyson s equation .
last , an update called reconnect ensures that all different topologies of a certain order are sampled .
this set of updates is ergodic and avoids sampling of first - order contributions .
the last remaining unphysical diagrams connect two adjacent @xmath3 matrices however , this is important for partially bold sampling ( cf .
section [ sec : sec3 ] ) .
if no self - consistent bold scheme is used , sampling of relevant diagrams can be enforced by assigning an additional penalty weight to those diagrams .
we will present the updates insert - mushroom , remove - mushroom and reconnect in the next subsections .
all other updates were designed in the same spirit .
illustration of inverse updates insert - mushroom and remove - mushroom . ]
this update is available for impurity propagator lines .
it attempts to insert the diagrammatic structure of fig .
[ fig : fake ] ( called mushroom ) on one of these lines .
if the current diagram order is denoted by @xmath29 , there are @xmath30 propagators available for insertion .
having selected one of those with imaginary time @xmath31 and momentum @xmath32 , internal time slices @xmath33 and @xmath34 are uniformly seeded ( probabilities : @xmath35 and @xmath36 ) , as well as a bath propagator momentum @xmath37 with @xmath38 ( probability : @xmath39 ) .
this fixes the time variable of the last piece to @xmath40 .
the whole process is illustrated in fig .
[ fig : insert_mushroom ] .
the metropolis acceptance ratio @xmath41 is @xmath42 the factor @xmath43 in the denominator is part of the diagrammatic weight of the new configuration . @xmath44 and @xmath45 are the probabilities of selecting the updates insert - mushroom or remove - mushroom , respectively .
remove - mushroom is the inverse update for insert - mushroom .
if the current diagram order is denoted by @xmath29 , there are @xmath29 @xmath3 matrices which could be removed together with the corresponding impurity propagator .
however , the first @xmath3 matrix can not be removed , because it is never constructed by insert - mushroom .
the same is true for the last @xmath3 matrix . that leaves @xmath46 possible @xmath3 matrices and balances the selection factors in the metropolis algorithm . being the inverse update of insert - mushroom , the acceptance ratio of remove - mushroom is given by @xmath47 illustration of the first and second case of update reconnect .
the dotted vertical bath propagator line is connected to an arbitrary @xmath3 matrix in the diagram . ]
illustration of the third case of update reconnect .
the dotted and full vertical bath propagator lines are connected to arbitrary @xmath3 matrices in the diagram ( as long as neither the full line of the left figure nor the dotted line of the right figure is connected to itself . ] reconnect is the key update of our procedure .
it randomly selects one of the @xmath3 matrices , except the last one . in diagram order
@xmath29 , there are @xmath30 possible choices .
suppose that a @xmath3 matrix with parameters @xmath48 and an impurity neighbor adjacent to the right with parameters @xmath49 is selected .
the update then proposes to swap the incoming bath propagator with the incoming bath propagator of its right neighbor .
there is a unique way of swapping as we are not working in cyclical representation .
accordingly , the former bath propagator times @xmath33 and @xmath34 are exactly mapped on new times @xmath50 and @xmath51 .
index @xmath52 labels the bath propagator linked to the selected @xmath3 matrix .
note that the mapping of the bath propagator momenta to the corresponding new momenta is not clear at this instant , as the shape of the current topology has to be reflected . in the moment of linking the new propagator configuration ,
the underlying momenta have to be adjusted in a manner described below .
last , the resulting diagram has to be checked for one - particle - irreducibility the update has to be rejected if any impurity propagator line turns out uncovered .
subsequent application of reconnect updates allows to reach every bath propagator configuration and guarantees ergodicity .
more precisely , the update separates into three different cases depending on the current diagram configuration .
the first diagram topology ( cf .
[ fig : reconnect1 ] ) is identified by having a mushroom - structure on the selected @xmath3 matrix its incoming bath line is connected with its outgoing bath line . since swapping will transfer a hole into a bath particle , it is necessary to create its new particle momentum @xmath37 .
this is done by uniform seeding on the interval @xmath53 $ ] for each component of @xmath54 , where @xmath55 introduces the momentum cutoff of our procedure .
the update is rejected if @xmath56 or if @xmath57 . concerning underlying momenta ,
the selected @xmath3 matrix is assigned the momentum @xmath32 of the right neighboring @xmath3 matrix , while its right impurity neighbor obtains @xmath58 .
it is easy to compute final times @xmath59 the acceptance ratio is @xmath60 the second topology is identified by a link between the outgoing end of the selected @xmath3 matrix and its right neighbor .
being the inverse of the latter update , only one more step is necessary . instead of seeding new particle momentum ,
now the hole momentum has to be created on the selected @xmath3 matrix , thus explaining the factor of @xmath12 in eq .
[ eq::db1 ] .
the acceptance ratio for the second topology is @xmath61 all other cases are included in the third topology ( cf .
[ fig : reconnect2 ] ) , defined by neither connecting the selected @xmath3 matrix with its right neighbor nor with itself . such cases are self - inverse .
no seeding is necessary , all bath momenta are purely swapped or added . determining final times
is straightforward @xmath62 note that holes are defined to have negative times .
concerning momenta , only the selected @xmath3 matrix and its right impurity neighbor have to be considered , yielding new momenta @xmath63 for @xmath3 matrix and @xmath64 for impurity line : @xmath65 this results in the acceptance ratio @xmath66
in addition to the modifications of the bare code , we propose some changes only affecting the bold diagrammatic monte carlo routine .
using full green s functions as self - consistent input has the disadvantage of increasing sampling space drastically by forcing the tabulation of the full green s function in both imaginary time and momentum .
it is therefore beneficial to use the fact that first - order contributions dominate the fully bold propagator and construct a partially bold diagrammatic series out of quantities that are easily tabulated .
it is possible to put the analytically known first - order self - energy into dyson s equation in matsubara frequency space to obtain the first - order green s function .
a fourier transform yields the basic propagator @xmath67 without stochastic errors . only a few modifications are necessary to adjust the basic monte carlo routine : first , every diagram containing at least one first - order self - energy diagram ( cf . fig .
[ fig : fake ] ) has to be excluded from measurements .
second , it is no longer forbidden to connect adjacent @xmath3 matrices , just as in fully bold code@xcite the only difference is that the associated impurity weight is now @xmath27 .
we also extended the molecule code to include partially bold propagators .
this extension is almost identical to the bold polaron code . linking @xmath3 matrices
is readmitted with the same ( now non - artificial ) weight @xmath27 for impurity propagators .
the only new feature concerns the first - order molecule diagram that is now sampled and has to be calculated with impurity weight @xmath27 . by this means , the ultraviolet divergence is cured and a second , independent molecule series is constructed which helps confirming robustness and reproducibility of the results .
resummation is still needed for this new series . as a last comment on bold sampling
, we would like to draw attention to the flaws of bold diagmc .
first , if the dyson series is not absolutely convergent , the rearrangement of this series into the fully or partially bold series is potentially harmful , as it constitutes a ( though physically motivated ) regrouping of terms which can yield any result for non - convergent series@xcite .
second , many series in quantum field theory are asymptotic expansions , implying that results begin to get better and better with increasing expansion order until some maximum expansion order @xmath68 is reached , after which the factorial growth of the number of diagrams leads to huge oscillations .
in such a case , using a bare series is a common procedure , whereas the bold approach captures diagrams of higher order from the start . summarizing this line of argument
, a bold diagrammatic approach seems only reasonable if the underlying series is convergent .
( color online ) resummation methods of type riesz with different exponents @xmath69 are compared for unitarity in a three - dimensional setup .
the plot shows polaron energies depending on maximum sampling order . ]
( color online ) resummation methods of type riesz with different exponents @xmath69 applied to the modified bare series ( see eq .
[ eq : resum ] ) are compared at unitarity in a three - dimensional setup .
the plot shows polaron energies depending on maximum sampling order . ]
as diagrammatic expansions are in general not absolutely convergent , an important tool to study the underlying series is resummation@xcite .
this resummation procedure requires a discussion in more detail .
typically we find the molecular energies to be stable , but the polaron energies in the bose - einstein - condensate limit are harder .
sharp resummations are potentially dangerous if the maximum sampling order is not high enough as can be seen as follows : with a strong resummation method , the produced curve is almost flat for low expansion orders and then bends down sharply for higher orders .
weak resummation methods on the other hand have more curvature for low expansion orders and flatten off if the order of divergence of the series is weak enough .
there is thus a risk with strong resummation methods if only low expansion orders can be reached in the sense that a possibly strong curvature is missed , resulting in an apparently converging but wrongly extrapolated result in close vicinity to the first - order result .
this effect is demonstrated@xcite in fig .
[ fig : resummationconventional ] at unitarity .
note that the bare series is monotonously decreasing which makes a high maximum resummation order necessary to extract the correct answer . summing it up : the more sign - alternating the bare series is , the better resummation works .
a typical resummation method is the riesz resummation method .
these resummations will act upon self - energy series which can be written as @xmath70 , where @xmath71 contains all contributions of diagrams of order @xmath29 .
the order of a self - energy diagram is defined as the number of interaction @xmath3 matrices .
the resummed self - energy series @xmath72 for some given maximum order @xmath73 is defined@xcite as @xmath74 where @xmath75 is given by the riesz coefficients @xmath76 @xmath69 fixes the strength of the resummation : for @xmath77 , no resummation is performed at all , while only the first - order contribution is maintained in the limit @xmath78 . if this method is used on molecular series , it might be beneficial to set @xmath79 as this ensures that the first contributing diagram ( which is of second order for molecules ) always contributes with full weight .
we introduce a regrouping technique which seems to saturate much faster .
provided the series is absolutely convergent , this is always allowed .
it is based on a regrouping of terms in the bare series in such a way that sign - alternation is maximized .
the technique consists in splitting @xmath71 , the self - energy contributions of order @xmath29 , into two parts : @xmath80 where @xmath81 collects all diagrams containing at least one @xmath3 matrix linked by a hole to itself , cf .
[ fig : fake ] , and @xmath82 gathers the rest .
we propose a new series @xmath83 which aims to maximize sign - alternation in @xmath84 .
the coefficients in the resummation procedure depend on the expansion order for @xmath85 as @xmath86 these coefficients are in principle arbitrary our choice was designed to show fast saturation as can be seen in fig . [ fig : resummationnewunitarity ] for the case of the fermi - polaron at unitarity : the application of conventional riesz resummations on the reordered series illustrates that the manually implemented oscillations make it possible to read off polaron energies reliably and allow for a clear statement whether the expansion order is high enough or not . when reverting the roles of reducible and irreducible in eq .
[ eq : resum ] , the same answer can be found but only after a stronger resummation .
note that although the sum of diagrams of a specific order of the unitary polaron series is vanishing within error bars , the different terms are not small . as the regrouped series agrees with the results of the standard bare series
, this might indicate that the polaron dyson series is convergent or that the maximum expansion order @xmath68 of its asymptotic expansion is essentially infinity at unitarity .
nevertheless , fig . 7 of ref .
suggests that the series might be asymptotic in fact , as a doubly bold regrouping shows clear signs of growing fluctuations for increasing expansion order after an initial improvement of results .
in this section , our diagrammatic monte carlo results for a mass - imbalanced polaron are presented .
( color online ) polaron energy at unitarity for different mass ratios @xmath87 .
the inset shows the flat part of the figure .
many - body effects get more pronounced for a lighter polaron . ]
( color online ) polaron residue at unitarity for different mass ratios @xmath87 .
the whole range of masses shows a clear difference between the first order result and the full diagrammatic answer .
the inset shows the curve for additional values of @xmath87 . ]
[ fig : energy ] plots the polaron energy at unitarity for different mass ratios @xmath88 .
while the variational chevy ansatz , here labeled as first order , captures the whole curve qualitatively , its quantitative accuracy gets less precise for low @xmath87 , i.e. , a light polaron .
note that this energy curve reproduces the correct infinite mass limit@xcite @xmath89 for an imbalance ratio as low as @xmath90 . for the case of an immobile impurity ,
the polaron is subject to anderson s orthogonality catastrophe@xcite and the quasiparticle description is no longer appropriate . for a light impurity
, the polaron energy decreases rapidly as the effective interaction is stronger for smaller reduced mass ( see appendix [ sec : appa ] ) at unitarity .
the next subsection will show that this effect is weakened for finite scattering length .
eventually , a very light impurity will be subject to relativistic effects so that our description will no longer be appropriate .
the extraction of error bars was based on conservative extrapolations of light riesz resummation with @xmath91 . for details ,
consult appendix [ sec : appb ] .
the quasiparticle residue @xmath92 can be extracted from the asymptotic decay of the full propagator@xcite @xmath93 , where @xmath94 is the ground state energy of the quasiparticle .
this form implies that a suitable monte carlo estimator for the residue is @xmath95 with @xmath96 @xmath97 is the sampled self - energy .
our results for @xmath92 are depicted in fig .
[ fig : residue ] . in this case , the first - order ansatz is both quantitatively and qualitatively different , predicting a different position of the mass - imbalance ratio of maximum residue .
this might indicate that the variational wave - function description works particularly well for energy based quantities , while it might take further particle - hole terms to capture the residue equally well .
therefore the quasiparticle with maximum residue can be found at higher @xmath87 than estimated by first order . at low @xmath87 ,
higher orders affect @xmath92 stronger and stronger , down to a ratio as low as @xmath98 which is sufficient for most mixtures , e.g. , @xmath0li@xmath1k . in this regime
, the differences between the diagrammatic answer and the first order result are most pronounced . for high @xmath87 ,
the diagmc solution yields a roughly constant shift to the first order answer . as a trimer state acquires more and more strength with respect to the polaron state for decreasing @xmath87 and as the molecular state
is strengthened for increasing@xcite @xmath87 , it seems natural that the polaron residue takes on a maximum value in between once it is no longer the ground state , its residue will decay quickly ( although it will not be zero because there is no simple decay channel@xcite ) .
the measured residue is strictly lower than the first order variational result .
this is remarkable as the functional renormalization group analysis of ref .
predicts a higher residue for unitarity at @xmath99 .
further investigation is needed to understand this discrepancy .
no resummation was used to extract the quasiparticle residues .
for @xmath100 , the series seemed to saturate within our maximum expansion order .
the extrapolation error was approximated to be twice the fluctuation of the saturating points .
for @xmath101 , the series changed and the saturation was not visible anymore .
these points are therefore only valid if a linear extrapolation to infinite expansion order is appropriate .
this extrapolation error was approximated by the method explained in appendix [ sec : appb ] .
( color online ) the polaron energy of two values of @xmath87 is plotted for various interaction parameters @xmath102 .
note that the binding energy @xmath103 is subtracted . for strong interactions ,
the light impurity acquires a higher effective dressing relative to the binding energy than the heavy impurity .
this is eventually reversed in the bec - regime . ]
[ fig : energy_coupling ] shows polaron energies at varying coupling strength in the bec - regime for two different mass - imbalance ratios @xmath87 .
both curves experience a peak of maximum dressing around @xmath104 . decreasing the coupling further
, this relative energy is decreased for both masses , although the light impurity is affected more strongly .
eventually , the heavy impurity has a higher effective dressing ( relative to the binding energy ) than the light impurity .
this is a consequence of the @xmath105 term in the denominator of the @xmath3 matrix ( see appendix [ sec : appa ] ) that strengthens the effective interaction between impurity and bath atoms for increasing @xmath87 at a given interaction strength .
note that these curves extend into the molecular sector@xcite where the polaron ceases to be the ground state .
concerning the residue , no qualitative difference could be seen between the @xmath106 and @xmath107 curves .
( color online ) different contributions to the full polaron energy are compared for three maximum expansion orders @xmath29 .
@xmath108 denotes contributions including up to n - particle - hole diagrams .
two of the curves have been offset by @xmath109 for clarity .
the points were measured for a mass - imbalance ratio @xmath110 at unitarity . ] for quasi - two - dimensional geometries , a remarkable quantitative precision of two - particle - hole wave - functions was found for polaron energies@xcite .
a natural question is whether this approach remains valid in three dimensions .
[ fig : hole_plot ] compares different particle - hole channels for three maximum expansion orders . for the selected measurement point ( unitarity with imbalance @xmath98 )
, the three - particle - hole channel contributes with slight quantitative differences , whereas four - particle - hole and five - particle - hole diagrams vanish within error bars .
this confirms the observation that the two - particle - hole result is essentially correct and can be used for quantitative calculations .
the classification of diagrams into particle - hole classes breaks down for the fully bold approach , because each bold diagram captures bare diagrams of different particle - hole order . for the partially bold scheme , this does not apply since the diagrammatic structure of the partially bold green s functions ensures that the propagation will start with zero holes and
is guaranteed to switch to the 1-ph sector at least once .
( color online ) the polaron spectral function is plotted for a mass - imbalance @xmath90 at unitarity .
note the quadratic dispersion and the repulsive polaron at positive energies . ]
it is possible to extract the polaronic spectral function @xmath111 from the sampled self - energy @xmath97 .
the full green s function can be calculated by fourier transform , subsequent application of dyson s equation and another fourier transform .
then , analytic continuation to real quantities is applied@xcite .
[ fig : spectral_function ] presents the spectral function for a mass - imbalance ratio @xmath90 .
although the energy corresponds to the infinite mass limit@xcite , the polaron remains a stable quasiparticle .
the dispersion follows a parabola with positive effective mass , whereas at higher energies , the repulsive polaron ( a metastable eigenstate of the hamiltonian ) can clearly be seen@xcite .
( color online ) molecule energy extrapolation for @xmath112 and @xmath113 .
the maximum expansion order is denoted by @xmath29 .
the solid line corresponds to a linear fit of the last five points .
the crosses mark the one - particle - hole result of ref . .
riesz resummation with exponent @xmath114 was used . ]
( color online ) polaron energy extrapolation for @xmath112 and @xmath113 .
the maximum expansion order is denoted by @xmath29 .
the solid line corresponds to a linear fit of the last five points , the dashed line is a special fitting function explained in appendix [ sec : appb ] .
the one - particle - hole result of ref . is identically with the point @xmath115 .
riesz resummation with exponent @xmath116 was used . ] in this subsection , the extraction of polaron and molecule energies by resummation is compared to variational one - particle - hole wave - functions .
we choose the point @xmath112 and @xmath113 of the mass - imbalanced phase diagram@xcite as it stays away from the peculiarities of unitarity and the trimer threshold .
molecular energies ( fig .
[ fig : massimbmol ] ) yield perfect agreement with the variational ansatz and motivate the quantitative correctness of variational wave - functions for the fermi - polaron problem . for the polaron ( fig .
[ fig : massimbpol ] ) , using a one - particle - hole wave - function underestimates its energy slightly .
hence , it would be beneficial to use at least two - particle - hole precision for the polaron sector for a precise mapping of the phase diagram .
altogether , the phase diagram of ref . can be expected to be nearly quantitatively exact . nevertheless ,
as the polaron phase is underestimated , it will be shifted into the molecular sector . as this will reduce the small size of the nonzero - momentum molecular phase ( labeled as fulde - ferrell - larkin - ovchinnikov phase , fflo ) further , it remains open whether this phase really exists .
we suggest to compute the phase boundary with high precision within a variational 2-ph polaron approach .
tan s contact coefficient@xcite @xmath117 for a strongly population - imbalanced fermi gas is linked to the dimensionless contact coefficient @xmath118 by @xmath119 here , @xmath120 is the fermi momentum of the minority species which is finite for the strongly imbalanced fermi gas .
the dimensionless contact coefficient @xmath118 can be accessed easily by calculating the derivative of the polaron energy with respect to the dimensionless coupling@xcite @xmath121 .
the resulting contact curve agrees with first order calculations within error bars .
our work extends the diagrammatic monte carlo polaron routines to the more general case of a mass - imbalanced polaron .
this is an important limiting case of population imbalanced fermi gases and allows to estimate key properties of its phase diagram .
the diagrammatic space can be drastically reduced by sampling the self - energy instead of the full green s functions and by using the ladder approximation as basic interaction element .
we presented a critical analysis and alternative check of diagrammatic monte carlo as well as a partially bold approach , thus broadening the toolbox of the diagrammatic monte carlo method .
an additional regrouping technique was presented to speed up extrapolation to infinite diagram order for absolutely convergent series . while the first - order variational ansatz could give qualitative and quantitative good results for the polaron energy at different polaron masses , discrepancies are more pronounced for the polaron residue . for this quantity ,
higher orders have to be included in order to capture the whole physics . concerning tan s contact coefficient ,
an excellent agreement was found with the chevy variational wave - function .
the polaronic spectral function was extracted from imaginary time representation of diagrammatic monte carlo data by means of analytic continuation .
it demonstrates a clean parabolic dispersion as well as the existence of the repulsive polaron .
the two - particle - hole wave - function ansatz , which was shown to be essentially exact in quasi - two - dimensional geometries@xcite , provides an equally good description of quasi - particle energies in three dimensions .
therefore , using one - particle - hole trial functions will lead to a phase diagram which overestimates molecular contributions and might lead to a weakening of the fflo state found in ref . .
we are grateful to m. bauer , t. enss , j. levinsen , m. parish , n. prokofev , r. schmidt , b. svistunov , k. van houcke and w. zwerger for valuable discussions .
we would also like to thank m. parish for sharing her data with us .
this work is supported by the excellence cluster nim , fp7/marie - curie grant no
. 321918 ( fdiagmc ) , fp7/erc starting grant
306897 ( qusimgas ) and by a grant from the army research office with funding from darpa .
it is possible to write the @xmath3 matrix in presence of the fermi sea in a self - consistent way@xcite @xmath122 since @xmath3 is replacing an interaction line in first approximation it has to follow the same sign convention as @xmath8 .
also note that for a point interaction , @xmath123 is constant and independent of momentum . rewriting yields @xmath124 the integral
is calculated by residue calculus ; as only the pole of @xmath125 is in the upper plane , it follows @xmath126 the problem of the latter expression is an ultraviolet divergence in @xmath54 , consequently it has to be regularized .
a similar series appears in the standard scattering lippmann - schwinger equation and can be used for regularization .
starting point is the lippmann - schwinger form of the relative motion schrdinger equation @xmath127 imposing @xmath128 on the left and introducing the two - body @xmath3 matrix @xmath129 via @xmath130 , it follows @xmath131 in the next step , @xmath132 is multiplied from the left with @xmath133 and a complete set of eigenfunctions of @xmath134 is inserted @xmath135 the matrix elements of the potential are trivial for a pseudo - potential and the newly inserted @xmath136 is an eigenstate of @xmath134 @xmath137 looking at @xmath138 more closely @xmath139 remarkably , this expression does not depend on the first element . as there is no @xmath140-dependence left , the @xmath3 matrix can be extracted from the integral @xmath141 the two - body @xmath3 matrix is related to the scattering length @xcite if the effective range can be set to zero @xmath142 @xmath143 is related to the energy of relative motion and can be generalized to off - shell behavior @xmath144 note that the total energy @xmath26 is measured with respect to @xmath13 and @xmath145 .
@xmath146 is the total momentum of the system .
inserting this into equation [ ft ] yields @xmath147 finally , a suitable expression for the @xmath3 matrix can be found by inserting equation [ tmatrixlipp ] into equation [ tmatrixdiag ] @xmath148 this can be cast into its final form by a shift @xmath149 in the first integral @xmath150
( color online ) the extrapolation procedure on resummed data is demonstrated for @xmath151 and @xmath152 .
the dotted curve shows the linear part , whereas the dashed line is a fit with @xmath153 . ]
( color online ) the extrapolation procedure on resummed data is demonstrated for @xmath151 and @xmath116 .
the dotted curve shows the linear part , whereas the dashed line is a fit with @xmath153 .
note that the resulting error bar exceeds the corresponding @xmath152 error bar . ] in this appendix , we explain the details of our resummation procedure .
its use is most delicate for cases where the maximum diagram order is small .
therefore , it is illustrative to use a quasi - two - dimensional fermi - polaron series ( characterized by the dimensionless parameter @xmath155 , where @xmath156 is the two - dimensional scattering length and @xmath157 is the two - dimensional fermi momentum ) to explain this technique , since the maximum expansion order is approximately 8 . for these systems
, it is additionally necessary to deal with large binding energies , hence aggressive resummation has to be applied to the bare series in order to be able to extrapolate to infinite expansion order .
however , this tends to conceal the curvature of the series in the first points , leading to an initially flat curve .
our extrapolation procedure is the following : for the upper value of the error bar , we apply linear extrapolation on the riesz - resummed data with riesz exponent @xmath69 . in this linear extrapolation ,
only the two points corresponding to highest and second highest expansion order are taken into account .
for the lower value of the error bar , we _ assume _ a worst - case scenario with large curvature of the extrapolated curve , according to the following fit function @xmath153 of parameters @xmath17 and @xmath158 : @xmath159 @xmath29 denotes the maximum expansion order .
we emphasize that the curvature of this function is empirically set by us .
this curve includes only the highest and second highest expansion points .
an important feature of @xmath153 is the dependence on the riesz exponent .
this ensures that a stronger resummation results in a bigger error bar due to extrapolation errors . the fit @xmath153 can also be used for bare data ( @xmath77 ) . in this case
, we replace @xmath69 by -1 in eq .
[ fitting ] . note that the error bars represent a variability of results due to systematic origins corresponding to the one-@xmath6-interval .
the result of this technique is shown in fig .
[ fig : resummation1 ] . for a maximum expansion order of 7 ,
the two ways of extrapolating are shown in comparison .
if the maximum expansion order was 6 , then this error bar would increase , just as one would expect regarding the loss of information .
[ fig : resummation2 ] uses a sharper resummation on the same data , demonstrating that the error bar increases with @xmath69 .
therefore , it becomes clear that the weakest possible resummation procedure ( among the ones resulting in a monotonously decaying series ) should be applied . as a final example
, we show our resummation for a polaron point inside the quasi - two - dimensional transition region in fig .
[ fig : resummation3 ] . here ,
resummation with @xmath152 is too weak , so @xmath116 has to be used , resulting in a stronger curvature .
it is important to stress that these two extrapolations represent assumed worst - case scenarios .
finally , as always for diagmc , the extrapolation result has to be checked with available experimental or theoretical results , thus justifying its application in retrospect . in our case
, these results would be variational 2-ph results which we expect to be quantitatively exact .
as an example for a system in which extrapolated error bars were underestimated for a similar system , consult fig .
22 of ref . . for molecular energies
, resummation is more straightforward .
as this series is typically alternating , resummed curves can often be extrapolated linearly .
this is demonstrated in fig .
[ fig : resummation4 ] : the last four points are well fit by a straight line
. however , as this resummation involves the same dangers as described above , we try to vary both the fitting ( e.g. , fitting three of the four last points ) and the resummation technique to test the variability of this result .
99 p. massignan , m. zaccanti and g. m. bruun , rep .
* 77 * , 034401 ( 2014 ) .
f. chevy , phys . rev .
a * 74 * , 063628 ( 2006 ) .
s. pilati and s. giorgini , phys .
* 100 * , 030401 ( 2008 ) . c. j. m. mathy , m. m. parish and d. a. huse , phys .
lett . * 106 * , 166404 ( 2011 ) .
n. v. prokofev and b. v. svistunov , phys . rev .
b * 77 * , 020408(r ) ( 2008 ) . n. v. prokofev and b. v. svistunov , phys . rev .
b * 77 * , 125101 ( 2008 ) .
r. combescot , a. recati , c. lobo and f. chevy , phys .
lett . * 98 * , 180402 ( 2007 ) .
r. combescot and s. giraud , phys .
* 101 * , 050404 ( 2008 ) .
m. punk , p. t. dumitrescu and w. zwerger , phys .
a * 80 * , 053605 ( 2009 ) . c. mora and f. chevy , phys .
a * 80 * , 033607 ( 2009 ) .
p. massignan and g. m. bruun , eur .
j. d * 65 * , 83 ( 2011 ) .
c. trefzger and y. castin , europhys
* 101 * , 3006 ( 2013 ) .
r. combescot , s. giraud and x. leyronas , europhys . lett . * 88 * , 60007 ( 2009 ) . j. e. baarsma , j. armaitis , r. a. duine , and h. t. c. stoof , phys
a * 85 * , 033631 ( 2012 ) .
o. i. kartavtsev and a. v. malykh , j. phys .
b * 40 * , 1429 ( 2007 ) .
d. blume , phys .
* 109 * , 230404 ( 2012 ) .
j. levinsen and m. m. parish , phys .
lett . * 110 * , 055304 ( 2013 ) .
r. schmidt and t. enss , phys .
a * 83 * , 063620 ( 2011 ) . c. lobo , a. recati , s. giorgini , s. stringari , phys .
lett . * 97 * , 200403 ( 2006 ) .
a. schirotzek , c. h. wu , a. sommer and m. w. zwierlein , phys . rev . lett . * 102 * , 230402 ( 2009 ) .
s. nascimbne , n. navon , k. j. jiang , l. tarruell , m. teichmann , j. mckeever , f. chevy , and c. salomon , phys .
lett . * 103 * , 170402 ( 2009 ) . c. kohstall , m. zaccanti , m. jag , a. trenkwalder , p. massignan , g. m. bruun , f. schreck and r. grimm , nature ( london ) * 485 * , 615 ( 2012 ) . j. vlietinck , j. ryckebusch and k. van houcke , phys .
b * 87 * , 115133 ( 2013 ) .
j. vlietinck , j. ryckebusch and k. van houcke , phys .
b * 89 * , 085119 ( 2014 ) .
p. kroiss and l. pollet , phys .
b * 90 * , 104510 ( 2014 ) .
x. cui and h. zhai , phys . rev .
a * 81 * , 041602(r ) ( 2010 ) . i. bloch , j. dalibard and w. zwerger , rev .
phys . * 80 * , 885 ( 2008 ) .
m. inguscio , w. ketterle and c. salomon _ ultra - cold fermi gases : varenna 20 - 30 june 2006 _ , ios pressinc , isbn 9781586038465 ( 2007 ) .
we define the order of a polaron ( molecule ) diagram by the number of @xmath3 matrices ( impurity propagators ) .
e. kozik , m. ferrero , a. georges , arxiv:1407.5687 .
we use the same energy estimator as ref . throughout the paper .
p. anderson , phys .
. lett . * 18 * 01049 ( 1967 ) . m. m. parish and j. levinsen , phys .
a * 87 * , 033616 ( 2013 ) .
m. jarrell and j. e. gubernatis , phys .
rep . * 269 * , 133 ( 1996 ) .
b. bauer , l. carr , h. g. evertz , a. feiguin , j. freire , s. fuchs , l. gamper , j. gukelberger , e. gull , s. guertler _ et al .
_ , j. stat .
( 2011 ) p05001 .
this section is taken from p. kroiss , diploma thesis ( 2013 ) and follows refs . in large parts .
h. stoof , k. gubbels and d. dickerscheid _ ultracold quantum fields _
( 1st edition ) , springer , isbn 9781402087639 ( 2009 ) .
j. levinsen , m. m. parish , arxiv:1408.2737 .
|
we apply the diagrammatic monte carlo approach to three - dimensional fermi - polaron systems with mass - imbalance , where an impurity interacts resonantly with a noninteracting fermi sea whose atoms have a different mass .
this method allows to go beyond frequently used variational techniques by stochastically summing all relevant impurity feynman diagrams up to a maximum expansion order limited by the sign problem . polaron energy and quasiparticle residue
can be accurately determined over a broad range of impurity masses .
furthermore , the spectral function of an imbalanced polaron demonstrates the stability of the quasiparticle and allows to locate in addition also the repulsive polaron as an excited state .
the quantitative exactness of two - particle - hole wave - functions is investigated , resulting in a relative lowering of polaronic energies in the mass - imbalance phase diagram .
tan s contact coefficient for the mass - balanced polaron system is found in good agreement with variational methods .
mass - imbalanced systems can be studied experimentally by ultracold atom mixtures like @xmath0li@xmath1k .
|
since the discovery of high-@xmath6 superconductivity in la@xmath4ba@xmath2cuo@xmath3 @xcite , the analysis of many experiments has led to a complex phase diagram of cuprate materials , in part originating from the competition between antiferromagnetic ( af ) and superconducting ( sc ) correlations .
both are induced by the coulomb repulsion on the copper @xmath7-orbitals in the cuo@xmath8 planes . at a critical hole - doping level antiferromagnetism and superconductivity
are in balance , and details of the material or its environment determine which order is realized or whether a regime of coexistence or local phase separation exists . at very low temperatures
most cuprates show a transition from an af to a sc state , if the density of charge carriers exceeds a critical value , and superconductivity vanishes again beyond a higher carrier density , when the coulomb interaction becomes less significant . above the critical hole doping level ,
af correlations on reduced time and length scales persist and fluctuating antiferromagnetism may coexist with cooper pairing .
one prominent phase with a combination of both has stripe order .
an ordered phase with static stripes was first observed in 1994 by tranquada _
et al_. in the nickelate la@xmath8nio@xmath9
@xcite and later on in the rare - earth doped cuprate la@xmath0nd@xmath1sr@xmath2cuo@xmath3 @xcite and in the original high-@xmath6 cuprate la@xmath4ba@xmath2cuo@xmath3 @xcite .
the cuprate systems evolve from a mott insulator into a homogeneous sc phase as @xmath10 is increased , but as @xmath10 approaches @xmath11 , antiferromagnetism returns in a characteristic spin - density - wave ( sdw ) pattern and superconductivity is suppressed @xcite . based on his neutron scattering data ,
tranquada suggested the existence of antiferromagnetically ordered spin ladders separated by metallic lines , thus forming stripes with a width of four lattice constants ( figure [ fig0 ] ) .
the metallic lines form anti - phase domain - wall boundaries between the af stripes , which optimizes the kinetic energy through virtual electronic hopping processes into the af stripes @xcite .
the appearance of static stripe order seems to be tied to the structural transition into the anisotropic low - temperature tetragonal phase ( ltt ) in which at low temperatures superconductivity possibly coexists with stripe order @xcite .
the spin dynamics of such systems , has been explained quite successfully using models of coupled spin ladders @xcite , although their microscopic origin has remained unresolved within this ansatz , and superconductivity was not incorporated . in this framework both , site and bond centered (
i.e. three- or two - legged ) ladders are compatible with the experimental results for the dynamical spin susceptibility , although distinct theoretical considerations suggest that site centered stripes are favored @xcite .
unidirectional charge- and spin - stripe order was indeed predicted for cuprate systems before their experimental discovery . in 1989 ,
zaanen and gunnarson @xcite and machida @xcite found af stripe formation in the mean - field solution of a two - band hubbard model .
after stripes had been discovered experimentally , the @xmath12@xmath13 model was identified as a more appropriate approach towards stripe formation in hole doped cuprates .
extensive numerical studies were performed using exact diagonalization @xcite , dmrg methods @xcite or monte carlo simulations @xcite , all indicating that stripes indeed form in the hole doped @xmath12@xmath13 model .
although these numerically exact results yielded a consistent picture , their range of validity is limited by small system sizes and boundary effects , and the inclusion of superconductivity has only recently become possible @xcite .
meanwhile , new approaches have been developed for a phenomenological characterization of the sc state that possibly coexists with spin- and charge - stripe order .
et al_. suggested a new type of sc state , termed pair density wave ( pdw ) , in which cooper pairs with center - of - mass momenta @xmath14 and @xmath15 coexist and the pair density oscillates with wave vector @xmath14 @xcite .
the latter oscillation is analogous to the fflo solution of a superconductor in magnetic fields presented by fulde and ferrell @xcite and by larkin and ovchinnikov @xcite , and is typically accompanied by a cdw with wave vector @xmath16 .
houzet and buzdin introduced a description of the fflo state on the ginzburg - landau level @xcite , which was later used as a model for the pdw by agterberg and tsunetsugu @xcite and by berg _
et al_. @xcite .
subsequently , the pdw state without antiferromagnetism was shown to be the energetically stable solution for a microscopic nearest - neighbor pairing model with sufficiently strong pairing interaction @xcite .
this state indeed explains qualitatively some of the observed properties of striped cuprates @xcite . to examine the nature of the sc state in coexistence with spin- and charge - stripe order ,
a simple microscopic model is desirable which allows to adequately describe magnetic and sc order simultaneously , merging the strong - coupling ansatz of the @xmath12@xmath13 model with the weak - coupling framework of bcs theory .
work in this direction was recently performed using a gutzwiller renormalized mean - field ansatz for the @xmath12@xmath13 model @xcite .
however , the pdw did not prove to be the groundstate , but rather a modulated @xmath7-wave state without phase shift of the sc order parameter between neighboring stripes was found as the lowest energy solution .
in contrast to the pdw , this modulated @xmath7-wave has a pair density that oscillates around a finite uniform @xmath17 component , and pair and charge densities oscillate with the same wave vector .
[ width=0.74]fig1.pdf in this paper we introduce a modified @xmath12@xmath18 model as an extension of the standard @xmath12@xmath13 model which allows for a straightforward mean - field decoupling .
this method of relaxing the no - double - occupancy constraint and extending the @xmath12@xmath13 model by an additional on - site repulsion was originally introduced by kagan and rice @xcite to demonstrate the possibility of @xmath7-wave superconductivity in the @xmath12@xmath13 model .
this ansatz enables us to study stripe formation and superconductivity in a wide range of parameters , and it merges two previously distinct approaches to theoretically describe cuprates in different regimes of the renormalized exchange coupling @xmath18 . in section [ sec : tj - bcs ] we introduce the @xmath12@xmath18 model and identify two parameter regimes where a mean - field description of magnetic correlations is appropriate for hole - doped systems : @xmath19 @xmath20 . in this limit superconductivity dominates and antiferromagnetism is controlled by the on - site coulomb repulsion rather than by the superexchange interaction . for small @xmath21 the @xmath12@xmath21 model therefore maps onto a simple nearest - neighbor pairing hamiltonian with an additional on - site repulsion @xmath22 ; this model is derived in section [ sec : bcs ] as the @xmath22-model .
@xmath23 @xmath24 . a strong exchange interaction tends to phase separate the system into a hole - free antiferromagnet and a hole - rich part @xcite .
we show that for @xmath25 an intermediate regime exists in which the charge is separated periodically in half - filled af regions and quarter - filled metallic regions . in this regime
the coulomb repulsion is less significant and the @xmath12@xmath21 model approaches the so called @xmath26-model , which was investigated in detail in @xcite in particular with respect to the breakdown of superconductivity in la@xmath0nd@xmath1sr@xmath2cuo@xmath3 near @xmath5 .
a similar regime has also been found in the @xmath12@xmath13 model itself using dmrg calculations @xcite and variational monte carlo simulations @xcite .
section [ sec : bdg ] describes the numerical strategy to obtain striped solutions within the framework of the bogoliubov - de gennes ( bdg ) equations .
although charge- and spin - stripe order coexisting with superconductivity is found for @xmath20 as well as for @xmath24 , there are important qualitative differences in the two regimes .
we provide a detailed comparison for the origin of stripe formation in both regimes in section [ sec : u - v ] .
the striped groundstate solution presented in this section is energetically favored among a considerable number of other possible solutions .
many of these energetically unfavorable solutions exist only in certain temperature ranges .
we therefore present a survey of the temperature dependences of the most regular solutions in section [ sec : t ] . in section [ sec : conc ] we draw conclusions and discuss our findings in relation to striped cuprate materials .
shortly after the discovery of high-@xmath6-cuprates it was agreed upon that the electronic structure of the cuo@xmath8 planes is well described by a three - band hubbard ( emery ) model @xcite , which can be further reduced to a one - band repulsive hubbard model for electron densities above @xmath29 @xcite . apart from numerically exact solutions , two different approaches have commonly been followed to treat the hubbard interaction @xmath30 : @xmath19 in the strong - coupling regime ( @xmath31 ) , the hubbard model is approximated by the @xmath12@xmath13 model with a finite superexchange coupling @xmath13 @xcite .
this effective interaction is non - local and responsible for superconductivity as well as for antiferromagnetism .
@xmath23 alternatively , effective non - local interactions in the finite @xmath22 hubbard model can be computed numerically using , for example , qmc simulations @xcite or determined diagrammatically , e.g. in the rpa approximation .
one thereby obtains a non - local attractive interaction between electrons on nearest - neighbor sites which may serve as a basis for a bcs type mean - field description . in this section ,
we derive mean - field hamiltonians for both of the above approaches , @xmath19 and @xmath23 , and show that they indeed describe the same physics and are formally equivalent .
we start from the @xmath12@xmath13 hamiltonian : @xmath33 where @xmath34 , @xmath35 , and @xmath36 is the projected annihilation operator which enforces the constraint that no lattice site is doubly occupied .
the spin operator is @xmath37 with @xmath38 containing the pauli matrices .
here we use the matrix elements @xmath39 for nearest and next - nearest neighbor hopping with @xmath40 , and @xmath41 denotes all pairs of nearest - neighbor sites @xmath42 and @xmath43 . note that the three - site terms are omitted in the hamiltonian ( [ tj1 ] ) .
the constraint on the site occupation is the main obstacle in solving the @xmath12@xmath13 model and also prevents a straightforward mean - field approximation .
an ansatz to solve this problem is the slave - boson technique ( see @xcite for a review ) . here
we follow alternatively a method by kagan and rice @xcite who introduced an extension of @xmath44 , in which an additional hubbard repulsion @xmath45 replaces the local constraints : @xmath46 in the limit @xmath47 , one recovers @xmath48 from ( [ tj2 ] ) with @xmath49 , if the three - sites terms are omitted . generally , mean - field solutions of @xmath50 overestimate antiferromagnetism as compared to superconductivity at large interaction strength .
therefore more physical results are expected for smaller values of @xmath45 ( c.f . the discussion in section [ sec : u - v ] ) .
the hamiltonian @xmath50 is readily decoupled on the hartree - fock level . with @xmath51 \label{tj3}\end{gathered}\ ] ] the decoupled hamiltonian becomes @xmath52+c_1.\nonumber\end{aligned}\ ] ] with @xmath53 the constant @xmath54 is given by @xmath55{$\displaystyle c_1\!=\!\sum_{\langle i , j\rangle}\!\left[\!\frac{\displaystyle\delta^{\!s*}_{ji}\delta^{\!s}_{ij}}{\hat j}-\frac{{\cal x}^*_{ji{\downarrow}}{\cal x}_{ij{\uparrow}}}{2\hat j}+\frac{\hat j}{2}\bar n_{i{\uparrow}}\bar n_{j{\downarrow}}\!\right]\!-\frac{\hat u}{2}\sum_{i , s}\bar n_{i , s}\bar n_{i ,- s}$}. \label{tj4.1}\end{aligned}\ ]
] @xmath54 is essential for determining the free energy correctly .
the superconducting order parameter @xmath56 represents nearest - neighbor electron pairing in the spin - singlet channel , while possible spin triplet contributions cancel within the hartree - fock decoupling of the @xmath12@xmath18 model .
@xmath57 is the spin resolved thermal average of the local charge density @xmath58 and @xmath59 renormalizes the nearest - neighbor hopping amplitude @xmath12 .
the pairing term in @xmath60 drives superconductivity with a maximum energy gain for an order parameter @xmath56 with @xmath7-wave symmetry , i.e. @xmath61 .
furthermore , @xmath60 contains two terms which favor antiferromagnetic order .
the hubbard term @xmath62 costs the energy @xmath45 for each doubly occupied lattice site and thus polarizes the spin on each site . the kinetic energy in turn is optimized by ordering the spins antiferromagnetically .
the term @xmath63 directly supports an opposite spin alignment on nearest - neighbor sites by the energy @xmath64 per bond , where @xmath65 is the local magnetic moment . for large values of @xmath18
the system approaches perfect af order and the hubbard term looses relevance . on the other hand ,
even if all lattice sites are fully spin polarized by a large @xmath45 , a finite @xmath18 term in ( [ tj4 ] ) leads to af correlations .
the above reasoning justifies to use @xmath66 with a finite @xmath45 instead of the original @xmath12@xmath13 hamiltonian for the following physical reason : in the limit @xmath67 , the half - filled one - band hubbard model decouples into independent localized electrons without any magnetic order . in the @xmath12@xmath13 model , this is reflected by @xmath68 .
if a multi - band hubbard model is mapped onto the @xmath12@xmath13 model , the superexchange terms lead to a finite and sizable @xmath13 and thus to antiferromagnetic order even in the limit of a large intra - orbital @xmath22 @xcite .
antiferromagnetism is in this case controlled by @xmath13 rather than by @xmath22 .
the dominant effective interactions derived from the qmc calculations in @xcite are the on - site repulsion @xmath22 and an attractive interaction @xmath26 for electrons on nearest - neighbor sites .
if only these two dominant interactions are kept , we arrive at the real - space bcs type hamiltonian @xmath71 with @xmath72 and @xmath73 .
the mean - field decoupling of @xmath74 generates the bilinear hamiltonian @xmath75+c_2 \label{tj8}\end{gathered}\ ] ] with @xmath76 and @xmath77-\frac{u}{2}\sum_{i , s}\bar n_{i , s}\bar n_{i ,-
s}. \label{tj8.1}\end{aligned}\ ] ] comparing @xmath50 to @xmath74 reveals that the mean - field decoupled hamiltonians are almost identical , if one identifies the interaction parameters @xmath78 in @xmath60 and @xmath79 , respectively . in @xmath79 , the term @xmath59 is missing , which renormalizes the nearest - neighbor hopping @xmath12 in the @xmath12@xmath18 model .
since it effectively decreases @xmath12 , it tends to stabilize superconductivity ; its qualitative influence on the groundstate solution is however marginal ( c.f .
more relevant are the differences in the definitions of the sc order parameters in @xmath60 and @xmath79 : whereas @xmath56 contains only the spin - singlet channel , @xmath80 includes also the @xmath81 triplet channel . in a non - magnetic system , the triplet component generally vanishes for a nearest - neighbor interaction , because the pair wave - function has even parity . in the presence of antiferromagnetism however ,
a finite triplet admixture appears ( c.f .
@xcite ) . in @xmath79 ,
this @xmath82-triplet has its own order parameter @xmath83 , the spin - triplet component of @xmath80 .
however , @xmath84 remains small compared to the spin - singlet component , as we have verified .
therefore we neglect the triplet component in the following calculations and discussions and use the spin - singlet order parameter @xmath85 alone with the coupling constant @xmath26 [ instead of @xmath18 as in ( [ tj5 ] ) ] .
furthermore , @xmath79 has an additional prefactor @xmath86 in the sc term .
therefore sc order is weighted stronger in @xmath60 than in @xmath79 and consequently @xmath18 has to be larger than @xmath26 in order to stabilize antiferromagnetism ( see the discussion in sec .
[ sec : u ] ) .
notice that the last term @xmath87 in ( [ tj8 ] ) , which appears identically also in @xmath60 , is absent in the classical bcs theory , because in a homogeneous system it only renormalizes the chemical potential @xmath88 used to control the particle number ( see section [ sec : bdg ] ) .
it is however a source of af order in the nearest - neighbor pairing model @xmath74 .
a mean - field approach does not necessarily capture all the physics contained in the underlying model hamiltonians .
the above decoupling scheme lacks e.g. the proper physics of the mott insulator to metal transition .
there are however two specific cases in which the mean - field solutions of @xmath74 or @xmath89 are indeed self - consistent : \1 .
if @xmath90 , the term @xmath91 contributes little to the emergence of antiferromagnetism , which in this case is controlled mainly by @xmath22 . if also @xmath22 is small , superconductivity is the dominating order and the sc gap is larger than the af gap
. therefore the failure of the mean - field theory to describe the quasiparticle peak at the fermi energy of a doped mott insulator is not relevant .
thus , for small @xmath26 , @xmath79 reduces to the @xmath22-model : @xmath92 \label{j11}\end{gathered}\ ] ] discussed in detail in section [ sec : u ] .
if @xmath26 is close to or larger than @xmath12 , the term @xmath93 is the dominant source of antiferromagnetism .
the system separates into nearly half - filled af regions ( stripes ) where superconductivity is suppressed , and into empty or quarter - filled metallic ( or superconducting ) regions , depending on the type of solution .
since commensurate antiferromagnetism appears only in the half - filled regions , a mean - field approach is consistent also in this case , and @xmath79 reduces to the @xmath26-model : @xmath94 \label{j12}\end{gathered}\ ] ] discussed in @xcite and in section [ sec : v ] . within the @xmath26-model , a finite @xmath22-term can be added without significant influence on the groundstate solution , since the @xmath26-term alone is sufficient to drive the af order parameter close to its maximum .
here we briefly summarize the basic steps to solve the bogoliubov - de gennes ( bdg ) equations for @xmath95 and @xmath96 and explain how striped ( or other inhomogeneous ) solutions are identified .
the bogoliubov transformation which diagonalizes the hamiltonians ( [ j11 ] ) and ( [ j12 ] ) , respectively , is given by @xmath97 , \label{j4.1}\\ c_{i{\downarrow}}&=\sum_n\left[u_{ni{\downarrow}}a_{n{\downarrow}}+v^*_{ni{\uparrow}}a^\dag_{n{\uparrow}}\right ] , \label{j4}\end{aligned}\ ] ] where the coefficients @xmath98 and @xmath99 are obtained from the eigenvalue equation @xmath100 with @xmath101 and the corresponding vector @xmath102 .
the sum over @xmath103 in ( [ j4.1 ] ) and ( [ j4 ] ) extends over all positive eigenvalues @xmath104 and the corresponding eigenvectors .
the operators @xmath105 and @xmath106 act on @xmath107 and @xmath102 as @xmath108 where @xmath43 labels all nearest - neighbor sites of @xmath42 and @xmath109 labels all sites for which @xmath110 is finite , i.e. nearest and next - nearest neighbors . using the symmetry of the energy spectrum @xmath104 around @xmath111 , it is sufficient to diagonalize ( [ j5 ] ) for a single spin component , say @xmath112 .
inserting the transformation ( [ j4 ] ) into ( [ tj6 ] ) and ( [ tj9 ] ) leads to the self - consistency equations for the order parameter and the spin resolved densities @xmath113 and @xmath114 : @xmath115{$\displaystyle \delta^{\!s}_{ij}=\frac{v}{2}\sum_n\left[u_{ni{\uparrow}}v^*_{nj{\downarrow}}f(e_n-\mu)+u_{nj{\uparrow}}v^*_{ni{\downarrow}}f(-e_n+\mu)\right],$ } \label{j7}\end{aligned}\ ] ] @xmath116 here we used that @xmath117 equals the number of holes with spin @xmath118 , and @xmath119 denotes the fermi function .
the chemical potential @xmath88 fixes the total particle number @xmath120 to the desired value .
note that in the presence of antiferromagnetism , @xmath56 has a finite spin - triplet component .
the bdg equations ( [ j5 ] ) and ( [ j7 ] [ j9 ] ) are solved iteratively until self - consistency in @xmath56 and @xmath57 is achieved . in each loop @xmath88 is adjusted to keep @xmath120 constant . to obtain a higher momentum and energy resolution
, we use the super cell method to block - diagonalize the eigenvalue problem ( [ j5 ] ) .
this procedure was introduced by wang and macdonald @xcite for magnetic super cells ( for more details , see @xcite and @xcite ) . here the size of a super cell must be chosen commensurate with the wavelength of the striped solution we attempt to find .
the bdg equations typically have more than one solution , into which the self - consistency cycle may converge .
an anticipated solution can usually be selected by choosing appropriate initial values for @xmath56 and @xmath57 ( and for @xmath88 ) , the search for new solutions is however always a demanding task . typically , two types of sc solutions compete in the presence of af stripes : the modulated @xmath7-wave and the pdw solution . both are characterized by an order parameter with local @xmath7-wave symmetry , i.e. @xmath121 has the opposite sign of @xmath122 , but the absolute values are different , if an extended @xmath123-wave component exists . in the pdw solution
, @xmath56 also has opposite signs on neighboring stripes .
such sign changes in @xmath56 or in the antiferromagnetic order parameter @xmath124 do not evolve continuously from a uniform initial state . in order to trace the anti phase striped solutions discussed in section [ sec : u - v ]
, the initial values of @xmath56 and @xmath125 must therefore have the same anti phase stripe pattern with equal wavelength .
[ width=0.635]fig2a.pdf ( 0,7)*(a ) * [ width=0.635]fig2b.pdf ( 0,7)*(b ) * [ width=0.635]fig2c.pdf ( 0,7)*(c ) * + [ width=0.65]fig2d.pdf ( 8,69)af order parameter @xmath125 ( 3,0)*(d ) * [ width=0.65]fig2e.pdf ( 8,69)charge density @xmath58 ( 3,0)*(e ) * [ width=0.65]fig2f.pdf ( 8,69)sc bond order parameter @xmath56 ( 3,0)*(f ) * the groundstate for each set of parameters is found by minimizing the system s free energy @xmath126 over all self - consistent solutions .
this means minimizing @xmath127 over stripe patterns with different wavelength , but also over bond- and site - centered stripes and different types of sc states , i.e. a modulated @xmath7-wave or a pdw state .
three general observations are made : \1 .
for those values of @xmath26 and @xmath22 for which striped solutions exist at all , solutions of the bdg equations exist for all stripe widths that are commensurate with the finite - size lattice , with af stripes separated by an anti phase hole - rich domain wall . \2 . for @xmath10 close to @xmath11
the striped groundstate has a stripe wavelength @xmath128 , where @xmath129 is the lattice constant .
for small values of @xmath26 and @xmath22 , the groundstate is a homogeneous non - magnetic @xmath7-wave superconductor , whereas phase separation occurs , if @xmath26 or @xmath22 exceeds a certain limit .
a sc state of pdw type occurs only , if the pairing interaction @xmath26 is large enough ( @xmath130 ) .
its energy however is _ always _ larger than the energy of a modulated @xmath7-wave state , although the energy difference becomes vanishingly small for large @xmath26 .
this result is in agreement to earlier work using either a gutzwiller approximation @xcite or dmrg calculations @xcite for the @xmath12@xmath13 model .
in contrast , the pure pdw in non - magnetic systems was shown to be the sc groundstate for a sufficiently strong nearest - neighbor pairing interaction @xcite .
a thorough analysis of the groundstate properties in different parameter regimes follows in section [ sec : u - v ] .
all calculations , except for section [ sec : t ] , were performed at temperature @xmath131 . in the grand canonical framework of bcs - type mean - field theories ,
the thermodynamically stable groundstate is defined by the global minimum of the grand canonical potential @xmath132 , where @xmath88 is the chemical potential and the entropy @xmath133 of the system is given by @xmath134 .
\label{j10}\end{aligned}\ ] ] self - consistent solutions of the bdg equations correspond to local minima in @xmath135 . to determine the groundstate or the thermodynamically stable state at finite temperatures , it is therefore necessary to compare different self - consistent solutions of the bdg equations for each chosen set of parameters ( e.g. , with or without sign changes in @xmath56 , site - centered or bond - centered stripe patterns , etc . ) , and to select the solution with the lowest @xmath135 .
this minimization procedure is valid for the grand - canonical ensemble with a fixed chemical potential @xmath88 and a variable particle number @xmath120 .
however , our calculations aim at solutions with fixed @xmath120 by adjusting @xmath88 , although all expectation values are evaluated grand canonically .
therefore the thermodynamically stable state is determined by the minimum of @xmath136 .
it is instructive to compare @xmath135 and @xmath127 for solutions of the bdg equations with fixed @xmath88 or with fixed @xmath120 , respectively . for fixed @xmath120
, one can prepare the system initially in a state where @xmath135 is minimal , but not @xmath127 .
specifically this would be a state where antiferromagnetism is absent . for fixed @xmath120
, @xmath135 has a global minimum without af order , whereas @xmath127 has two degenerate minima with af order . as @xmath127 is not minimal for this state ,
the non - magnetic solution is unstable and the bdg equations will eventually converge into one of the two minima of @xmath127 . by contrast , the roles of @xmath135 and @xmath127 are interchanged , if @xmath88 is fixed and @xmath120 varied .
therefore fixing @xmath120 instead of @xmath88 is numerically equivalent to a legendre transformation from the grand canonical back to the canonical ensemble .
[ width=0.44]fig3a.pdf ( -8,92)*(a ) * [ width=0.44]fig3b.pdf ( -8,92)*(b ) * + [ width=0.44]fig3c.pdf ( -8,92)*(c ) * [ width=0.44]fig3d.pdf ( -8,92)*(d ) * if @xmath90 ( say @xmath137 ) , the energy gain originating from @xmath138 upon ordering antiferromagnetically is smaller than the accompanying cost in kinetic energy .
if also @xmath139 , the groundstate will be a homogeneous , non - magnetic @xmath7-wave superconductor , whereas larger values of @xmath22 induce af order . for moderate values of @xmath22 ,
@xmath7-wave superconductivity is still the dominating order , which coexists with weak antiferromagnetism above a critical value of @xmath22 depending on the pairing interaction strength .
this regime is described by the @xmath22-model given by the simplified mean - field hamiltonian @xmath95 in ( [ j11 ] ) .
the @xmath22-model supplemented with non.magnetic impurity potentials has been widely and successfully used to describe af correlations in disordered cuprate superconductors @xcite or around vortex cores in magnetic fields @xcite ; andersen and hedegrd showed the existence of striped solutions for this model @xcite
. in particular disorder induced antiferromagnetism appears above a critical value of @xmath22 . here
we show that this model also allows for striped groundstate solutions in clean systems for @xmath140 , where the critical value @xmath141 is slightly larger than for anitferromagnetism in disordered systems . for values of @xmath142
, we obtain a homogeneous af solution without superconductivity , i.e. a doped af mott insulator , for which the mean - field treatment is not adequate .
[ width=0.44]fig4a.pdf ( -8,92)*(a ) * [ width=0.44]fig4b.pdf ( -8,92)*(b ) * + [ width=0.44]fig4c.pdf ( -8,92)*(c ) * [ width=0.44]fig4d.pdf ( -8,92)*(d ) * figure [ fig1 ] shows the typical groundstate solution for hole doping @xmath5 , @xmath143 , and @xmath144 .
the local magnetization @xmath65 ( a ) forms the well known site - centered spin - stripe structure with an anti - phase domain wall between the stripes and an overall periodicity of @xmath145 . upon increasing @xmath22 through @xmath146 ,
the maximum polarization @xmath147 abruptly increases to @xmath148@xmath149 , and rises further to @xmath148@xmath150 towards @xmath151 .
thus the system is quite far from the ideal spin - ladder structure with fully polarized af stripes ( in contrast to the @xmath26-model , c.f .
section [ sec : v ] ) .
holes are expelled from the af stripes , leading to a charge density modulation with a relative amplitude of @xmath152@xmath153 [ see figure [ fig1 ] ( b ) and ( e ) ] .
a solution with bond - centered stripes in a two - legged spin - ladder structure also exists , but it is higher in energy for the parameters chosen above .
if however @xmath22 is increased towards @xmath151 , the groundstate changes to the bond - centered spin structure .
the groundstate for @xmath140 is a modulated @xmath7-wave(mdsc ) state .
the @xmath7-wave projection of the sc order parameter @xmath154 [ see figure [ fig1 ] ( c ) ] shows a similar stripe pattern as the charge density . because of the @xmath10 - @xmath155 asymmetry of the striped system , a finite extended @xmath123-wave component @xmath156 is also induced . to fully characterize the sc state , @xmath10 and @xmath155 bond - order parameters are shown in figure [ fig1 ] ( f ) over a cross section of two stripe periods . for comparison
, we also include a solution of the pdw type with a sign change in @xmath157 from one af stripe to the next .
the pdw solution has a reduced sc order parameter as compared to the mdsc state and a slightly higher free energy .
although the maximum af order parameter @xmath125 is larger for the pdw solution , the gain in magnetic energy is not sufficient to compensate for the smaller sc condensation energy caused by the zeros in the sc order parameter in the center of the af stripes [ see figure [ fig1 ] ( f ) ] .
[ width=0.44]fig5a.pdf ( -8,92)*(a ) * [ width=0.44]fig5b.pdf ( -8,92)*(b ) * if superconductivity is completely suppressed ( by setting @xmath158 artificially ) , the modulations in @xmath159 and @xmath58 are barely larger than in the mdsc state and almost identical to the pdw solution [ see figure [ fig1 ] ( d ) and ( e ) ] .
this indicates that the spin polarization is limited mostly by the kinetic energy , which energetically favors a homogeneous charge and spin density , and less limited by the competition with superconductivity .
indeed , antiferromagnetism and superconductivity rather avoid each other , as becomes evident from the momentum - space characterization discussed below . a complementary description of the intertwined sc and af orders , together with the resulting anisotropy of the system , is possible in momentum space .
note that although we measure all length in real space in units of the lattice constant @xmath129 , we will set @xmath160 for the following discussions in momentum space .
figure [ fig3 ] characterizes the groundstate solution of the mdsc state .
the occupation probability @xmath161 in figure [ fig3 ] ( a ) and the pair density @xmath162 in figure [ fig3 ] ( b ) are almost identical to the known distributions of a pure @xmath7-wave state without magnetism ( c.f .
@xcite ) , except for a small @xmath10-@xmath155 asymmetry .
the pair density squared @xmath163 \label{pd}\end{gathered}\ ] ] measures the average correlated occupation of the electron states with momenta @xmath164 and @xmath165 .
thus @xmath166 is largest at those momenta in the brillouin zone where electron pairs predominately form . within the mean - field decoupling scheme ,
equation ( [ pd ] ) reduces to @xmath167 .
the pair density is concentrated around the fermi surface of the normal state and vanishes at four nodal points near the zone diagonal . around these nodal points ,
af correlations emerge .
they are quantified by the spin order in momentum space , that is by @xmath168 [ see figure [ fig3 ] ( c ) ] .
@xmath169 indeed has its maxima at the nodal points , reflecting the competition between , or rather avoidance of superconductivity and antiferromagnetism , i.e. antiferromagnetism is strong where pair formation is weak . @xmath170
[ see figure [ fig3 ] ( d ) ] is the analog for the charge order , and @xmath171 is non zero only in the presence of charge modulations .
it is instructive to compare figure [ fig3 ] to the corresponding results for the pdw solution in figure [ fig4 ] or to a state where superconductivity is suppressed artificially ( figure [ fig4.0 ] ) .
if superconductivity is absent , the spin polarization is slightly stronger in the af stripes which leads to the characteristic occupation probability @xmath172 , shown in figure [ fig4.0 ] ( a ) : most of the fermi surface is gapped by the af order with a continuous @xmath172 across , but discontinuities in @xmath172 remain on the border of a horizontal bar between @xmath173 and @xmath174 .
these discontinuities constitute disconnected fermi surface arcs which are clearly visible in the spectral function @xmath175 shown in figure [ fig4.0 ] ( b ) . in the superconducting pdw solution ( see figure [ fig4 ] ) , antiferromagnetism is almost as strong as in the absence of superconductivity [ c.f .
figure [ fig1 ] ( d ) ] , thus @xmath172 , shown in figure [ fig4 ] ( a ) , has the same structure as in figure [ fig4.0 ] ( a ) , but the entire fermi surface is now gapped due to superconductivity .
the pair density @xmath162 of the pdw state shown in figure [ fig4 ] ( b ) is strongly anisotropic because of the line zeros of the order parameter enforced by the @xmath82 phase shift between neighboring stripes .
@xmath162 is similar to the pure pdw without antiferromagntism described in @xcite , the latter solution is however not gapless .
the pairing free pieces of the reconstructed fermi surface of the pure pdw @xcite are now gapped due to antiferromagnetism and extend towards the @xmath176 points where superconductivity is weak for vertically oriented stripes [ see figure [ fig4 ] ( c ) ] .
although the pairing contributions around @xmath176 are weak compared to the vicinity of @xmath177 , they are essential for a globally phase coherent sc state .
[ width=0.8]fig6a.pdf ( -0,59)*(a ) * + [ width=0.8]fig6b.pdf ( -0,59)*(b ) * it is instructive to analyze also charge and pair correlations in real space . the effective hopping amplitude @xmath178 for hopping from site @xmath179 to site @xmath42 is obtained by fourier transforming @xmath180 .
the pairing amplitude in real space is @xmath181 .
figure [ fig4.5 ] shows @xmath182 and @xmath183 for the @xmath22-model .
both quantities exhibit an oscillating behavior in which @xmath182 decays over the distance of the sc coherence length @xmath184 , which is roughly 10 lattice constants for the chosen parameters .
the oscillations in @xmath182 , as well as the friedel type of oscillations in @xmath183 , extend over almost the same length in @xmath10 and @xmath155 direction , which verifies the 2d character of the solution of the @xmath22-model .
[ width=0.8]fig7a.pdf ( -0,59)*(a ) * + [ width=0.8]fig7b.pdf ( -0,59)*(b ) * the dominance of the sc order in the @xmath22-model is most apparent in the local density of states ( ldos ) on magnetic and non - magnetic lattice sites as shown in figure [ fig5 ] for the same parameters as above . for the modulated @xmath7-wave solution
[ see figure [ fig5 ] ( a ) ] the ldos vanishes linearly upon approaching the fermi energy , which is the characteristic of @xmath7-wave superconductivity , but it has a slightly asymmetric shape , caused by the particle - hole asymmetry induced by the presence of a charge density modulation .
the ldos of the pdw solution [ figure [ fig5 ] ( b ) ] has a smaller but @xmath123-wave like gap , since it has a strong extended @xmath123-wave contribution . in both solutions
the gap scales with the sc order parameter , and its site independence is characteristic for the @xmath22-model . if @xmath22 is further increased to values above @xmath151 , the magnetism induced gap starts to dominate over the sc gap and superconductivity breaks down .
the @xmath22-model therefore does only allow for sc solutions with superconductivity dominating over af order .
this behavior is typical for the @xmath22-model and quite different to the @xmath26-model , as we show below .
[ width=0.65]fig8a.pdf ( 8,69)af order parameter @xmath125 ( 3,0)*(a ) * [ width=0.65]fig8b.pdf ( 8,69)charge density @xmath58 ( 3,0)*(b ) * [ width=0.65]fig8c.pdf ( 8,69)sc bond order parameter @xmath56 ( 3,0)*(c ) * a detailed description of the @xmath26-model @xmath96 at @xmath5 was given in @xcite .
here we extend this analysis to clarify its relation to the @xmath22-model and the characteristic differences . in the @xmath26-model , the interaction strength @xmath26 controls both the sc and the af correlations . upon turning on and increasing @xmath26
, we find a similar sequence of phases as in the @xmath22-model with increasing @xmath22 at @xmath5 : for small @xmath26 , the groundstate is a homogeneous , non - magnetic @xmath7-wave superconductor . at a critical value @xmath185 , af stripe order with wavelength @xmath145 sets in , similar to the solution of the @xmath22 model ; this solution is characterized in real space in figure [ fig6 ] for @xmath143 at @xmath5 . above a second critical value @xmath186
, superconductivity disappears and the system becomes susceptible to phase separation .
figure [ fig6 ] shows that the af stripes are almost half filled with a staggered spin polarization that is close to maximal , whereas superconductivity vanishes in the center of the af stripes for both the @xmath10 and @xmath155 bonds .
the system can therefore be described as locally phase separated into half - filled af stripes and 1d metallic lines acting as anti phase domain walls .
the crossover from the homogeneous @xmath7-wave superconductor to the striped solution with amplitudes as in figure [ fig6 ] is rather sharp .
[ width=0.44]fig9a.pdf ( -8,92)*(a ) * [ width=0.44]fig9b.pdf ( -8,92)*(b ) * the 1d character of the @xmath26-model solution is most apparent in momentum space , if superconductivity is again artificially suppressed .
the momentum distribution @xmath161 [ see figure [ fig7 ] ( a ) ] consists of a horizontal bar with occupied states for @xmath187 , and a diffuse cloud of occupied states in the center of the brillouin zone far below @xmath188 . the states with momenta inside the bar
arise from the quasi 1d metallic lines between the af stripes , whereas the cloud near the zone center originates from the af stripes themselves .
both structures are tied to separate energy windows in the ldos [ c.f .
figure [ fig8 ] ( a ) ] .
the borders of the bar form two 1d fermi surfaces at @xmath189 , visible in the spectral function @xmath190 in figure [ fig7 ] ( b ) .
pairing can occur exclusively around these two fermi surfaces and is therefore restricted to the metallic hole - rich lines . as for the @xmath22-model
, we also calculated the pairing and effective hopping amplitudes @xmath182 and @xmath183 in real space , shown in figure [ fig7.5 ] .
the pairing amplitude @xmath182 exhibits a sc correlation length which is highly anisotropic with @xmath191 .
the cooper pair motion is therefore restricted to 1d channels along the stripes and consequently the phases of the sc order parameters on adjacent non - magnetic lines become decoupled . in other words , the 2d phase coherent state is degenerate with uncorrelated 1d superconducting stripes .
a globally phase coherent state may still be favorable by a josephson coupling of the 1d sc stripes , however on the hartree - fock level josephson coupling is not included . since charge transport in the @xmath10-direction transverse to the stripes is absent , also the charge correlations do not extend from one stripe to the next , as is visible from @xmath183 in figure [ fig7.5 ] ( b ) .
the dimensionality of the sc state therefore constitutes a qualitative disparity between the @xmath22- and @xmath26-model . the formation of energetically separated states on metallic ( or sc ) and af sites is the essential distinction between the @xmath26 model and the @xmath22 model .
the ldos of the non - sc system in figure [ fig8 ] ( a ) shows a broad metallic band on the metallic sites around @xmath188 , whereas on the af sites all states are far from @xmath188 .
a sc gap therefore appears only on the non - magnetic sites [ see figure [ fig8 ] ( b ) ] . if @xmath26 is lowered below @xmath192 , the magnetism induced gap , which reaches from @xmath193 to @xmath12 in figure [ fig8 ] , becomes smaller than the sc gap , as in the @xmath22-model .
here however magnetism vanishes below @xmath192 .
the @xmath26-model behaves therefore contrarily to the @xmath22-model , where superconductivity is the dominant order until it vanishes when the af gap becomes too large .
[ width=0.8]fig10a.pdf ( -0,59)*(a ) * + [ width=0.8]fig10b.pdf ( -0,59)*(b ) *
in section [ sec : bdg ] we mentioned already the existence of different self - consistent solutions of the bdg equations in both the @xmath22- and the @xmath26-model .
they correspond to different local minima in the parameter space of the free energy @xmath127 above the global minimum representing the states described above .
some of these additional solutions can only be obtained in special parameter ranges , or within a specific temperature regime . in this section
we present the temperature evolution of a selection of these states . the list of states presented here is incomplete , but contains the most stable and regular patterns that are obtainable in both the @xmath22- and the @xmath26-model whereas in the @xmath26-model the contribution to the free energy @xmath127 from magnetism is by far larger than the condensation energy of superconductivity , in the @xmath22-model both contributions are of similar size in the range of @xmath22-values where superconductivity and af coexist .
therefore local minima in @xmath127 with distinct magnetic structures are much closer in energy and thus a large variety of solutions is obtained . on the other hand , in the @xmath26-model
the dominant af order allows only for a few self - consistent solutions .
[ width=0.8]fig11a.pdf ( -0,59)*(a ) * + [ width=0.8]fig11b.pdf ( -0,59)*(b ) * figure [ fig9 ] shows the maximum values of the sc order parameter @xmath56 ( a ) , the af order parameter @xmath125 ( b ) , and the maximum amplitude of the cdw @xmath194 ( c ) as a function of temperature for the @xmath22-model for @xmath5 . for better characterization of the various solutions of the @xmath22-model
, we present the corresponding real - space patterns of @xmath159 , @xmath58 and @xmath195 in the upper panel . for the @xmath22-model we find solutions with stripe order of wavelength @xmath128 and @xmath196 , and also homogeneous solutions .
the striped solutions exist both in the pdw and the mdsc type , whereas homogeneous antiferromagnetism may coexist with homogeneous @xmath7-wave superconductivity ( dsc ) .
some solutions disappear abruptly at certain temperatures . more specifically , superconductivity vanishes at temperatures where antiferromagnetism still persists .
it is therefore instructive to trace the high - temperature states without superconductivity back to lower temperatures .
these special solutions are indicated by white fields in the upper panel of figure [ fig9 ] where the corresponding order ( superconductivity or antiferromagnetism ) is suppressed .
[ width=2]fig12x.pdf ( 34,32)*@xmath69-model with @xmath197 and @xmath198 * ( 18.7,28.5)[1]@xmath128 ( 6.4,26)[1]af + pdw ( 16.5,26)[1]af + mdsc ( 29.8,26)[1]af ( 50.5,28.5)[1]@xmath196 ( 38.0,26)[1]af + mdsc ( 50.5,26)[1]pdw ( 61.8,26)[1]af ( 79.8,28.5)[1]homogeneous ( 70.5,26)[1]af + dsc ( 83,26)[1]dsc ( 93.6,26)[1]af [ width=0.65]fig12a.pdf ( 3,70)*(a ) @xmath199 * [ width=0.65]fig12b.pdf ( 3,70)*(b ) @xmath200 * [ width=0.65]fig12c.pdf ( 3,70)*(c ) @xmath201 * within the @xmath22-model , the order for the onset of superconductivity and antiferromagnetism is not fixed .
close to @xmath146 , superconductivity sets in at a higher temperature than antiferromagnetism , while for the parameters used in figure [ fig9 ] , antiferromagnetism survives to slightly higher temperatures . in this regime
, the sudden disappearance of superconductivity is characteristic for the @xmath22-model , because , as discussed above , superconductivity breaks down , if the sc gap becomes smaller than the af gap as the temperature is raised .
this phenomenon is visible for the @xmath128 af + pdw and af + mdsc solutions as well as for the @xmath196 af + mdsc solution . as the mdsc state vanishes with raising the temperature , the af stripe order changes spontaneously from a site- to a bond - centered pattern .
a remarkable observation is the disappearance of antiferromagnetism in the @xmath196 pdw solution .
this state is stable only in a certain temperature window , while for lower temperatures it decays into an af tartan pattern . for the @xmath26-model
the magnetic order parameter is dominant and at @xmath5 only solutions with wavelengths of @xmath145 or @xmath202 can be stabilized .
the temperature dependence of these solutions is shown in figure [ fig10 ] . here
we also find non - superconducting states with af stripes ( @xmath128 sdw and @xmath203 sdw ) as well as the striped superconducting state with wavelength @xmath128 ( @xmath128 sdw+pdw ) ; note that the phase difference of @xmath80 on neighboring stripes is not uniquely determined in the mean - field solution of the @xmath26-model as discussed in detail in @xcite .
since the non - magnetic sites in the @xmath203 sdw state are nearly empty , superconductivity is absent in this solution .
a @xmath196 sdw state can be stabilized for higher doping levels , e.g. @xmath204 . within the @xmath26-model ,
the superconducting @xmath6 is typically much smaller than the onset temperature of antiferromagnetism , and the two temperatures become equal , if @xmath26 is lowered to @xmath192 .
the @xmath128 sdw+pdw solution merges smoothly into the non - sc @xmath128 sdw solution above @xmath6 , since @xmath125 and @xmath194 are almost temperature independent at @xmath6 .
[ width=0.65]fig13a.pdf ( 128,80)*@xmath70-model with @xmath198 * ( 3,70)*(a ) @xmath199 * [ width=0.65]fig13b.pdf ( 3,70)*(b ) @xmath200 * [ width=0.65]fig13c.pdf ( 3,70)*(c ) @xmath201 * the sequence of phase transitions with decreasing temperature is complicated in the striped cuprate materials .
it is best investigated in la@xmath205ba@xmath206cuo@xmath3 @xcite , where a cdw sets in at @xmath207@xmath208k at the transition into the ltt phase , which is joined by an af stripe order at a slightly lower temperature .
superconductivity appears only below 10@xmath208k , which suggests the dominance of af order as in the @xmath26-model .
a direct comparison to our mean - field models is however difficult , since these models do not describe fluctuating superconductivity , e.g. non phase - coherent stripes , which is likely to be present at much higher temperature close to @xmath209 .
in this article , we have explored how far a mean - field approach of the @xmath12@xmath13 model and of closely related models can assist our understanding of static spin- and charge - stripe order observed in underdoped cuprates in coexistence with superconductivity . for hole doping @xmath5
the free energy is minimized by an sdw order with wavelength @xmath128 with concomitant cdw order with @xmath210 , which agrees with neutron and x - ray scattering experiments for , e.g. , la@xmath205ba@xmath206cuo@xmath3 or la@xmath211nd@xmath1sr@xmath206cuo@xmath3 .
this type of order is stable for a wide range of the interaction parameter @xmath18 in the @xmath12@xmath18 model and in both its descendants for small and large @xmath18 , the @xmath22-model and the @xmath26-model , respectively .
these latter models represent two limiting cases where a mean - field treatment of the @xmath12@xmath18 model is well justified .
while the energetically favored spin structure is a robust property of the @xmath212 filled @xmath12@xmath18 model , the nature of the sc state coexisting with the spin order changes between different limits . for small values of @xmath18 , an almost homogeneous and isotropic @xmath7-wave sc state
is realized which recedes to one dimensional non - magnetic lines for larger @xmath18 and also in the @xmath26-model . for larger @xmath18 ,
phase coherence perpendicular to the stripes is lost or very fragile , which naturally explains the breakdown of superconductivity in the presence of static spin - stripe order , stabilized by the anisotropic hopping amplitudes in the ltt phase of rare earth doped 214 cuprates @xcite .
if phase coherence is established with decreasing @xmath18 , the favored sc order parameter does not acquire a phase shift between neighboring stripes and therefore has a finite homogeneous @xmath213 component .
it is therefore not of the pure pdw type as suggested in @xcite , but rather similar to the state found in @xcite .
a definitive answer to the phase relation of the superconducting order parameter on neighboring stripes can not be given within the presented mean - field description .
a phase sensitive josephson - coupling term is needed , in the spirit of the phenomenological theory of @xcite .
the analysis of the @xmath12@xmath18 model led us to two qualitatively distinct regimes for the coupling constant @xmath18 .
its appropriate value for high @xmath6 cuprates is generally believed to be in the range of @xmath214 @xmath215 .
the physics of the cuprates therefore falls somewhere in between the two limiting regimes where the two mean - field approaches are reliably applicable .
while the two regimes are not continuously connected on the mean - field level , the model systems may well show a smooth crossover and features of both regimes .
for example , arpes measurements on la@xmath205ba@xmath206cuo@xmath3 @xcite show four - fold symmetric fermi - arc like structures at zero energy , compatible with the almost isotropic @xmath7-wave solution , whereas la@xmath211nd@xmath1sr@xmath206cuo@xmath3 with @xmath216 , which has a much stronger anisotropy in the hopping amplitudes along the @xmath10- and the @xmath155-direction , shows a quasi one dimensional fermi surface with a breach at the brillouin - zone center @xcite .
these features fit well to the solution of the @xmath26-model in @xcite .
a related issue is the influence of disorder and vortices on the phenomenon of stripe formation .
our initial calculations for disordered systems indicate that impurities act as pinning centers for af stripes by slowing down spin fluctuations and thereby freezing static stripe order . in the @xmath22-model ,
the stripe structures flexibly adjust in wavy forms to the impurity pattern , while superconductivity seems only slightly affected .
on the other hand , impurities with a moderate scattering potential do not disorder the stripe pattern in the @xmath26-model at all , although they are detrimental for superconductivity .
since stripes in the cuprates are not perfectly ordered , the @xmath26-model limit is probably not reached , but rather an intermediate regime . in the presence of antiferromagnetic stripes ,
the motion of vortices is confined to the stripe direction .
their stable position is at the center of an antiferromagnetic stripe , where superconductivity is weakest .
as has been shown within the @xmath22-model for homogeneous @xmath7-wave superconductors @xcite , antiferromagnetism is enhanced around the vortex core , which , similarly to a strong impurity potential , forces the neighboring superconducting stripes to bend around the vortex .
|
even before the experimental discovery of spin- and charge - stripe order in la@xmath0nd@xmath1sr@xmath2cuo@xmath3 and la@xmath4ba@xmath2cuo@xmath3 at @xmath5 , stripe formation was predicted from theoretical considerations .
nevertheless , a consistent description of the complex coexistence of stripe order with superconductivity has remained a challenge .
here we introduce a hartree - fock decoupling scheme which unifies previous approaches and allows for a detailed analysis of the competition between antiferromagnetism and superconductivity in real and momentum space .
we identify two distinct parameter regimes , where spin - stripe order coexists with either one- or two - dimensional superconductivity ; experiments on different striped cuprates are either compatible with the former or the latter regime .
we argue that the cuprates at @xmath5 fall into an intermediate coupling regime with a crossover to long - range phase coherence between individual superconducting stripes .
|
among numerous phases of two - dimensional electron gases ( 2deg ) , the even - denominator quantum hall states with the filling factors @xcite @xmath6 and @xmath7 are particularly interesting .
in contrast to odd - denominator fractional quantum hall ( fqh ) liquids , they can not be explained by a straightforward generalization of the laughlin variational wave function . an early attempt to understand their nature led to the beautiful idea of non - abelian states of matter @xcite . in non - abelian systems , the types and positions of quasiparticles do not uniquely determine the quantum state .
this results in unusual physics and may open a road to topological quantum computing @xcite . however , the existence of non - abelian quasiparticles has not been proven and the nature of the 5/2 state remains a puzzle . both abelian and non - abelian
candidate states were proposed as possible theoretical explanations of the 5/2 fqh effect @xcite . a number of methods @xcite were invented and several experiments @xcite were performed in an attempt to determine the right ground state .
one approach @xcite consists in the measurement of the tunneling current through a quantum point contact ( qpc ) between the edges of a 5/2 fqh liquid .
the low - temperature conductance exhibits a power - law behavior @xmath8 , where the exponent @xmath9 depends on the topological order in the bulk @xcite .
the exponent @xmath9 was measured in two recent experiments @xcite .
the results of the earlier experiment @xcite were interpreted as supporting the non - abelian anti - pfaffian or @xmath3 states .
the best fit for the second experiment @xcite comes from the abelian 331 state .
we argue that the discrepancy between those results can be explained by different geometries of point contacts in refs . and . the geometry of the device from ref .
is shown schematically in fig .
1b and 1c illustrate the two geometries from ref . .
in all cases , transport is supported by fqh edge channels defined by top gates . in fig .
1a and for the upper gate in fig .
1b , the gate widths are comparable to their distances from 2deg .
the gates are considerably wider in fig . 1c .
thus , the distance between the edge points , connected by the dotted lines in figs . 1a and 1b , is shorter than the screening length due to the metallic gates . in other words ,
the repulsive coulomb interaction between the fqh edge segments on the opposite sides of qpc is not fully screened by the gates in the geometries fig .
1a , b but is screened completely in the geometry of fig .
it is well known that repulsive interaction suppresses tunneling in quantum wires @xcite .
similar physics has been addressed in the context of edge transport in laughlin fqh states in refs . , .
we find that in our case the coulomb interaction drives the tunneling exponents @xmath10 and @xmath11 in the geometries 1a and 1b above the exponent @xmath12 in the geometry 1c .
only the latter assumes a universal value while the former depend on the strength of the interaction across the gates .
hence , information about the nature of the 5/2 state can be extracted from a comparison of the experimental @xmath12 ( ref . ) but not @xmath10 and @xmath11 with universal theoretical predictions in various proposed phases .
such predictions are available for the abelian @xmath5 and spin - polarized and unpolarized @xmath4 states@xcite , the non - abelian @xmath3 and pfaffian states @xcite and the particle - hole conjugate of the pfaffian state , the anti - pfaffian state @xcite . since ref .
provides support for the existence of contra - propagating edge channels in the 5/2 fqh liquid , it is particularly interesting to consider candidate states with upstream edge modes . in the above list
, only the anti - pfaffian state has such property .
the particle - hole conjugates of the other proposed states also exhibit upstream edge transport and we compute @xmath12 in them below .
the comparison with the experiment shows that the halperin 331 state agrees best with the tunneling data @xcite .
this paper is organized as follows .
we review theoretically proposed @xmath13 fqh states in section ii .
three appendices complement that section . in appendix
a we discuss the edge theory of the anti-331 state which can be understood as the bonderson - slingerland state @xcite built on the particle - hole conjugate of the laughlin 1/3 state .
appendix b addresses the anti-@xmath3 state . in appendix c , we argue that the anti-@xmath5 state does not exhibit a universal tunneling exponent .
effects of the coulomb interaction across the gate are discussed qualitatively in section iii .
detailed calculations are given in section iv and appendix d. in the final section we compare our results and the experimental findings @xcite with the information obtained from other types of experiments .
here we review several simplest topological orders , proposed for the @xmath6 state , and address their edge properties which we will need in the subsequent sections . in all candidate states ,
the lowest quasiparticle charge is @xmath14 , where @xmath15 is the electron charge .
this is rather different from the odd denominator states , where we usually expect that the lowest quasiparticle charge of the @xmath16 state is @xmath17 .
the @xmath14 charge follows a general rule @xcite for even denominator states : the ratio of the quantized hall conductance to the lowest possible charge must be @xmath18 , where @xmath19 is an integer .
the @xmath14 value of the charge agrees with the experiments @xcite on the 5/2 fqh liquids .
the best fits for the quasiparticle charge in the tunneling experiments @xcite are @xmath20 in the geometry 1a , @xmath21 in the geometry 1b and @xmath22 in the geometry 1c . in all cases
the nearest possible excitation charge is @xmath14 .
thus , we conclude that in all geometries the tunneling current is carried by @xmath14-quasiparticles .
we start our overview with the simplest abelian topological orders . to simplify our notation we set @xmath23 to 1 .
we assume that electrons form pairs in the second landau level and single - electron excitations are gapped .
in particular , no single - electron tunneling into the fqh edge channel of a @xmath6 system is possible at low energies .
the filling factor for bosonic electron pairs is @xmath24 .
thus , we use the edge theory of a laughlin state of bosons at that filling factor .
the lagrangian density @xcite @xmath25+\mathcal{l}_{2}+\mathcal{l}_{\rm int},\ ] ] where @xmath26\ ] ] describes two integer edge modes , @xmath27 contains information about the interaction of the integer and fractional modes , the charge density on the fqh edge is @xmath28 and the charge density on the integer edges is @xmath29 . the operator @xmath30 annihilates a charge-@xmath14 anyon .
the pair annihilation operator on the fractional edge is @xmath31 .
we do not make assumptions about the spin of the electron pairs in the @xmath5 state .
the same edge theory describes spin - polarized and unpolarized fqh liquids . certainly , the particle - hole conjugate system has the filling factor 5/2 only for the spin - polarized @xmath5 liquid .
this state comes in two shapes with identical signatures in the tunneling experiment but different spin polarizations @xcite .
the spin - unpolarized version can be understood as a halperin bilayer state with spin - up and -down electrons playing the role of two layers . the lagrangian density on the edge @xmath32+\mathcal{l}_{2}+\mathcal{l}_{\rm int},\ ] ] where the @xmath33-matrix@xcite @xmath34 and @xmath35 and @xmath27 are defined in the same way as in the previous subsection .
the charge density on the fractional edge is @xmath36 and the most relevant operators that annihilate quasiparticles with charge @xmath14 are @xmath37 .
there are two independent electron operators : @xmath38)$ ] and @xmath39)$ ] .
the spin - unpolarized 331 state is not an eigenstate of the total spin and is related @xcite to the physics of superfluid he-3 .
the spin - polarized version of the 331 state emerges from the condensation of the charge-@xmath40 quasiparticles on top of the laughlin state with the filling factor 1/3 .
the @xmath33-matrix @xmath41 can be expressed as @xmath42 with @xmath43 and hence describes the same topological order as the matrix ( [ dima4 ] ) .
note that the particle - hole conjugate of the 331 state has the filling factor 5/2 only for its spin - polarized version .
the non - abelian pfaffian state @xcite can be seen as a @xmath44-wave superconductor with the wave function @xmath45\pi_{i < j}(z_i - z_j)^2\exp(-\frac{1}{4}\sum|z_i|^2),\ ] ] where @xmath46 is the position of the @xmath47th electron , @xmath48 stays for the pfaffian of a matrix and we ignore the two filled landau levels for a moment .
the edge theory is described by the following lagrangian density @xmath49 + i\psi(\partial_t+v_\psi\partial_x)\psi+\mathcal{l}_{2}+\mathcal{l}_{\rm int},\ ] ] where the charge density on the fractional edge is @xmath50 , @xmath51 is a neutral majorana fermion , @xmath52 and @xmath53 are the velocities of the charged and neutral modes and the meaning of @xmath54 and @xmath27 is the same as in the preceding subsections . the charge @xmath14-quasiparticle annihilation operator @xmath55 , where the operator @xmath56 changes the boundary conditions for the majorana fermion from periodic to antiperiodic and has the scaling dimension @xmath57 .
the pfaffian state is spin - polarized .
this is another spin - polarized non - abelian state .
its wave function can be derived from the parton construction @xcite .
we split an electron into a fermion @xmath58 of charge @xmath59 and two fermions @xmath60 , @xmath61 of charge @xmath14 : @xmath62 the filling factor for the @xmath59-partons is 1 . for each of the two sorts of @xmath14-partons
the filling factor is 2 .
for decoupled partons the wave function would simply be a product of three integer quantum hall ( iqh ) wave functions .
taking into account that the three types of partons occupy exactly the same positions , we find the ground state wave function of the form @xcite
@xmath63 ^ 2\pi_{k < l}(z_k - z_l)\exp(-\frac{1}{4}\sum|z_k|^2),\ ] ] where @xmath64 is the polynomial factor in the wave function of two filled landau levels .
the decomposition ( [ dima9 ] ) exhibits @xmath65 gauge symmetry ( the factor @xmath66 acts on @xmath67 and @xmath61 ; @xmath68 describes the freedom to choose the phase of @xmath58 ) . for decoupled partons ,
the edge theory would contain 5 iqh edge channels .
the theory of the fqh edge is obtained by keeping only gauge invariant states and is described by the lagrangian density @xcite @xmath69+i \psi ( \partial_{t}+v_{\psi}\partial_{x})\psi , \end{aligned}\ ] ] where @xmath70 is a bosonic mode carrying electric charge , @xmath71 is an electrically neutral bosonic mode and @xmath51 is a neutral majorana fermion .
the charge density is @xmath72 .
the neutral mode @xmath73 describes `` pseudo - polarization '' , i.e. , the difference in the occupation of the two landau levels for @xmath14-partons .
the two integer edge modes should be included in the same way as in eqs .
( [ dima1],[dima3],[dima8 ] ) .
the @xmath14-quasiparticle annihilation operators are @xmath74 , where @xmath56 twists the boundary conditions for the majorana fermion and has the scaling dimension 1/16 .
three operators annihilate an electron on the fractional edge : @xmath75 and @xmath76 . in the presence of disorder
the `` pseudo - polarization '' does not conserve and the action can contain tunneling operators of the form @xmath77 , where @xmath78 are random functions of the coordinate . their effect is the same as the effect of similar operators in the anti - pfaffian state @xcite and the anti-331 and anti-@xmath3 states ( appendices a and b ) . the theory acquires an emergent @xmath79 symmetry .
the neutral modes should be described as three majorana fermions with equal velocities .
in contrast to appendices a and b , however , this does not affect the scaling dimension of the most relevant quasiparticle operators .
we now turn to the particle - hole duals of the above four states . in appendix c
we argue that the anti-@xmath5 state does not exhibit universal behavior in the tunneling experiment with a broad range of possible values for the tunneling exponent @xmath9 .
moreover , we do not expect the quantized conductance @xmath80 in that state at sufficiently low temperatures and voltages .
its physics is thus quite different from the physics of the other seven states considered in this paper .
the edge theory of the anti-331 state can be constructed by reversing the direction of the two fqh edge modes and adding another integer edge channel in the opposite direction to the reversed fractional channels .
the presence of impurities ensures electron tunneling between the fractional edge and the additional integer edge . at weak interaction
this tunneling turns out to be irrelevant and the physics become nonuniversal just like in the anti-@xmath5 state . in contrast to the anti-@xmath5 case , the anti-331 state exhibits universal conductance and tunneling exponents at stronger interactions .
this is the situation addressed in appendix a. the fqh edge theory possesses the @xmath81 symmetry and contains a bosonic charge mode and four upstream majorana fermions with identical velocities .
the lagrangian density is @xmath82+i\sum_{k=1}^4\tilde{\psi}_k ( \partial_t-\bar{v}_n \partial_x)\tilde{\psi}_k \nonumber\\ & + \frac{\pi } { 6}v_{n_1n_2}\varepsilon^{ijkl}\tilde{\psi}_i\tilde{\psi}_j\tilde{\psi}_k\tilde{\psi}_l+\mathcal{l}_2+\mathcal{l}_{\rm int},\end{aligned}\ ] ] where @xmath83 are four majoranas , the charge density on the fqh edge is @xmath72 and the contributions @xmath54 and @xmath27 describe two integer channels as above .
the most relevant quasiparticles carry the electric charge @xmath59 .
there are four most relevant @xmath14-particle annihilation operators of the structure @xmath84 ( @xmath85 ) , where @xmath86 changes the boundary conditions from periodic to antiperiodic for all four majorana fermions ( see appendix a for an explicit expression ) and has the scaling dimension 1/4 . the anti - pfaffian state @xcite is obtained from the pfaffian state in exactly the same way as the anti-331 state can be obtained from the 331 state .
the edge theory is very similar to ( [ dima12 ] ) and exhibits the @xmath79 symmetry .
the edge lagrangian density is @xmath87\nonumber\\ + \sum_{k=1}^3 i\psi_k(\partial_t-\bar{v}_n \partial_x)\psi_k+\mathcal{l}_2+\mathcal{l}_{\rm int}.\end{aligned}\ ] ] the two most relevant @xmath14-quasiparticle operators express as @xmath88 ( @xmath89 ) , where @xmath86 changes the boundary conditions from periodic to antiperiodic for all majorana fermions and has the scaling dimension 3/16 .
the edge theory is similar to the anti - pfaffian and anti-331 cases .
its derivation is given in appendix b. there are 5 majorana fermions on the edge .
this corresponds to the emergent @xmath90 symmetry .
the edge lagrangian density @xmath91\nonumber\\ + \sum_{k=1}^5 i\psi_k(\partial_t-\bar{v}_n \partial_x)\psi_k+\mathcal{l}_2+\mathcal{l}_{\rm int}\end{aligned}\ ] ] differs from eq .
( [ dima13 ] ) only by the number of upstream neutral modes .
the four most relevant @xmath14-anyon operators can be written as @xmath88 ( @xmath85 ) , where @xmath86 changes the boundary conditions from periodic to antiperiodic for all majorana fermions and has the scaling dimension 5/16 .
in luttinger liquid systems with position - independent interaction constants , such as the edge theories discussed in section ii , the low - energy tunneling density of states scales as @xmath92 , where @xmath9 depends on the details of the system @xcite .
we argue below that @xmath9 assumes different values in different geometries .
the tunneling conductance is proportional to the product of the tunneling densities of states @xmath93 and @xmath94 on the upper and lower edges at @xmath95 ( ref . ) . in general , the edges are described by two different exponents @xmath96 and @xmath97 . at low temperatures , the linear conductance for the tunneling current @xmath98 .
hence , the experimentally measured @xmath99 . as we discussed in section
ii , in all three geometries the tunneling current is carried by quasiparticles with charge @xmath14 .
the best fits @xcite for the tunneling exponent at this quasiparticle charge are @xmath100 , @xmath101 and @xmath102 .
we want to connect these numbers with the nature of the topological order . since the electric current is carried by the edges , the exponent @xmath9 depends on the edge physics near the qpc .
the size of the relevant region near the point contact is set by the distance a charge can travel during the tunneling event : @xmath103 , where @xmath104 is the velocity of the edge excitation and @xmath105 is the duration of the tunneling event .
the latter can be estimated from the uncertainty relations , @xmath106 , where @xmath107 is the uncertainty of the quasiparticle energy . assuming that @xmath108 cm / s , we find @xmath109 @xmath110 m at relevant values of the temperature and voltage bias .
let us look at the geometry of the edges within @xmath111 @xmath110 m from the qpc .
the edges are defined by gates at the distance of 200 nm from 2deg .
the widths of both gates in the geometry 1a and the upper gate in the geometry 1b have the same order of magnitude of hundreds nanometers @xcite .
the width is 2200 nm for the lower gate in the geometry 1b and both gates in the geometry 1c .
the distance between the edge channel segments on the opposite sides of a gate differs from its width . in our case
there are several edge channels that can run at different locations .
it is not easy to determine those locations . in particular , the annealing procedure used in refs
. changes the device electrostatics but it is unclear what its effect on the geometry is .
a crude estimate of the edge channel positions can be obtained from ref . .
we find that the edge channels run within the distance of hundreds nanometers from the gates .
this agrees with the upper bound one can derive from the gate geometry in fig .
the tip of the upper gate is at 600 nm from the lower gate .
the distance between the outermost edge channel and the gate is thus expected to be below 600/2 nm @xmath112 nm . we conclude that the gate width , the gate distance from 2deg and the distance between the edge modes on the opposite sides of the gate all have the same order of magnitude for both gates in the geometry 1a and the upper gate in the geometry 1b .
this means that there is a significant unscreened coulomb interaction between the segments of the edge on the opposite sides of the gates . on the other hand , in the geometry 1c and for the lower gate in the geometry 1b
, the width of the gates is close to the distance between the edge channels on their opposite sides and much greater than the gate distance from the 2deg plane .
thus , we expect that the coulomb interaction across the gates is almost completely screened in the latter geometries .
one can also neglect the electrostatic interaction between the edge channels defined by two different gates in all geometries .
indeed , an edge point at the distance @xmath113 @xmath110 m from the tip of the gate is much further from the edge on the other side of the qpc than from the gate . what is the effect of the electrostatic interaction across the gate on the tunneling current ?
it is easy to see that repulsive coulomb interaction suppresses tunneling . indeed , after a tunneling event , the tunneling charge must move away from the qpc .
due to the chiral transport on the quantum hall edge , it moves along segment oa in fig .
2 . when the excess charge arrives at point a , it pushes charge from point b due to their repulsive electrostatic interaction . because of chirality the charge from point b can only move towards the tip o of the gate .
thus , excess charge accumulates in point o. this means lower tunneling density of states than for a noninteracting system where charge rapidly distributes over a large region .
hence , one expects higher values of the tunneling exponent @xmath9 in the geometries 1a and 1b than in the geometry 1c . this is exactly what has been observed .
the lower edge has the same geometry in figs .
1b and 1c and its tunneling density of states is described by the same exponent @xmath97 .
the same exponent also describes the upper edge in fig
thus , @xmath114 .
the upper edge in the geometry 1b is described by a different exponent @xmath96 .
hence , @xmath115 . if two identical gates of the same geometry as the upper gate in fig .
1b were used in fig .
1a then one would obtain @xmath116 .
in such situation @xmath10 , @xmath11 and @xmath12 form an arithmetic series .
interestingly , this agrees with experiment . at the same time , device 1a was made in a different sample and its geometry details are different from the upper gate in fig .
besides , the gate width changes linearly @xcite with the distance @xmath117 from the qpc at the distances @xmath118 in the geometry 1a .
this means that the effective interaction strength depends on the temperature since it depends on the gate width at the distance @xmath119 .
hence , there are corrections to the power law for the tunneling density of states in the geometry 1a .
this may be a reason for a poor fit for the quasiparticle charge @xmath14 from the data in that geometry .
-shaped gate defines an inner and outer quantum hall channels .
dotted lines illustrate coulomb interaction across the gate.,width=163 ] the above discussion has focused on the geometry 1a and the upper gate in fig .
we now briefly discuss some features of the lower gate in the geometry 1b and both gates in the geometry 1c .
those gates are @xmath120-shaped , as shown in fig .
3 , with the width of the metal strips on the order of 200 nm .
thus , each gate gives rise to two fqh edge channels inside and outside the gate .
we are interested in the tunneling in the outer channel .
its coulomb interaction with the inner channel affects the tunneling exponent @xmath9 .
we investigate that effect in section iv.d below and show that it can be neglected , provided that the interaction across the gate is such that the system is not close to the interaction - driven instability .
our main conclusion is that only @xmath12 is unaffected by the electrostatic interaction between different edge segments .
thus , the information about the topological order at the filling factor 5/2 can be obtained from the comparison of @xmath102 with theoretical predictions for the models from section ii .
table i shows the predictions for @xmath9 .
we find that the 331 state gives the best fit with an excellent agreement between the theoretical @xmath121 and the experimental value @xmath122 . as discussed in section iv.d , for a system on the verge of instability , all three exponents @xmath10 , @xmath11 and @xmath12 considerably exceed the universal value in the absence of interactions across the gates . in this unlikely scenario , the data @xcite do not exclude the pfaffian state .
.exponent @xmath9 in the tunneling density of states @xmath92 for a straight edge in various 5/2 states .
the values for the @xmath5 , 331 , pfaffian , @xmath3 and anti - pfaffian states can be found in ref . .
the anti-331 state is addressed in appendix a. appendix b contains calculations for the anti-@xmath3 state . [ cols="^,^,^,^,^,<,<,<",options="header " , ]
in this section we compute the tunneling exponents @xmath9 for various models of the 5/2 state in the presence of unscreened coulomb interaction across the gate .
this will allow us to estimate the strength of the electrostatic interaction across the gate . even though the difference of the tunneling exponents is rather small in the geometries 1b and 1c , we find that the electrostatic interaction is strong : the interaction between the nearest points of the edge segments on the opposite sides of the upper gate in fig .
1b turns out comparable with the interaction between the charges , placed in those points in the absence of a gate .
this agrees with what one expects from the geometry of the gates .
the general structure of the action of a straight edge is @xmath123 + l_{\psi},\end{aligned}\ ] ] where @xmath124 denote various edge bose - modes and @xmath125 is the action of the majorana fermion degrees of freedom in the pfaffian , @xmath3 , anti-@xmath3 , anti-331 and anti - pfaffian states .
modes @xmath124 include two fields describing integer quantum hall channels .
stability requires that the symmetric matrix @xmath126 is positive definite . as seen from section ii
, there are no relevant or marginal interactions between the bose and majorana degrees of freedom . and @xmath127 , the charged fqh mode @xmath128 and the neutral mode @xmath51 travel around the shaded gate .
electrostatic interaction across the gate is illustrated by dotted lines.,width=211 ] consider now the geometry of the edge , defined by a narrow gate , fig .
4 . we can still extend integration in eq .
( [ dima15 ] ) from minus to plus infinity , assuming that @xmath117 is measured along the edge , i.e. , @xmath117 is negative on the edge segment above the gate and positive below the gate , fig .
4 . however , eq . ( [ dima15 ] )
is no longer the only contribution to the action .
we must also include the electrostatic interaction across the gate . in the presence of the gate
the interaction is short range and hence can be described by local operators coupling fields at the points @xmath117 and @xmath129 . since we consider low temperature physics , we are interested in relevant and marginal operators only . in the following discussion
we will assume that the list of such operators remains the same as in the limit of the weak coulomb interaction .
this will allow us to use the same scaling dimensions as in the absence of the electrostatic force across the gate .
the most obvious interaction contribution to the action is @xmath130 .
all other contributions made of derivatives of bose fields are irrelevant .
what about contributions with operators of the form @xmath131 which shift charge between different edge channels on each side of the gate ( certainly , no charge tunneling across the gate is allowed ) ?
one might think that some such operators must be included .
indeed , consider the operator @xmath132)$ ] which moves one electron between the two integer modes on each side of the gate .
it appears to be marginal .
however , there is inevitable disorder in our system .
disorder can be described by adding contributions of the form @xmath133 with random @xmath134 to the lagrangian density .
we omitted all such contributions in eq .
( [ dima15 ] ) since one can gauge out disorder by redefining @xmath135 .
we expect that disorder contains a contribution with no correlation between the opposite sides of the gate .
then our redefinition of the fields introduces randomness in the operator @xmath136 and makes it irrelevant .
a similar argument excludes operators , containing @xmath137 in the @xmath3 state .
what about other neutral - mode contributions such as operators with majorana fermions ?
the list of possible operators is restricted by the requirement that the topological charge conserves on each side of the gate : indeed , topological charge can not travel under the gate where there is no fqh liquid .
one might still construct some marginal operators that conserve topological charge on both sides in some of the states from section ii .
for example , the operator @xmath138 is marginal in the anti-331 state .
however , all such operators are prohibited by symmetry .
indeed , as discussed in appendix a , disorder gives rise to the emergent @xmath81 symmetry in the anti-331 state .
the electrostatic coupling between the sides of the gate has the same form as for an infinitely long gate . in the latter case , the system is invariant with respect to two independent @xmath81 symmetry groups acting on the two electrostatically coupled edges .
the combined @xmath139 symmetry excludes the operator @xmath140 .
a similar argument applies in the anti - pfaffian and anti-@xmath3 states .
we conclude that the action can be chosen in the form @xmath141 the tunneling at the qpc at @xmath142 is described by the contribution to the action @xmath143 where @xmath144 and @xmath145 are quasiparticle operators on the upper and lower edges , @xmath146 and @xmath147 distinguish different quasiparticle operators , @xmath148 are the tunneling amplitudes and @xmath149 is the ultraviolet cut - off energy .
certainly , every allowed tunneling process conserves the electric and topological charges . in order to find the low - temperature conductance , it is convenient to perform the renormalization group procedure .
we integrate out fast degrees of freedom and rescale @xmath150 . only the terms with the most relevant @xmath14-excitation operators @xmath151 and @xmath145 with the scaling dimensions @xmath152 and @xmath153 must be kept .
we stop at the energy scale @xmath95 . at that scale @xmath154 .
the linear conductance @xmath155 can now be found from the perturbation theory : @xmath156 .
we conclude that @xmath157 , @xmath158 and @xmath159 . thus , to find @xmath9 we need to compute the scaling dimensions of quasiparticle operators of a general structure @xmath160 , where the operator @xmath56 acts on the majorana sector ( and is just an identity operator in some models ) and @xmath161 are constants ( see section ii ) .
since the majorana sector is decoupled from the bose modes , the scaling dimension of the operator @xmath56 is not affected by the interaction across the gate and can be taken from section ii .
we thus focus on the exponential factor in @xmath162 . in all models of section
ii , the symmetric @xmath33-matrix @xmath163 is positive definite . in the rest of this subsection
we also assume that @xmath164 , eq . ( [ dima16 ] ) , is symmetric .
this is automatically satisfied , if the gate configuration has a symmetry plane such that reflections in the plane transform the segments of the edge above and below the gate into each other .
the gate configuration @xcite , fig .
1a , has such a symmetry plane .
the configuration of the upper gate , fig .
1b , used in ref . , does not have a symmetry plane but is approximately symmetric beyond about 10 @xmath110 m from the qpc .
note that @xmath164 can be symmetric even in the absence of a symmetry plane .
this happens if one can neglect all interactions across the gate except the interaction of the charged modes .
then @xmath164 becomes a @xmath165 matrix .
this is the case in the model of the next subsection .
thus , in subsection iv.c , we can rely on subsection iv.b even without assuming the existence of a symmetry plane .
certainly , our qualitative discussion in section iii is also free from that assumption .
calculations are similar but more cumbersome without the symmetry .
we focus on the symmetric situation because it allows a proof of two general theorems : 1 ) in appendix d we demonstrate that @xmath164 is positive definite ; 2 ) we show that the positive definite @xmath164 drives the tunneling exponent above the universal value without interaction across the gate .
the physical origin of the latter claim has been addressed in section iii . at this point
it is convenient to change notation .
we define @xmath166 and @xmath167 , @xmath168 .
the bose - mode contribution to the action now reads @xmath169\partial_x\theta_i\partial_x\theta_j+[v_{ij}-\frac{u_{ij}}{2}]\partial_x\bar\phi_i\partial_x\bar\phi_j \}.\end{aligned}\ ] ] the quasiparticle operator becomes @xmath170 .
stability implies that @xmath171 are positive definite matrices .
we next diagonalize the bilinear form @xmath172 with the transformation @xmath173 , where the matrix @xmath174 is such that @xmath175 .
the bose contribution to the action becomes @xmath176\partial_x\tilde\theta_i \partial_x\tilde\theta_j + [ \tilde v_{ij}-\tilde\lambda_{ij}]\partial_x\tilde\phi_i \partial_x\tilde\phi_j \},\end{aligned}\ ] ] where @xmath177 and @xmath178 are positive definite symmetric matrices . integrating out the fields
@xmath179 one gets @xmath180 the quasiparticle operator assumes the form @xmath181 .
we introduce another piece of notation now .
consider a symmetric positive definite matrix @xmath182 .
we define @xmath183 in the following way .
we first find such orthogonal matrix @xmath184 that @xmath185 , where @xmath186 is a diagonal matrix with positive diagonal entries .
next , we define @xmath187 as a diagonal matrix with positive diagonal entries such that @xmath188 . finally , we set @xmath189 . obviously , @xmath183 is a symmetric positive definite matrix whose square is @xmath182 ; @xmath190 .
we make the following change of variables @xmath191{\sqrt{\tilde v+\tilde\lambda}(\tilde v-\tilde \lambda)\sqrt{\tilde v+\tilde \lambda}}}\right ] ^{-1}\right)_{ij}\hat\phi_j.\ ] ] the action becomes @xmath192 where @xmath193 at this point we trace back the steps that led us to eq .
( [ dima19 ] ) with eq .
( [ dima21 ] ) as the starting point .
we introduce auxiliary fields @xmath194 and rewrite the action in the form , similar to eq .
( [ dima18 ] ) : @xmath195 next , we define new fields @xmath196 according to @xmath197 and @xmath198 , @xmath168 , and end up with the action @xmath199,\ ] ] in eq .
( [ dima23 ] ) we return to our initial definition of the coordinate @xmath200 and the points @xmath117 and @xmath129 no longer interact .
the quasiparticle operator @xmath201 $ ] .
it is now easy to write down the scaling dimension @xmath202 of @xmath162 and @xmath203 : @xmath204 where @xmath205 is the scaling dimension of the operator @xmath56 .
we now prove that for any choice of @xmath206 , i.e. , for any quasiparticle operator , @xmath9 in eq .
( [ dima24 ] ) exceeds the value one would obtain without coulomb interaction between the edge segments on the opposite sides of the gate . in the above expression
the electrostatic interaction across the gate enters only through the symmetric positive definite matrix @xmath207 .
we thus wish to prove that @xmath9 , eq .
( [ dima24 ] ) , is greater at @xmath208 than at @xmath209 .
@xmath205 does not depend on @xmath207 .
the symmetric positive definite matrix @xmath210 that enters ( [ dima24 ] ) reduces to the identity matrix at @xmath209 .
hence , it is enough to prove that all eigenvalues @xmath211 of @xmath212 are greater than 1 . for this end
, we notice that @xmath213 let us work in the eigenbasis of @xmath212 and look at the diagonal elements of the above eq .
( [ dimae1 ] ) .
we find @xmath214 and hence @xmath215 .
since both matrices @xmath207 and @xmath216 are positive definite , their diagonal matrix elements are positive .
hence , @xmath217 and @xmath9 ( [ dima24 ] ) increases indeed at nonzero @xmath207 compared to the case of @xmath209 .
this agrees with the relation between the exponents @xmath10 , @xmath11 and @xmath12 in refs . .
in this subsection we address an apparent paradox : the geometry suggests a rather strong interaction across the upper gate in geometry 1b ; why is then the tunneling exponent @xmath11 close to @xmath12 ?
the expression ( [ dima24 ] ) depends on several unknown parameters .
they can not be extracted from a single observable @xmath9 . in particular
, one can not compute the strength of the interaction across the gate .
we can only estimate its order of magnitude from @xmath12 and @xmath11 .
our estimates show that despite a small difference @xmath218 , the interaction is not weak .
interestingly , our estimates do not depend on the nature of the bulk topological order .
the estimates will be based on two simple models . in the first model
we will assume that the interaction contribution @xmath219 to the action reduces to the product of the charged modes on the opposite sides of the gate , @xmath220 where @xmath221 stays for the linear charge density .
such model is legitimate , if all edge channels on each side of the gate run much closer to each other than to the gate . in our second model
the across - the - gate interaction includes only the charge density in the fqh channel : @xmath222 where @xmath223 is the linear charge density on the edge channel , separating the @xmath224 and @xmath13 regions .
model 2 is legitimate , if the gate is close to 2deg and all integer channels run under the gate while fractional channels are sufficiently far from the gate .
it is unlikely that either of the above two sets of assumptions accurately describes the experimental system . at the same time ,
our estimates of @xmath225 in eqs .
( [ dima25 ] ) and ( [ dima26 ] ) will give an idea of the range of possible interaction strength .
for simplicity we will assume that the mode @xmath226 in model 1 and the mode @xmath227 in model 2 do not interact with any other edge modes on the same side of the gate .
the application of eq .
( [ dima24 ] ) then becomes very easy .
consider first model 1 .
we can always make a linear change of the variables @xmath124 such that one of the new variables is @xmath226 and all other new variables commute with @xmath228 .
then the dynamics of the charged mode is completely independent from all other modes .
the form of the charged mode action is dictated @xcite by the quantum hall conductance @xmath229 : @xmath230,\ ] ] where @xmath231 is the velocity of the mode .
since the quasiparticle charge is @xmath14 , the field @xmath226 must enter the quasiparticle operator as @xmath232)=\exp(i\phi_\rho/10)$ ] . finally , from eq .
( [ dima24 ] ) one finds : @xmath233.\ ] ] in model 2 we similarly assume that the charged mode of the fqh edge is decoupled from all other modes .
it is controlled by the action @xmath234,\ ] ] where the coefficient 2 reflects @xcite the fqh conductance @xmath235 .
the quasiparticle operator contains @xmath236 in the exponential factor @xmath237)=\exp(i\phi^f_{\rho}/2)$ ] .
hence , eq . ( [ dima24 ] ) yields @xmath238}.\ ] ] substituting the experimental @xmath239 in eqs .
( [ dima28],[dima30 ] ) we find @xmath240 and @xmath241 .
these values are similar to the estimate @xcite in a related geometry in the 1/3 state . certainly , the above two models are not very realistic .
besides , even small uncertainties in @xmath11 and @xmath12 result in major uncertainties in @xmath242 . note also that we neglect the repulsive interaction between the upper and lower edges of the 2200 nm - wide constriction in the geometry 1c since the constriction width is much shorter than the relevant thermal length .
this interaction decreases @xmath12 and hence decreases the estimates ( [ dima28],[dima30 ] ) for @xmath242 .
finally , in the above discussion we disregarded the repulsive interaction between the edge modes on the inner and outer sides of the gates in the geometry 1c and the lower gate in the geometry 1b ( fig .
this interaction increases both @xmath11 and @xmath12 compared to the situation without an inner edge . in the next subsection
we show that it is safe to neglect the interaction of the inner and outer channels unless it is so strong that the system is on the verge of instability .
this possibility is not a concern for us here because our goal is to demonstrate that the interaction is not weak .
overall , it appears safe to conclude that physical @xmath243 is within an order of magnitude from the above values .
this corresponds to a significant interaction across the gate . in this subsection
we address the interaction between the inner and outer edge channels , fig .
3 , in the geometry 1c and around the lower gate in the geometry 1b . we find that for most values of parameters this interaction has considerably less effect on the tunneling exponents @xmath11 and @xmath12 than the interaction discussed in the previous subsection .
this justifies neglecting the interaction of the inner and outer channels . to estimate the change in the exponents
@xmath11 and @xmath12 we consider a model , similar to the second model of the previous subsection .
the action reads @xmath244,\end{aligned}\ ] ] where @xmath245 and @xmath246 describe the charge density in the outer and inner fractional edge channels .
we are interested in the tunneling into the outer channel .
@xmath245 enters the tunneling operator through the factor @xmath247 .
thus , we wish to compute the correction @xmath248 to the scaling dimension of the above operator due to a nonzero @xmath249 .
it is convenient to change the variables : @xmath250 where @xmath251{\frac{v_\rho-\lambda_2}{v_\rho+\lambda_2}}$ ] . in the new variables
, we discover the action of two noninteracting chiral fields : @xmath252\}.\end{aligned}\ ] ] computing @xmath248 is now easy .
one finds @xmath253 the experimentally measured @xmath9 differs from the universal value @xmath254 for a system without an inner channel along a wide gate and coulomb interaction across a narrow gate : @xmath255 and @xmath256 , where @xmath257 is given by eq .
( [ dima30 ] ) and reflects the effect of the interaction across the upper gate in fig .
1b . in the previous subsection we estimated @xmath241 .
this corresponds to @xmath258 . substituting the same value of @xmath249 in eq .
( [ 4d4 ] ) , one finds @xmath259 .
this justifies neglecting @xmath248 above .
it is easy to see that @xmath248 is much less than 1 and considerably smaller than @xmath257 as long as @xmath260 is not close to 1 .
the unlikely regime @xmath261 describes a system on the verge of an instability due to coulomb interaction . even in that regime @xmath262 for all @xmath249 and
hence @xmath248 can be ignored in crude estimates of , e.g. , the coulomb interaction strength . at the same time , the expressions ( [ dima28 ] ) , ( [ dima30 ] ) and ( [ 4d4 ] ) all diverge as @xmath263 and all three experimental exponents @xmath10 , @xmath11 and @xmath12 provide only upper bounds for @xmath254 in that unlikely limit .
one can avoid complications due to the inner channel by changing the geometry of the gates . instead of @xmath120-shaped gates , one can use gates that contain not only the perimeter but also the inside of a rectangle .
we find that the 331 state with the theoretical @xmath264 gives the best fit to the experimental @xmath102 .
it is instructive to compare this conclusion with the lessons from other types of experiments .
the data on the spin polarization are controversial @xcite .
there is support for both 0 and 100% polarization .
the 331 state comes in both spin - polarized and unpolarized versions with identical transport signatures in the tunneling experiment .
most other proposed 5/2 states are spin - polarized .
some features of the thermopower data @xcite are qualitatively compatible with theoretical predictions for the pfaffian state @xcite .
however , even abelian states may exhibit qualitatively similar behavior , if different quasiparticle species are degenerate or close in energy .
the fabry - perot interference experiments @xcite were interpreted @xcite as supporting the pfaffian or anti - pfaffian state . at the same time , it was shown that the 331 state can produce identical signatures in a fabry - perot interferometer @xcite ( but interestingly , not in a mach - zehnder interferometer @xcite ) .
the anti-331 state may also exhibit identical signatures .
the results of the transport experiment @xcite were explained as a sign of an upstream neutral mode .
such interpretation is incompatible with all proposed states except anti - pfaffian , anti-331 and anti-@xmath3 .
in particular , the 331 state has no upstream edge modes . as discussed in section iii , in the unlikely case of a system on the verge of coulomb - interaction
driven instability , the data @xcite do not exclude the pfaffian state .
that state also does not possess contra - propagating edge modes .
thus , a conflict appears between the interpretation of the experiments from refs . and .
our explanation of the data @xcite is based on the assumption that the system is in the scaling regime , where universal predictions apply .
we also assume that the edges can be described by a chiral luttinger liquid model .
it may happen that such description fails due to , e.g. , dissipation @xcite .
at the same time , edge reconstruction @xcite or bulk transport @xcite may affect the interpretation of the experiments on upstream modes .
on the other hand , various proposed 5/2 states are rather close in energy @xcite and it was suggested that more than one 5/2 state might be present at different conditions or in different samples @xcite .
it is also possible that the true 5/2 state is not one of the simplest states considered above .
more light could be shed on the nature of the 5/2 state by tunneling experiments at lower temperatures . to reduce the effect of coulomb interaction across the gate
, one needs to modify the 1c setup , fig .
the inner edge channel can be eliminated by filling the inside of the @xmath120-shaped gate , fig .
3 , with metal . other methods that could show unique signatures of abelian and non - abelian states include thermopower measurements @xcite and mach - zehnder interferometry @xcite . the non - equilibrium fluctuation - dissipation theorem @xcite would provide an independent test of the existence of upstream modes .
our key idea about the role of gating in the experiments @xcite can be tested with a `` smoking gun '' experiment illustrated in fig .
the width of the gate is different at short distances @xmath265 and long distances @xmath266 from the gate .
the coulomb interaction across the gate is strong at @xmath265 and negligible at large @xmath117 .
in such situation our theory predicts two different power dependencies of the conductance on the temperature at @xmath267 and @xmath268 , where @xmath104 is the excitation speed on the edges .
the higher temperature regime is similar to the geometry 1a and the low temperature limit corresponds to the geometry 1c .
thus , one can go between two fixed points by simply changing the temperature ( or voltage ) .
even though the agreement between the experimental @xmath102 and theoretical @xmath264 looks excellent , one should be cautious about data accuracy .
for example , electron tunneling experiments into the @xmath269 edge have routinely shown a 10% discrepancy with the theory @xcite
. a relative fragility of the 5/2 edge may lead to even lower experimental resolution . nevertheless , the difference between @xmath12 in the 331 and other competing states exceeds 30% in all cases and this lends credibility to the 331 state as the best fit .
a considerable body of numerical work supports pfaffian or anti - pfaffian states ( see , e.g. , refs . ) .
as discussed above , the data @xcite are not compatible with the anti - pfaffian state but do not exclude the pfaffian state in the unlikely case of the system on the verge of instability . on the other hand , numerical studies of small simplified model systems have limitations and only experiment can determine the right state .
in particular , the existing numerical results @xcite for the energy gap at @xmath13 significantly exceed experimental findings .
one limitation of numerics is due to the fact that most studies assume full spin polarization outright .
this would exclude the spin - unpolarized 331 state .
at the same time , several papers @xcite lend support to a spin - polarized ground state .
another limitation comes from the incomplete understanding of the landau - level mixing effects @xcite .
disorder apparently plays a major role in the discrepancy of the theoretical and experimental results for the energy gap @xcite and may affect the nature of the ground state .
we are not aware of any numerical studies that include disorder .
moreover , the numerical energy difference between competing states @xcite is so small that even a tiny and generally neglected effect of the spin - orbit interaction @xcite might be relevant . in conclusion
, we have proposed an explanation of the discrepancy of the tunneling exponents in refs .
two of the three geometries used in refs .
are affected by the electrostatic interaction across the gates that changes the exponent @xmath9 .
we compare the exponent @xmath12 in the third geometry 1c with the theoretical predictions for seven states with the simplest topological orders at @xmath13 .
for some of those states , the tunneling exponents were computed earlier .
in addition , we compute the tunneling exponents for the particle - hole conjugates of the non - abelian @xmath3 state and abelian 331 state in appendices a and b. the latter state can also be viewed as the bonderson - slingerland state built on top of the particle - hole conjugate of the laughlin 1/3 state .
we find that the 331 state gives the best fit to the experiment .
we thank c. dillard , m. a. kastner , x. lin , n. e. staley and c. wang for useful discussions .
this work was supported by the nsf under grant no .
in this appendix we formulate the edge theory of the particle - hole conjugate of the spin - polarized @xmath4 state which we call the anti-331 state .
we only consider the fqh edge between the @xmath224 and @xmath13 regions .
the two additional integer edge modes will play little role in our discussion .
we use an approach similar to the treatment of the disorder - dominated @xmath270 state@xcite and the anti - pfaffian state@xcite .
we start from a clean particle - hole conjugate of the @xmath4 state which is obtained by condensing hole excitations of the @xmath271 quantum hall state into the @xmath4 state . in our picture
, the 2deg has an annular shape with the inner edge being the interface between the hole @xmath4 state and its parental @xmath271 iqh state and the outer edge being the interface between the @xmath271 quantum hall state and the vacuum .
the edge is the combination of a right - moving @xmath271 edge and a left - moving fqh edge of the @xmath4 state with the lagrangian density @xmath272,\ ] ] where the @xmath33-matrix @xcite and the potential matrix are @xmath273 @xmath274 is the right - moving @xmath271 charge density mode , @xmath275 and @xmath276 are two left - moving modes in the edge theory of the @xmath4 state .
the charge vector that determines how the modes ( @xmath277 ) couple to the external gauge field and contribute to the electric current is @xmath278 .
the @xmath33-matrix and the charge vector @xmath279 together characterize the topological order of the fqh state . to simplify the discussion , we assume at this point that @xmath280 .
we will also assume that the @xmath271 edge mode only couples with the edge mode in the bottom hierarchy of the hole @xmath4 state so that @xmath281 .
this approximation reflects the fact that only the bottom level hierarchy contains electrons which interact with the @xmath271 iqh liquid through the coulomb potential . at the end of the appendix
, we will relax the above assumptions and will see that they do not change any conclusions .
the lagrangian @xmath282 describes the low energy edge physics of a clean particle - hole conjugate of the @xmath4 state . in reality , there are always impurities on the edge .
they destroy the translational invariance on the fqh edge and cause inter - edge tunneling which leads to the edge mode equilibration .
nontrivial topological charges can not travel between the hole @xmath4 edge and the @xmath271 channel .
hence , only electrons can tunnel .
the electron operator on the @xmath271 edge is @xmath283 .
the most relevant electron operators on the hole @xmath4 edge are @xmath284 and @xmath285 .
the appropriate term for the electron tunneling in the lagrangian density is @xmath286 where we have suppressed the klein factors that are necessary to ensure correct statistics among different electron operators . @xmath287 and @xmath288 are complex variables characterizing the strength of random impurities .
we assume for simplicity that their distribution is gaussian and they are @xmath289-correlated : @xmath290 and @xmath291 . at weak disorder @xmath292 and @xmath293 ,
the effect of electron tunneling on the fqh edge is determined by the scaling behavior of the tunneling operators .
using @xmath282 , we find that the two tunneling operators have the same scaling dimension @xmath294 where @xmath295 .
the leading order renormalization group ( rg ) equations for the disorder strength are@xcite @xmath296 where @xmath297 . for @xmath298 ,
the electron tunneling is irrelevant and the low temperature edge physics is described by @xmath282 in eq .
( [ dimaa1 ] ) in which no equilibration occurs between the @xmath271 edge and the hole @xmath4 edge .
the physics is very similar to the physics of the anti-@xmath5 state , addressed in appendix c. quasiparticle tunneling is nonuniversal . at low voltages and temperatures the quantum hall conductance
is quantized at @xmath299 instead of the right value @xmath80 .
this happens because the integer mode conductance @xmath300 and the conductance @xmath301 of the fractional edge must be added and not subtracted in the quantum hall bar geometry
. a detailed discussion of a similar point can be found in ref . .
on the other hand , if @xmath302 the electron tunneling is relevant and we end up with a disorder - dominated phase with equilibrated edge modes at low temperatures .
this is the situation we consider below .
to study the disorder - dominated phase , it is useful to rewrite the edge dynamics in terms of a charged mode , represented by the charge vector @xmath279 , and two independent neutral modes , represented by the neutral vectors @xmath303 , @xmath304 .
if we treat the matrix @xmath305 as a metric then neutral vectors are those with vanishing inner products with the charge vector , @xmath306 , @xmath307 .
we choose @xmath308 , @xmath309 .
the corresponding charged @xmath310 and neutral @xmath311 , @xmath312 boson fields are @xmath313 in the basis ( @xmath314 ) , the lagrangian density of the tunneling problem @xmath315 , where @xmath316\ ] ] and @xmath317 \nonumber \\ & + [ \xi_{1}(x ) e^{i(\phi_{n_{1}}-\phi_{n_{2}})}+\xi_{2}(x ) e^{i(\phi_{n_{1}}+\phi_{n_{2}})}+ \textrm { h.c . } ] ; \nonumber \\ \mathcal{l}_{\textrm{sb}}= & -\frac{1}{4\pi } \sum_{i=1,2}\delta v_{n_i } ( \partial_{x}\phi_{n_{i}})^{2}-\frac{2}{4\pi}v_{\rho n_1}\partial_{x}\phi_{\rho}\partial_{x}\phi_{n_{1}}.\end{aligned}\ ] ] the velocities of @xmath70 , @xmath311 and @xmath318 are @xmath319 , @xmath320 and @xmath321 , respectively . in writing the lagrangians ,
we have defined the average velocity @xmath322 and moved the anisotropic part of the velocities @xmath323 into @xmath324 .
the reason for this is that @xmath325 now has a hidden @xmath81 symmetry whereas @xmath324 is the symmetry - breaking term .
the interaction between the charged mode @xmath70 and the neutral mode @xmath311 is @xmath326 .
note that after the basis change the electron tunneling only couples the neutral modes and contributes to their equilibration .
the charged mode is left alone with its own velocity @xmath327 . to view the @xmath81 symmetry in @xmath325 more clearly
, we fermionize @xmath311 and @xmath318 into chiral fermions by setting @xmath328 and @xmath329 .
we then further break the two chiral fermions into four real majorana fermions defined as @xmath330 , @xmath331 , @xmath332 and @xmath333 . in terms of the majorana fermions , the lagrangians @xmath325 and @xmath324 are @xmath334 where @xmath335 and @xmath336 .
the new random variables are defined as @xmath337 , @xmath338 , @xmath339 , @xmath340 , and all other @xmath341 .
the hermitian matrices @xmath342 are the generators of the @xmath81 group in the fundamental representation , @xmath343 with @xmath344 .
the other two matrices in the lagrangians are @xmath345 and @xmath346 . as the last step , we perform a local rotation of the majorana fermions @xmath347 , where @xmath348 and @xmath349 is the path - ordering operator . clearly , @xmath350 and @xmath351 . after the rotation ,
the lagrangians become @xmath352 where @xmath353 and @xmath354 are spatially random matrices . on the second line of the above equation
we omit a term with the derivative @xmath355 .
its structure is similar to the second term on the first line of eq .
( [ dimaa9 ] ) and it can be removed with a variable change , similar to eq .
( [ dimaa11 ] ) .
the @xmath81 symmetry in @xmath325 is now clearly manifest .
let us study the symmetry - breaking terms in @xmath324 . naively , both operators in @xmath324 have scaling dimension 2 and they are marginal .
however , the random coefficients @xmath356 and @xmath357 make them irrelevant under a perturbative rg analysis.@xcite hence , both terms in @xmath324 scale to zero at low temperature . in general , our initial simplifying assumptions for
the potential matrix may not hold and we will have two more terms in the lagrangian , @xmath358 where @xmath359 is the levi - civita tensor with @xmath360 , and @xmath212 is a spatially random matrix .
@xmath361 is a marginal term that respects the @xmath81 symmetry .
it should be retained in the edge theory of the anti-331 state . on the other hand ,
the operator in @xmath362 is irrelevant .
any @xmath81-invariant combination of the majorana fermions can be constructed from the majorana operators and their derivatives , the kronecker tensor @xmath363 and the levi - civita tensor @xmath359 . keeping all possible relevant terms we find the full lagrangian that describes the low energy edge physics of the anti-331 state : @xmath364+i\tilde{\psi}^t ( \partial_t-\bar{v}_n \partial_x)\tilde{\psi}\nonumber \\ & + \frac{\pi } { 6}v_{n_1n_2}\varepsilon^{ijkl}\tilde{\psi}_i\tilde{\psi}_j\tilde{\psi}_k\tilde{\psi}_l.\end{aligned}\ ] ] the edge theory consists of one right - moving charged density mode and four left - moving majorana fermions obeying an explicit @xmath81 symmetry .
the density mode and the majorana fermions decouple .
the thermal conductance of a fqh state is a universal quantity that is determined by the bulk topological order .
it depends only on the numbers of right - moving and left - moving edge modes and is robust with respect to interactions ( for example , the off - diagonal elements in the potential matrix ) and disorder.@xcite using the lagrangian in eq .
( [ dimaa1 ] ) or eq .
( [ dimaa14 ] ) , we obtain the thermal conductance in the anti-331 state as @xmath365 , in units of @xmath366 ( the first factor of 2 comes from the two integer modes ) .
this is different from the thermal conductance @xmath367 in the 331 state , indicating different topological orders of the 331 and anti-331 state .
a generic quasiparticle operator on the edge of the anti-331 state is @xmath368 with the electric charge @xmath369 in units of @xmath15 , where @xmath370 , @xmath371 , @xmath372 are independent integers .
its scaling dimension @xmath373 can be obtained by computing the correlation function , @xmath374 at zero temperature , using the @xmath81 invariant lagrangian from eq .
( [ dimaa14 ] ) .
equivalently yet more easily , we can use the boson lagrangians from eqs .
( [ dimaa7],[dimaa8],[dimaa13 ] ) by setting @xmath375 and @xmath376 .
this is because the rotation @xmath377 is local and does not change the equal - space correlation functions of quasiparticle operators .
the interaction @xmath378 between copropagating neutral bosons can not modify the scaling behavior of a quasiparticle @xcite .
using @xmath379 , @xmath380 and @xmath381 , we obtain the universal scaling dimension of @xmath382 as @xmath383 the most relevant quasiparticles are those with the minimal scaling dimension ( [ dimaa15 ] ) . here , it is @xmath384 with the electric charge @xmath59 and scaling dimension @xmath385 .
the next - most relevant quasiparticles are @xmath386 , @xmath387 , @xmath388 and @xmath389 with the electric charge @xmath14 and scaling dimension @xmath390 .
the above quasiparticle operators can all be represented in the form @xmath391 , where the last two exponential factors correspond to the operators @xmath86 in section ii.f .
in this appendix we discuss the edge physics of the anti-@xmath392 state , which is the disorder - dominated particle - hole conjugate of the @xmath392 state . as in the previous appendix
, we focus on the fqh edge separating @xmath224 and @xmath13 regions .
our approach is the same as in appendix a. as a starting point , we write down the edge lagrangian of a clean particle - hole conjugate of the @xmath392 phase @xmath393\nonumber \\ & -\frac{2}{4\pi}v_{0\rho}\partial_{x}\phi_{0}\partial_{x}\phi_{\rho}+i \psi ( \partial_{t}-v_{\psi}\partial_{x})\psi , \end{aligned}\ ] ] where @xmath274 is the right - moving @xmath271 edge mode and @xmath70 , @xmath71 , @xmath51 are left - moving @xmath392 fqh edge modes .
we first assume that only charged modes interact via coulomb potential @xmath394 .
at the end of the appendix we will see that our conclusions do not depend on that assumption .
the charge vector of the boson fields ( @xmath395 ) is @xmath278 .
the majorana fermion @xmath51 is neutral . in the presence of disorder
, electrons may tunnel between the integer edge and the fqh edge .
the electron operator on the @xmath271 edge is @xmath283 .
the most relevant electron operators on the fqh edge are @xmath396 , @xmath397 and @xmath398 .
the lagrangian density that describes the electron tunneling due to impurities on the edge is @xmath399 where @xmath400 are complex gaussian variables characterizing the strength of disorder , @xmath401 , where @xmath402 . to simplify notation
we suppress klein factors as in appendix a. from @xmath282 in eq .
( [ dimab1-new ] ) , we find that the tunneling operators in eq .
( [ dimab_add ] ) have the same scaling dimension @xmath403 , where @xmath404 . the scaling dimension @xmath202 determines the edge physics at low temperatures . the leading order rg equations are @xmath405.@xcite when @xmath298 , the electron tunneling is irrelevant and the low temperature edge physics is described by @xmath282 . when @xmath302 , we arrive at the disorder - dominated phase with equilibrated edges at low temperature .
the physics of the non - equilibrated case @xmath302 is nonuniversal . just like in appendix c
we find that the quasiparticle tunneling exponent @xmath9 can assume a broad range of values .
moreover , the quantum hall conductance at low temperatures and voltages does not assume the correct value @xmath80 .
it becomes @xmath299 in the quantum hall bar geometry .
this happens for the same reasons as in appendices a and c : one has to add the conductance @xmath300 of the integer quantum hall subsystem and the conductance @xmath301 of the fqh subsystem .
below we focus on the disorder dominated equilibrated phase .
let us separate the charged and neutral degrees of freedom on the edge by defining @xmath406 , @xmath407 and @xmath408 .
the full lagrangian @xmath409 , where @xmath410 ; \nonumber \\ \mathcal{l}_{\textrm{sym}}= & -\frac{1}{4\pi } \sum_{i=1,2 } [ -\partial_{t}\phi_{n_{i}}\partial_{x}\phi_{n_{i}}+\bar{v}_{n } ( \partial_{x}\phi_{n_{i}})^{2 } ] + [ \xi_{1 } \psi e^{i\phi_{n_{1 } } } \nonumber \\ & + \xi_{2}\textrm{re}\{e^{-i\phi_{n_2}}\ } e^{i\phi_{n_{1 } } } + \xi_{3 } \textrm{im}\{e^{-i\phi_{n_2}}\ } e^{i\phi_{n_{1 } } } + \textrm { h.c . } ] \nonumber \\ & + i \psi ( \partial_{t}- \bar{v}_n\partial_{x})\psi ; \nonumber \\ \mathcal{l}_{\textrm{sb}}= & -\frac{1}{4\pi } \sum_{i=1,2}\delta v_{n_i } ( \partial_{x}\phi_{n_{i}})^{2}-\frac{2}{4\pi}v_{\tilde{\rho } n_1}\partial_{x}\phi_{\tilde{\rho}}\partial_{x}\phi_{n_{1}}\nonumber \\ & -i \psi ( \delta v_{\psi } \partial_{x})\psi.\end{aligned}\ ] ] the velocities of @xmath411 , @xmath311 , @xmath312 are @xmath412 , @xmath413 and @xmath414 , respectively . the interaction parameter @xmath415 . here
we have split the lagrangian into @xmath325 , which respects a hidden symmetry , and the symmetry - breaking part @xmath324 .
the velocity anisotropies @xmath416 and @xmath323 ( @xmath297 ) are defined with respect to the average velocity @xmath417 .
we now fermionize the neutral bosons into majorana fermions according to @xmath418 , @xmath419 , @xmath420 , @xmath421 and define @xmath422 . in terms of the majorana fermions , @xmath423 where @xmath424 and @xmath425 .
random variables @xmath426 are defined as @xmath427 , @xmath428 , @xmath429 , @xmath430 , @xmath431 , @xmath432 and all other @xmath341 .
hermitian matrices @xmath433 are the generators of the @xmath90 group in the fundamental representation , @xmath434 with @xmath435 .
the other matrices in the lagrangians are @xmath436 and @xmath437 .
finally , we make a local rotation of the majorana fermions @xmath347 , where @xmath438 and @xmath349 is the path - ordering operator .
the lagrangians after the rotation are@xmath439 where @xmath353 and @xmath354 are spatially random matrices . as in appendix
a , we omit a term with @xmath355 . it can be removed with another transformation of the type ( [ dimar - new ] ) .
now , the @xmath90 symmetry in @xmath325 is clear .
the symmetry - breaking terms in @xmath324 are irrelevant due to their random coefficients , in exactly the same way as in the anti-331 state , appendix a. in the most general case , the neutral boson @xmath73 in eq .
( [ dimab1-new ] ) may interact with charged bosons @xmath440 and @xmath70 .
this gives rise to two more symmetry - breaking terms in the lagrangian .
however , both terms are irrelevant and disappear at low temperatures .
the edge physics of the anti-@xmath392 state is described by @xmath441 that contains all relevant operators allowed by the symmetry .
in contrast to the anti-331 case with its @xmath81 symmetry group , we do not need to keep any four - fermion operators in the action .
a generic quasiparticle operator on the edge of the anti-@xmath392 state is @xmath442 with the electric charge @xmath443 in units of @xmath15 , where @xmath444 is one of the three fields @xmath445 , @xmath56 , and @xmath51 with the scaling dimensions @xmath446 , @xmath447 and @xmath448 .
@xmath370 is an arbitrary integer . @xmath449 and @xmath450 are integers or half integers .
we are interested in @xmath451 charged quasiparticles .
the quasiparticle charge gives the constraint @xmath452 , which is satisfied only if @xmath449 is a half integer given that @xmath370 is an integer .
in addition , we must require that the quasiparticle is local with respect to electrons , which means there is no branch cut in the correlation function between the quasiparticle operator and any of the electron operators in the theory @xcite .
hence , @xmath450 must also be a half integer and @xmath453 .
the most general @xmath451 charged quasiparticle is then @xmath454 with @xmath449 and @xmath450 being half integers and @xmath452 an integer .
the scaling dimension @xmath455 can be computed using the boson lagrangians in eq .
( [ dimab2 ] ) by setting @xmath456 and @xmath457 .
this gives @xmath458 the most relevant @xmath14 and @xmath459 charged quasiparticles are @xmath460 , @xmath461 and @xmath462 , @xmath463 , all with the scaling dimension @xmath464 .
all these operators are products of a factor @xmath465 and one of the twist operators @xmath86 , section ii.h .
the anti-@xmath5 state is the particle - hole conjugate of the @xmath5 state , with the fqh edge lagrangian density @xmath466 -\frac{1}{\pi}v_{01}\partial_{x}\phi_{0}\partial_{x}\phi_{1},\end{aligned}\ ] ] where @xmath440 is the right - moving integer edge mode and @xmath467 is the left - moving fractional edge mode from the @xmath5 state . as in appendices
a and b we focus on the fqh edge between the @xmath224 and @xmath13 regions .
the charge vector of ( @xmath468 ) is @xmath469 , which reflects that the @xmath5 state is a laughlin state of electron pairs .
the lagrangian that describes the tunneling of electron pairs between the integer and fractional edges due to impurities is @xmath470 where @xmath471 is a complex gaussian random variable that describes local disorder , @xmath472 .
the scaling dimension of the tunneling operators is @xmath473 , where @xmath474 .
positive - definiteness of the hamiltonian requires @xmath475 , and hence @xmath476 .
the leading order rg equation for the disorder strength @xmath477 is @xmath478.@xcite by inspection , the scaling dimension @xmath202 is always greater than @xmath479 . hence
, electron - pair tunneling is always irrelevant and a disorder - dominated phase does not exist .
the edge physics of the anti-@xmath5 state is described by @xmath282 . a generic quasiparticle @xmath480 ( @xmath370 , @xmath371 are independent integers ) on the edge of the anti-@xmath5 state has the electric charge @xmath481 in units of @xmath15 .
its scaling dimension @xmath373 is nonuniversal and depends on the parameters in the hamiltonian . with @xmath282 , we find @xmath482 possible values of @xmath373 for charge-@xmath14 excitations range between @xmath57 and @xmath483 .
the low temperature quantum hall bar conductance is quantized at @xmath299 just like in other nonequilibrated states with the same filling factor , ref . .
the above discussion ignores the two integer edge modes @xmath484 , @xmath485 always present in the second landau level states .
if we allow tunneling between those modes and the fqh modes and include operators that transmit three or more electrons then it is possible to find relevant tunneling operators at certain choices of the interaction constants in the hamiltonian .
one example would be the operator @xmath486)$ ] , where @xmath484 describes the integer mode with the same spin polarization as the fqh edge .
we expect the amplitude of such many - body operators to be small and neglect them even if the interaction constants are such that they are relevant .
plane . the mirror - symmetric gate with the potential @xmath487 is shaded.,width=259 ] in this appendix we prove that the interaction matrix @xmath164 in eq .
( [ dima16 ] ) is positive definite .
we consider a system with mirror symmetry , fig .
6 . in fig . 6
the @xmath117-axis is perpendicular to the plane of the figure and the mirror plane is defined by the equation @xmath488 .
the only assumption about the dielectric constant outside the mirror symmetric gates is the mirror symmetry of its coordinate dependence .
our effective models for edge states are one - dimensional and contain only the coordinate @xmath117 but the physical charge distribution is always three - dimensional . the energy of the electrostatic interaction between the edge segments on the left and right of the gate , fig .
6 , depends on the whole three - dimensional distribution of the charges .
the derivatives of the bose fields @xmath489 in eq .
( [ dima16 ] ) are determined by the local charge distribution .
it will be convenient for us to consider a situation in which the charge density @xmath490 does not depend on @xmath117 and exhibits the mirror symmetry @xmath491 .
then all @xmath489 remain constant on the left and right of the gate with @xmath492 .
the interaction energy ( [ dima16 ] ) becomes @xmath493 .
the same energy can be found from electrostatics .
it is just the interaction energy of the charge distribution on the left of the gate , @xmath494 , @xmath495 , with the mirror symmetric charge distribution @xmath494 , @xmath496 .
it is now clear that to prove the positive definiteness of @xmath219 it is sufficient to prove that the electrostatic interaction energy of a set of charges on the left of the mirror plane with the mirror symmetric set of charges on the right of the plane is always positive .
below we compute such electrostatic energy in the presence of mirror symmetric metallic gates .
we set the electrostatic potential of the gates to zero .
the electrostatic potential @xmath497 outside the gates can be represented as the sum of two contributions .
@xmath498 is the electrostatic potential , created by the charges on the left of the mirror plane ( i.e. , the charges from the points with negative @xmath499 ) . by definition @xmath500
includes the effect of the screening charges on the gate surface , i.e. , satisfies the boundary condition @xmath501 .
@xmath502 is created by the charges on the right of the mirror plane in the presence of a screening gate .
the mirror symmetry implies that @xmath503 .
the total electrostatic energy is thus \2 ) alternatively , let us change the sign of all charges on the right of the mirror plane : @xmath508 , @xmath496 .
the potential becomes @xmath509 .
we have created a situation with @xmath510 .
hence , the total electrostatic potential at @xmath495 would not change after all charges with @xmath496 are removed and the region @xmath496 is filled with metal . in the latter situation ,
the total electrostatic energy becomes according to a theorem of electrostatics @xcite , the energy always goes down if a piece of metal is introduced into a system at fixed positions of free charges .
hence , @xmath514 .
subtracting @xmath515 from @xmath516 we find that @xmath517 and hence @xmath164 is positive definite .
|
two recent experiments [ i. p. radu _ et al . _ , science @xmath0 , 899 ( 2008 ) and x. lin _ et al .
_ , phys .
rev .
b @xmath1 , 165321 ( 2012 ) ] measured the temperature and voltage dependence of the quasiparticle tunneling through a quantum point contact in the @xmath2 quantum hall liquid .
the results led to conflicting conclusions about the nature of the quantum hall state . in this paper , we show that the conflict can be resolved by recognizing different geometries of the devices in the experiments .
we argue that in some of those geometries there is significant unscreened electrostatic interaction between the segments of the quantum hall edge on the opposite sides of the point contact .
coulomb interaction affects the tunneling current .
we compare experimental results with theoretical predictions for the pfaffian , @xmath3 , @xmath4 and @xmath5 states and their particle - hole conjugates .
after coulomb corrections are taken into account , measurements in all geometries agree with the spin - polarized and spin - unpolarized halperin 331 states .
|
the existence of non - gaussianity in primordial density fields has the potential to provide a unique observational probe that will enable discrimination among wide variety of inflationary models of the early universe .
versions of the inflation scenario based on the idea of a single slow - rolling scalar field predict levels of non - gaussianity too small to be observed . on the other hand , multi - field inflation models and models with a non - standard kinetic term for the inflaton may yield larger non - gaussian effects which could in principle be detected in current or next - generation observations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in this paper
we focus on the local parametrisation of primordial non - gaussianity by including quadratic corrections to the curvature perturbation during the matter era ( e.g. * ? ? ?
* ) : @xmath3 where @xmath4 represents an auxiliary random - gaussian field and @xmath5 characterizes the amplitude of the non - linear contribution to the overall perturbation .
this local form is motivated by the simple slow - rolling single scalar inflation scenario and other models , including curvaton models ; for an alternative parameterization of @xmath0 , see @xcite .
are not sufficiently sensitive to detect the wavelength dependence of @xmath0 so a constant @xmath0 provides a reazonable parameterization of the level of non - gaussianity . analysis of the angular bispectrum for the wmap 3-year data provides a constraint on @xmath0 to lie between @xmath6 and @xmath7 at the @xmath2 c.l . @xcite .
@xcite obtains more stringent constraint @xmath8 . on the other hand , @xcite recently reported a detection of primordial non - gaussianity at greater than 99.5% significance .
further detailed analyses of non - gaussianity is clearly necessary in order to reconcile and understand the various constraints and claimed detections .
different approaches to the study of non - gaussianity exploit different statistical properties and will be sensitive to different aspects of the behaviour of the pattern being tested . in general , there is no unique statistic to describe the non - gaussian nature of a sample in a complete manner .
a given method may have strong discriminatory power for one particular form of non - gaussianity , but this is not necessarily the case for all possible alternative distributions .
testing non - gaussianity therefore requires a battery of complementary techniques rather than a single approach .
it is particularly important to use different statistical approaches in the context of primordial non - gaussianity , because the physical mechanism responsible remains unknown .
furthermore , in the real world , issues including survey masks and inhomogeneous noise have to be taken into consideration .
different statistics may be sensitive to different systematics and foreground artefacts , so that complementary analysis using different statistics are essential for a robust detection .
analyses using different statistical methods are useful to validate or refute basic theoretical models and constrain model parameters more accurately .
the most commonly used statistic for non - gaussian analysis is the bispectrum ( or even trispectrum ) which focuses on information contained within three - point ( or four - point ) correlations .
other approaches represented by minkowski functionals ( mfs ) and genus statistics ( one of mfs ) utilize information concerning the integrated morphology and topology of the density structure , and are dependent on all order of correlation functions .
their robustness and generality therefore makes them ideal complements to standard correlation analyses . in this analysis ,
we focus on the _ local _ model of primordial non - gaussianity characterized by @xmath0 in equation ( [ eq : ngpotential2 ] ) .
@xcite show that the bispectrum is the optimal statistics in the estimation of @xmath5 and then other statistics ( e.g. , trispectrum ) are useless even if there are different foreground contaminations .
this is , however , only the case when the local model exactly describes the real universe .
other forms of non - gaussianity , different from the one purely characterized by @xmath0 , which may exist in a real observation , could make the fit of the theoretical estimation as a function of @xmath0 to the observation worse and also influence the estimation of @xmath0 among different statistics .
different statistical approaches are , therefore , still useful to test the assumed model of primordial non - gaussianity and also possible observational systematics by checking if they have a reasonable goodness of the fit to observations and thus give a consistent limit on @xmath0 compared to that that from the bispectrum . in this paper , we present a measurement of primordial non - gaussianity from the mfs of the wmap three - year temperature maps .
we apply perturbative formulae recently derived by @xcite to do the comparison with observations ; previous analyses rely on non - gaussian simulations @xcite .
the agreement of the theoretical predictions with non - gaussian simulations has already been established in sachs - wolfe limit @xcite . in this paper
, we apply non - gaussian simulations based on full radiative transfer computations and then demonstrate that the analytical predictions accurately reproduce the simulation results .
@xcite already derived an analytical formula for the genus statistic in the sachs - wolfe approximation to compare with wmap data .
our analysis takes more detailed physics into account and is consequently more accurately applicable to a wider range of scales .
observational effects ( including antenna beam pattern , inhomogeneous noise and the survey mask ) could be other sources of confusion . from a comparison with simulations including these observational issues , we find that the observational systematics are negligible to estimate the primordial non - gaussianity from wmap data directly using our method .
the organization of this paper is as follows . in [ sec : data ] the wmap three - year data studied here are briefly introduced . in [ sec : comp_sim ] we test whether the perturbative formulae well describe the mfs for the non - gaussian simulation maps even including the various observational effects mentioned above . in
[ sec : wmaplimit ] , we show the mfs for wmap three - year temperature maps compared with theoretical formulae and give constraints on @xmath0 .
[ sec : summary ] is devoted to a summary and the conclusions .
the cmb temperature maps derived from the wmap observation are pixelized in healpix format with the total number of pixels @xmath9 @xcite . in our analysis
, we use the maps for q , v and w frequency bands with @xmath10 .
the linearly co - added maps are constructed using an inverse weight of the pixel - noise variance @xmath11 , where @xmath12 denotes the pixel noise for each differential assembly ( da ) given in @xcite and @xmath13 represents the full - sky average of the effective number of observations per each pixel .
we adopt two maps with different combinations of frequency bands : v and w ( written as `` v+w '' ) and q , v and w ( written as `` q+v+w '' ) .
the co - added maps are masked with the _ kp0 _ galaxy mask including point - source mask provided by @xcite , which leaves @xmath14 of the sky available for the data analysis . in comparison with wmap observations to give constraint on @xmath0 in [ sec : wmaplimit ] , a @xmath15cdm cosmology is assumed with the cosmological parameters at the maximum likelihood peak from the wmap three - year data only fit @xcite : @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , and @xmath21 .
the amplitude of the primordial fluctuations has been normalized by the first acoustic peak of the temperature power spectrum , @xmath22 at @xmath23 @xcite .
the topology of random fluctuation fields is generally studied using their excursion sets , i.e. regions where the field exceeds some threshold level . in a two - dimensional random field such as a cmb temperature map ,
three mfs are defined : the fraction of area @xmath24 exceeding the threshold , the total circumference @xmath25 of all the entire excursion set , and the corresponding euler characteristic @xmath26 @xcite .
we measure mfs for cmb temperature maps as a function of the threshold density @xmath27 , defined as the temperature fluctuation @xmath28 normalized by its standard deviation @xmath29 . based on the general formalism of perturbation theory for mfs @xcite , @xcite derived perturbative formulae of the mfs as a function of the non - linear coupling parameter @xmath0 ( eq.[[eq : ngpotential2 ] ] ) .
the mfs are separately written with the amplitude and the function of @xmath27 as follows . @xmath30
the amplitude @xmath31 , which is determined only by the angular power spectrum @xmath32 , is given by @xmath33 @xmath34^j c_l w^2_l,\ ] ] where @xmath35 gives @xmath36 , @xmath37 , @xmath38 and @xmath39 represents the smoothing kernel determined by the pixel and beam window functions and any additional smoothing ( e.g. a gaussian kernel ) . in weakly non - gaussian fields , the function @xmath40 can be divided into the gaussian term @xmath41 and the non - gaussian term at lowest order @xmath42 : @xmath43 each term has the following form @xmath44 @xmath45\sigma_0 + { \cal o}(\sigma_0 ^ 2)\right\},\end{aligned}\ ] ] where @xmath46 represent the @xmath47-th hermite polynomials and the skewness parameters @xmath48 are given in equations [ 27 - 29 ] of @xcite .
the amplitude @xmath31 ( eq . [ [ eq : mfamp ] ] ) is not directly relevant to non - gaussianity but is dependent on the shape of @xmath32 .
we therefore concentrate on the non - gaussian term @xmath49 hereafter .
the quantity @xmath42 is the same as the relative difference of mfs , which are plotted in fig . 2 in @xcite , except for its normalization factor ; in this paper the difference of mfs is normalized by @xmath31 ( [ eq : mfamp ] ) , while the maximum value of mfs for gaussian fields is used in @xcite .
the above analytical formulae have already been found to match accurately the mfs for non - gaussian maps in sachs - wolfe limit ( appendix c in * ? ? ?
here we test them against non - gaussian simulations including the full radiative transfer function @xcite . as we mentioned in the introduction
, actual observations of cmb also involve different effects which may produce other confusions : the pixel window function , beam smearing , the inhomogeneous noise , survey mask and so on .
we include these observational effects into the simulations to check whether they could have a systematic effect on our topological measures .
the cosmology in the non - gaussian simulations is based on lambda cdm , but the cosmological parameters have slightly different values from wmap three - year best - fit ; @xmath50 , and @xmath51 .
the amplitude of primordial fluctuations is , however , set to be same as wmap three - year best - fit value @xmath22 .
observational effects related to wmap data are included as follows .
first we convolve the original simulation maps with the q+v+w co - added beam transfer function with inverse weight of the full - sky averaged pixel - noise variance in each da .
next we add independent gaussian noise realizations following the noise pattern co - added with the same weight .
the simulation map is then masked with the @xmath52 galaxy mask .
finally we smooth the simulation maps using a gaussian filter with a smoothing scale of @xmath53 , @xmath54.\ ] ] the mfs are sensitive to the resolution ( or smoothing ) scale of a density field and thereby we can obtain a variety of information from density fields using different levels of smoothing . the information extracted from varying smoothing scales is nevertheless limited because they are all derived from the same original field ; the smoothed fields are not independent . here
we focus on the field smoothed by three different smoothing scales 10 , 20 and 40 , where the limit on @xmath0 is sufficiently converged . to remove the effect of the survey mask near the boundary of the mask
, we only use the pixels more than @xmath55 away from the boundary .
the sky fraction used in the analysis for each smoothing scale is 41% for @xmath56 , 62% for @xmath57 and 73% for @xmath58 .
the mfs for the measured cmb temperature anisotropy are computed from the integral of the curvature of iso - temperature contour lengths ( the details are described in appendix a.1 . of * ? ? ?
the binning range of @xmath27 is set to be @xmath59 to @xmath60 with @xmath61 equally spaced bins of @xmath27 per each mf . this binning way produces well converged results irrespective of other choices of the range of @xmath27 and the number of bins .
we obtain the normalized mfs ( eq . [
[ eq : mfamp ] ] ) with the amplitude @xmath31 computed from @xmath32 of each realization .
then the residuals of the normalized mfs from gaussian predictions @xmath62 are calculated at each bin of @xmath63 , which denote a threshold value @xmath27 , a kind of mf @xmath64 , and a smoothing scale parameterized with @xmath53 or @xmath65 .
even when the mfs of gaussian realizations are computed , however , the function @xmath62 are not exactly equal to @xmath66 due to the effect of pixelization , survey mask and other numerical artifacts .
we therefore measure the deviations from the average of the measurements over gaussian realizations and subtract them as @xmath67_{\rm gaussian , mean}.\ ] ] in fig .
[ fig : cmbsim ] , we compare the analytical predictions of variance , skewness , and mfs with the measurements from the simulations for @xmath68 .
the simulated results are the average over 200 realizations and the error - bars represent the error for the average ( the sample variance divided by the square - root - of @xmath69 ) .
the averaged measurements for gaussian cmb maps are subtracted from those for non - gaussian maps in the simulated plots including variance and skewness as well as mfs ( see eq . [ [ eq : mfsub ] ] ) . in the plots ,
we adopt the gaussian maps which are generated from the same realizations of linear potential fields ( @xmath4 in eq .
[ 1 ] ) as the non - gaussian cmb maps .
the sample variances , represented by the error - bars , are cancelled very well in such plots so one can focus on the systematic effect of primordial enon - gaussianity .
the analytical formulae are found to agree with the simulations extremely well even including all observational effects .
this indicates that we can measure @xmath0 from direct comparison of the analytical formulae with observations without having to worry excessively about the presence of such systematics .
we also check that both the artificial systematics @xmath70_{\rm gaussian , mean}$ ] and covariance matrix are not strongly dependent on the details of cosmology .
these results are encouraging , but not unexpected : being based on integrated properties , the minkowski functionals are expected to be robust to such effects .
we have adopted a maximum likelihood method to estimate the best - fit value of @xmath0 and its associated uncertainty . in ` nearly gaussian ' fields , the distribution functions of @xmath71 are well described as multivariate gaussians .
the likelihood function of @xmath0 is , therefore , simply proportional to @xmath72 where @xmath73 is computed using the theoretical formulae ( eq . [ [ eq : delmf_pb ] ] ) as @xmath74 \sigma^{-1}_{\alpha\alpha ' } \nonumber \\ & \times & [ \delta v_{\alpha'}^{\rm ( obs)}-\delta v_{\alpha'}^{\rm ( theory)}(f_{\rm nl})],\end{aligned}\ ] ] where @xmath63 and @xmath75 denote the binning number of threshold values @xmath27 , different kinds of mf @xmath64 , and smoothing scale parameterized with @xmath53 or @xmath65 . the full covariance matrix @xmath76 is required because mfs are strongly correlated between different @xmath27 , different kinds of mf , and also different @xmath65 or @xmath53 .
we estimate the covariance matrix of mfs from 1000 gaussian simulation maps including the pixel and beam window function , _
kp0 _ survey mask , and inhomogeneous noise for wmap three - year maps .
the mfs contain information about fluctuations at different scales , so the results depend on the choice of window function . here , two different types of window functions are adopted ( in addition to the beam window functions ) .
one is a gaussian window function with the scale characterized by @xmath53 ( which is chosen to be sufficiently large compared with the pixel size ) .
the other is just the pixel window function in healpix format with a scale characterized by @xmath65 .
the multipole components with @xmath77 higher than @xmath78 are cut because they suffer from serious aliasing effects . before applying these ideas to the observational data , we check if our method based on @xmath79 analysis is valid using simulations with primordial non - gaussianity .
we apply the non - gaussian simulations to check that the likelihood function using the equation ( [ eq : covmatrix ] ) reproduces a valid probability distribution of the true value @xmath80 . here
we consider that the likelihood function of @xmath81 in each realization follows a gaussian distribution around the best - fit value @xmath82 as @xmath83 \nonumber \\ & & \\
\sigma_{f_{\rm nl}}&=&\left[\sum_{\alpha\alpha'}\frac{\partial v_\alpha}{\partial f_{\rm nl}}(\sigma^{-1})_{\alpha\alpha'}\frac{\partial v_{\alpha'}}{\partial f_{\rm nl}}\right]^{-1/2}\end{aligned}\ ] ] where the binning number @xmath63 represents a threshold @xmath27 for @xmath64-th mf at a given scale @xmath65 ( or @xmath53 if the gaussian smoothing is added ) and the covariance matrix @xmath76 is numerically estimated from gaussian simulations .
the function @xmath84 is independent of @xmath0 ( see equation [ [ eq : delmf_pb ] ] ) and thus the uncertainty @xmath85 is independent of @xmath0 . according to bayes theorem
, @xmath86 should distribute around @xmath80 in the same way : @xmath87 we estimate the distribution function of @xmath88 from 200 non - gaussian cmb simulated maps at a given @xmath89 and then compare with the equation ( [ eq : likefunc_fnl ] ) .
the simulated maps include observational effects represented by pixel and beam window functions , noise , and _
kp0 _ survey cut for wmap three - year data .
[ fig : fnlfit ] shows the theoretical predictions of @xmath90 at @xmath91 ( solid ) and 100 ( dotted ) . from the mfs for the combined maps at gaussian smoothing scales @xmath57 and @xmath92 where the uncertainty is @xmath93 .
the histograms show the distribution of @xmath86 from 200 ( non-)gaussian realizations .
the averages of the best fit values of @xmath0 from the simulations are respectively @xmath94 ( for @xmath91 ) and @xmath95 ( for @xmath96 ) .
the simulations reproduce the theoretical predictions of the likelihood function very well .
our method is thus well established to give constraints on @xmath5 from wmap three - year map .
the three mfs for the cmb temperature maps from wmap three - year data are respectively plotted with symbols in each column of fig .
[ fig : mf3year_ga ] ( left three columns for the `` q+v+w '' map and right three columns for the `` v+w '' map ) . in each column , the top panel shows the mf @xmath97 at a representative scale ( @xmath57 ) .
and the lower three panels illustrate @xmath42 at different @xmath98 and @xmath99 .
the perturbative formulae with the best - fit value of @xmath0 to each observed mf are plotted with lines .
the best - fit values and @xmath100 uncertainty are written in the left - bottom side of each panel . in top panels ,
all of the amplitude of observed mfs are found to be smaller than the theoretical estimations .
this comes from the deficit of the observed power at low @xmath77 which generates the larger amplitude of mfs determined by @xmath101 ( eq.[[eq : mfamp ] ] ) , as pointed out by @xcite .
[ fig : mf3year_th ] shows the same plot but for the mfs with the pixel window function only .
the top panel shows the mf at @xmath102 and the lower three panels illustrate @xmath42 for @xmath65=256 , 128 , and 64 .
it is interesting that all mfs at @xmath65=64 have large positive values of @xmath0 , though the significance is less than @xmath103 .
table [ tab : fnl3year_ga ] lists the best - fit values and the @xmath100 uncertainty of @xmath0 for each mf and their combined values at different sets of gaussian smoothing scales @xmath53 .
the @xmath100 uncertainty of @xmath0 is estimated from the range of @xmath0 with @xmath104 .
the minimum of chi - square @xmath105 and the goodness - of - fit @xmath106 are listed for each fit .
the results for the pixel window function only are shown in table [ tab : fnl3year_th ] .
the goodness - of - fit values are reasonable for all the fits , which means that the simple form of the primordial non - gaussianity ( equation [ [ eq : ngpotential2 ] ] ) well describes the behaviour of the observed mfs . in other words ,
present observations are too uncertain to allow the extraction of any further information about primordial non - gaussianity ( e.g. scale dependence of @xmath0 ) .
the constraint @xmath1 at 95% c.l .
is obtained from all mfs for the q+v+w co - added map at combined different gaussian smoothing scales of 10 , 20 and 40 arcmin .
a similar constraint is obtained from the mfs with pixel - window only as @xmath107 .
the results from q+v+w co - added map are consistent with the previous ones @xcite .
there is some friction ( but not disagreement ) between our results and those by @xcite ; our v+w analysis finds @xmath108 whereas they find @xmath109 . moreover our averaged @xmath0 decreases from q+v+w to v+w whereas their @xmath0 increases .
this is very interesting because there is the possibility that e.g. , foregrounds and point sources might be biasing one of the two results .
@xcite show in their analysis that these effects do not seem to contaminate the primordial bispectrum measurement significantly
. it will be then important to check their effect on the mfs statistics in order to verify if this can explain the differencies among the two results .
however the observed discrepancies show already how analyzing non - gaussianity using different statistics can provide additional interesting information . [
cols="^,^,^,^,^,^,^ " , ] [ tab : fnl3year_th ]
we have presented an analysis of mfs for wmap the three - year temperature maps to limit the primordial non - gaussianity characterized by the nonlinear coupling parameter @xmath0 . to do this
we compared perturbative formulae for mfs of weakly non - gaussian fields directly with the observations .
the analytical formulae are found to be in excellent agreement with results from non - gaussian simulations of cmb maps including full radiative transfer effects .
the agreement is still very good when including systematic observational effects including the _ kp0 _ survey mask , pixel and beam window functions , and inhomogeneous noise distribution for wmap three - year data .
we have performed a @xmath79 analysis to the comparison of the analytical formulae with wmap three - year data .
the fits of the analytical formulae to the observations are acceptable and we thus obtain a robust constraint of @xmath1 at 95% c.l . from the q+v+w coadded maps with gaussian filter at different scales 10 , 20 and 40 combined .
the result is consistent with previous results @xcite .
the behaviour of the results for the v+w maps raises some interesting issues ; our constraint is negatively shifted @xmath110 while @xcite find a more positive range @xmath111 .
the difference between the two results should be clearer in the near future survey represented by planck .
it is worth investigating this result in further detail through a careful analysis of foregrounds and point source effects .
this will be the subject of future work .
we thank the anonymous referee for providing very useful comments and suggestions .
we deeply appreciate eiichiro komatsu who originally proceeded with the project together . c. h. acknowledges support from the particle physics and astronomy research council grant number pp / c501692/1 . c.h . also acknowledges support from a jsps ( japan society for the promotion of science ) fellowship .
t. m. acknowledges the support from the ministry of education , culture , sports , science , and technology , grant - in - aid for scientific research ( c ) , 18540260 , 2006 , and grant - in - aid for scientific research on priority areas no .
467 `` probing the dark energy through an extremely wide & deep survey with subaru telescope '' .
s. m. acknowledges asi contract planck lfi activity of phase e2 , for partial financial support . 99 alishahiha m. , silverstein e. , tong d. , 2004 , phys .
d. , 70 , 123505 arkani - hamed n. , creminelli p. , mukohyama s. , zaldarriaga m. , 2004 , jcap , 4 , 1 bartolo n. , matarrese s. , riotto , a. , 2002 phys . rev .
d 65 , 103505 bartolo , n. , komatsu , e. , matarrese , s. , & riotto , a. , 2004 , phys .
rept . , 402 , 103 battefeld d. , battefeld t. , 2007 , jcap , 5 , 1 bennett , c. l. et al . , 2003b , apjs , 148 , 97 bernardeau f. , uzan , j .-
, 2002 , phys .
d. , 66 , 103506 chen x. , richard e , eugene a. l. , 2007 , jcap , 6 , 23 coles p. , 1988 ,
mnras , 234 , 509 creminelli , p. , senatore , l. , zaldarriaga , m. , & tegmark , m. , 2006 , jcap , 03 , 005 creminelli , p. , senatore , l. , & zaldarriaga , m. , 2007 , jcap , 03 , 019 dvali g. , gruzinov a. , zaldarriaga m. , 2004 , phys .
d. , 69 , 083505 hikage , c. , komatsu , e. & matsubara , t. , 2006 , apj , 653 , 11 g@xmath112rski , k. m. , hivon , e. , banday , a. j. , wandelt , b. d. , hansen , f. k. , reinecke , m. , & bartelman , m. , 2005 , apj , 622 , 759 gott iii , j. r. , colley , w. n. , park , c. g. , park , c. , & mugnolo , c. , 2007 , mnras , 377 , 1668 hinshaw , et al . , 2007 ,
apjs , 170 , 288 komatsu , e. & spergel , d. n. , 2001 , phys .
d , 63 , 63002 komatsu , e. et al . , 2003 , apjs , 148 , 119 koyama , k. , mizuno , s. , vernizzi , f. , wands , d. , 2007 , jcap , 11 , 24 liguori , m. , matarrese , s. , moscardini , l. , 2003 , apj , 597 , 57 liguori , m. , hansen , f. k. , komatsu , e. , matarrese , s. , & riotto , a. , 2006 , phys .
d , 73 , 043505 liguori , m. , yadav , a. , hansen , f. k. , komatsu , e. , matarrese , s. , wandelt , b. , 2007 , phys .
d , 76 , 105016 lyth d.h .
, ungarelli c. , wands d. , 2003 , phys .
d. , 67 , 23503 matsubara , t. , 2003 , apj , 584 , 1 spergel d.n .
et al . , 2007 ,
apjs , 170 , 377 yadav , a. p. s. , & wandelt , b. , d. , 2008 , phys .
lett . , 100 , 181301
|
we present an analysis of the minkowski functionals ( mfs ) describing the wmap three - year temperature maps to place limits on possible levels of primordial non - gaussianity .
in particular , we apply perturbative formulae for the mfs to give constraints on the usual non - linear coupling constant @xmath0 .
the theoretical predictions are found to agree with the mfs of simulated cmb maps including the full effects of radiative transfer .
the agreement is also very good even when the simulation maps include various observational artifacts , including the pixel window function , beam smearing , inhomogeneous noise and the survey mask .
we find accordingly that these analytical formulae can be applied directly to observational measurements of @xmath0 without relying on non - gaussian simulations .
considering the bin - to - bin covariance of the mfs in wmap in a chi - square analysis , we find that the primordial non - gaussianity parameter is constrained to lie in the range @xmath1 ( @xmath2 c.l . ) using the q+v+w co - added maps .
cosmology : early universe cosmic microwave background
methods : statistical analytical
|
the observational evidence for dark matter indicates the intriguing possibility of a `` dark sector '' extension to the standard model ( sm ) .
dark matter in fact may be a small part of the dark sector or indeed many dark sectors could exist , each with their own `` dark forces '' and constituent particles .
dark matter may be accompanied by hereto unknown gauge bosons ( `` dark force '' carriers , ) which can couple dark matter particles and ordinary particles with exceptionally weak couplings .
modern colliders can be blind to such new forces , even though the mass of the `` dark force '' carriers can be quite small .
this is because the cross - sections of relevant processes for ordinary matter are so small that the `` dark force '' events are simply statistically insignificant and are discarded in high - energy experiments .
dark sector light weakly - coupled particles that interact with ordinary matter have been proposed as explanations of astronomical anomalies @xcite as well as discrepancies between calculated and measured muon magnetic moment @xcite .
such interactions would be inevitably below the weak force scale , ergo , the dark sector has so far escaped detection .
there are several proposed inroads into the detection of weakly - coupled particles and their associated dark forces @xcite .
one such example is the dark photon @xcite that is hypothesized to be a massive particle which couples to electromagnetic currents just like the photon does . in addition , dark z bosons have been proposed @xcite that couple to the weak neutral currents , ( i.e. , their interactions are parity violating . ) in a sense dark photons are massive photons while dark z bosons are light - versions of z bosons . from here on we will refer to this type of dark force mediator particle as the light gauge - boson . motivated by such `` dark force '' ideas here we place constraints both on couplings and masses of dark force carriers ( light gauge bosons ) by reinterpreting results of experiments on searches for permanent electric dipole moments ( edms ) of diamagnetic atoms .
specifically we focus on dark forces generated by the p , t - odd interaction of electrons and nucleons through the exchange of a massive light gauge boson .
we will refer to the carrier as @xmath1 .
effectively , the usually - employed contact interactions are replaced with yukawa - like interactions . standard model ( sm ) predicts the existence of intrinsic permanent electric dipole moments ( edm ) in particles as varied as quarks , leptons , and baryons .
these sm predictions , however , are below the current levels of experimental accuracy . as an example , in the sm framework , the electron edm is estimated to be of the order of @xmath2 ( see e.g. , ref @xcite ) while the most stringent experimental limit stands at @xmath3 from the tho molecular search @xcite . remarkably , however , there are many theoretical extensions to the sm that predict edm values comparable to the present experimental constraints .
overall , the searches for atomic edm can be classified into two major categories : edm of paramagnetic atoms and molecules and diamagnetic atoms .
paramagnetic atoms , such as tl and cs , have an unpaired valence electron and the atomic edm in this category is attributed to the edm of the unpaired electron .
diamagnetic atoms , on the other hand , are closed - shell atoms . in discussions of diamagnetic atomic edms , the edm is usually associated with the intrinsic edm of an unpaired nucleon ( schiff moment or p , t - odd electron - nucleon interactions ) . the best limit on a diamagnetic atom so far is @xcite @xmath4 while we will use the hg edm result for putting constraints on the light mediators , the formalism and derived analytical expressions are applicable to other diamagnetic systems , such as the atoms of current experimental interest : xenon @xcite , ytterbium @xcite , radon @xcite , and radium @xcite .
we start by reviewing the structure of contact interactions formed out of products of bi - linear forms .
the entire set of ten unique semi - leptonic lorentz - invariant products is tabulated in ref .
@xcite . in this paper
we focus on the most commonly used parity , time - violating tensor current term ( see , e.g. , @xcite ) @xmath5(\mathbf{r } ) [ \bar{\psi}_{n } i g'^p_n \gamma_{5 } \psi_{n}](\mathbf{r})\label{eq : ptoddsp } \,,\\
% w_{p_e s_n}=&\int d\mathbf{r}[\bar{\psi}_{e}i g'^p_e \gamma_{5 } \psi_{e}](\mathbf{x } ) [ \bar{\psi}_{n } i g'^s_n \psi_{n}](\mathbf{r})\label{eq : ptoddps } \,,\\
w_{t_e t_n}= & \int d\mathbf{r } \varepsilon_{\chi \lambda \mu \nu}[\bar{\psi}_{e}i
g'^t_e \sigma_{\chi\lambda } \psi_{e}](\mathbf{r } ) [ \bar{\psi}_{n } i g'^t_n \sigma^{\mu \nu } \psi_{n}](\mathbf{r})\label{eq : ptoddtensor } \ , .\end{aligned}\ ] ] here @xmath6 are the electron / nucleon dirac bi - spinors and @xmath7 are their adjoints , while @xmath8 are coupling constants .
further , @xmath9 , where @xmath10 s are dirac matrices and @xmath11 is the four - dimensional levi - civita tensor . from eq.([eq :
ptoddtensor ] ) one could easily read off the interaction hamiltonians acting in the electron space by removing averaging over @xmath12 and `` rubbing off '' @xmath6 and @xmath13 , @xmath14_e [ \bar{\psi}_{n}(\mathbf{r}_e ) i g'^t_n \sigma^{\mu \nu } \psi_{n}(\mathbf{r}_e ) ] \label{eq : h - ptoddtensor } \
, , \ ] ] where the subscript @xmath15 emphasizes that the operators inside @xmath16_e$ ] act on the electron degrees of freedom .
the interaction ( [ eq : ptoddtensor ] ) is of the contact nature , i.e. , it is constructed in the limit of the infinite mass of the carrier .
for the finite mass @xmath17 of the mediator @xmath1 the interaction need to be modified by sandwiching the currents with the yukawa - type interaction ( see e.g. , ref .
@xcite ) @xmath18 the `` upgraded '' eq .
( [ eq : ptoddtensor ] ) reads @xmath19(\mathbf{r } ) \times \nonumber \\ & v_\chi ( \mathbf{r},\mathbf{r ' } ) [ \bar{\psi}_{n } i g'^t_n \sigma^{\mu \nu } \psi_{n}](\mathbf{r ' } ) \label{eq : ptoddtensorchi}\end{aligned}\ ] ] or @xmath20_e \times \label{eq : h - ptoddtensorchiveri } \\ & \int d\mathbf{r}_n v_\chi ( \mathbf{r}_e,\mathbf{r}_n ) [ \bar{\psi}_{n}(\mathbf{r}_n ) i g'^t_n \sigma^{\mu \nu } \psi_{n}(\mathbf{r}_n ) ] \ , .
\nonumber\end{aligned}\ ] ] it is easily verified that in the limit of large propagator mass , @xmath17 , the yukawa potential in the above equation becomes @xmath21 recovering eq.([eq : ptoddtensor ] ) .
we add a superscript @xmath1 to the interaction , ( @xmath22 ) to distinguish between the contact and finite - range interactions .
the structure of the expression ( [ eq : h - ptoddtensorchiveri ] ) suggests that the nuclear property @xmath23 $ ] is `` carried out '' beyond the nucleus by the yukawa potential .
thus one anticipates the p , t - odd forces would `` leak out '' of the nucleus on characteristic distances
@xmath24 equal to the compton wavelength of the mediator .
we are interested in the atomic permanent electric dipole moments of diamagnetic systems induced by the p , t - odd semi - leptonic interaction .
the induced edm @xmath25 of the atomic state @xmath26 of energy @xmath27 can be expressed as @xmath28 where c.c .
stands for the complex conjugate of the preceding term and @xmath29 and @xmath30 are the atomic energies and wave functions . @xmath31
and @xmath32 is the operator of electric dipole moment for atomic electrons and the sum is over all atomic electrons .
tensor interaction ( [ eq : h - ptoddtensorchiveri ] ) can be simplified further @xcite as the nucleon motion can be well approximated as being non - relativistic .
the result is @xmath33_e \cdot \label{eq : h - ptoddtensorchi } \\
\sum_{\rm nucleons } \int d\mathbf{r}_n & v_\chi ( \mathbf{r}_e,\mathbf{r}_n ) i g'^t_n \bm
{ \sigma}_n \psi^\dagger_{n}(\mathbf{r}_n ) \psi_{n}(\mathbf{r}_n ) \ , , \nonumber\end{aligned}\ ] ] i.e. , it is proportional to the linear combination of weighted scalar products between nucleon and electron spins . here
we explicitly introduced the summation over the nucleons .
we further define @xmath34 since @xmath35 is the contribution of an individual nucleon to the nuclear density @xmath36 .
the two sides of this equation can be related from nuclear structure calculations ( see , e.g. , ref .
@xcite ) which would define the constant @xmath37 .
thereby , the effective form of the interaction can be represented as a scalar product of the nuclear spin and a rank-1 irreducible tensor acting in the electron space @xmath38 with @xmath39 where we introduced the parameterization @xmath40 in terms of the fermi constant , @xmath41 , and the mass - dependent coupling constant @xmath42 . in the contact approximation
this parameterization recovers the conventional form of the semi - leptonic operator ( @xmath43 @xmath44 the finite - mass operator ( [ eq : t - y ] ) can be recast in the form analogous to the above equation , @xmath45 by introducing the effective `` yukawa - weighted '' nuclear density @xmath46 the essential difference between the infinite - mass ( [ eq : t - c ] ) and the finite - mass ( [ eq : t - y - rho ] ) cases is the replacement of the nuclear density @xmath47 with the effective nuclear density @xmath48 .
this effective nuclear density has been introduced earlier in ref .
@xcite in the context of atomic parity violation mediated by a light gauge boson .
it also plays an important role in our analysis . for a uniform nuclear distribution contained inside a sphere of radius @xmath49 ( i.e. , @xmath50 )
the effective nuclear density can be evaluated analytically ( see appendix ) , @xmath51 & e^{-\frac{r}{\lambda_\chi}}\left(\frac{r}{\lambda_\chi } \cosh(\frac{r}{\lambda_\chi})-\sinh(\frac{r}{\lambda_\chi } ) \right),\ , r > r. \nonumber \end{cases}\end{aligned}\ ] ] notice that outside the nucleus , @xmath52 , i.e. , as expected , the interaction ( [ eq : t - y ] ) would sample electronic cloud at distances @xmath53 beyond the nuclear edge .
the values of @xmath54 are ( 4,2,0.2 ) fm for @xmath17 = ( 50,100,1000 ) mev / c@xmath55 . in fig .
( [ fig : mercdens ] ) we plot @xmath48 for these mediator masses for the @xmath56 nucleus .
the tendency of the effective density @xmath48 to further `` leak out '' of the nucleus as the @xmath17 values are decreased is apparent .
when the range of the force , @xmath54 , is comparable to an atomic size ( @xmath57 ) , the interaction would extend over the entire atom and would sample atomic shell structure .
this happens at the characteristic value of @xmath58 .
, the effective `` yukawa - weighted '' nuclear density for the @xmath56 nucleus is shown as a function of the radial distance @xmath59 for specified values of @xmath17 , the mass of the carrier .
the nuclear cut - off radius is @xmath60 fm .
[ fig : mercdens],scaledwidth=50.0% ]
now we focus on the atomic - structure aspect of the problem .
@xmath56 is an 80-electron closed - shell system , with the electron configuration [ xe ] @xmath61 .
one needs to evaluate the induced atomic edm ( [ eq : edm - general ] ) with the finite - mass mediator interaction ( we refer to its value as @xmath62 ) and compare it with the contact - interaction result @xmath63 .
in particular we focus on the ratio @xmath64 we employ two atomic - structure methods to evaluate this ratio : dirac - hartree - fock ( dhf ) and relativistic random - phase approximation ( rrpa ) .
rrpa improves upon dhf s independent - particle approximation by including major correlation effects .
both methods are _ ab initio _
relativistic , as they are based on solutions of the dirac equation .
the relativistic approach is important especially for large carrier masses for which the interaction is lumped in the nuclear region where the atomic electrons move at relativistic velocities .
while there are more advanced techniques available @xcite , the dhf and rrpa methods should provide an adequate qualitative understanding of how the atomic edm responds to the finite - range forces . in the independent - particle approximation ( dhf )
, the induced atomic edm ( [ eq : edm - general ] ) becomes @xmath65 where the summation is carried over atomic orbitals @xmath66 and @xmath67 .
@xmath66 are core orbitals occupied in the ground state @xmath26 and @xmath67 are un - occupied ( excited or virtual ) orbitals .
@xmath68 are the dhf energies of these orbitals .
each orbital @xmath69 is characterized by the principle quantum number @xmath70 , relativistic angular momentum number @xmath71 and magnetic quantum number @xmath72 .
@xmath71 encodes the total angular momentum @xmath73 and the orbital angular momentum @xmath74 .
the reduced matrix elements of @xmath75 are @xmath76 here @xmath77 and @xmath78 are the radial large and small components from the parameterization @xmath79 with @xmath80 being the spinor spherical harmonics .
numerical procedure can be described as follows .
first we solve the dhf equations for the ground state of hg atom using the finite - differencing techniques @xcite .
next , we use the obtained dhf self - consistent potential to construct a finite basis set of atomic orbitals using the dual - kinetic - ballance @xmath81-spline technique @xcite .
this set of basis functions is finite and numerically complete ( i.e. , excited and continuum states are included in the set . ) with such a set , the summation over atomic orbitals in eq.([eq : edm - angular ] ) becomes a straightforward exercise . in a typical calculation
we use a set of basis functions expanded over 80 @xmath81-splines of order 9 , in a cavity of spherical radius of 30 bohr and a 1000-point grid , out of which 68 points reside inside the nucleus , providing adequate numerical accuracy for both large and small carrier masses . compared to the dhf method
, the more sophisticated rrpa approach accounts for a linear response of an atom to a perturbing interaction ( @xmath82 ) . as a result of solving the rrpa equations @xcite using the described dhf basis set we determined a quasi - complete set of particle - hole excited states and their energies , required for evaluating the sum over intermediate states in the edm expression ( [ eq : edm - general ] ) .
the developed codes are an extension of the dhf and rrpa codes of ref .
we start by describing numerical results for the ratio , @xmath83 ( [ eq : ratio ] ) , and then present analytical formulae . our calculated ratios @xmath83 are plotted in fig .
[ fig : ratio ] as a function of the carrier mass @xmath17 . for @xmath0hg atom on the carrier mass @xmath17 .
, scaledwidth=50.0% ] to test the quality of the developed code , we first perform edm calculations with the usual contact interaction of eq.([eq : t - c ] ) , along with the nuclear fermi distribution , in the dhf approximation .
the resulting edm , @xmath84 , recovers the earlier result @xcite obtained in the same approximation .
next , to simplify the integration in eq .
( [ eq : t - y - reduced ] ) , we replace the fermi - type distribution with the uniform nuclear distribution , shown with the solid line in fig .
[ fig : mercdens ] . such calculation yields @xmath85 slightly deviating from the quoted fermi - distribution value .
since all our calculations of the ratio ( [ eq : ratio ] ) are carried out with the uniform nuclear density distribution , we fix this value as the infinite - carrier - mass value @xmath63 .
this result is consistent with the earlier value @xcite next we perform the edm calculations in the rrpa approximation .
our calculation with the uniform nuclear density distribution yields @xmath86 , agreeing with ref .
@xcite value .
the results of our numerical calculation of the ratio @xmath83 are plotted in fig .
[ fig : ratio ] as a function of the carrier mass @xmath17 .
the ratio tends to zero for small masses .
it monotonically increases to unity as the mass increases , as in this limit the effective `` yukawa - weighted '' nuclear density @xmath48 approaches the true nuclear density @xmath47 ( see fig .
[ fig : mercdens ] ) , thereby @xmath87 and @xmath88 . for small masses
we clearly observe a constant slope on the log - log plot .
we will comment on this scaling law below . in general ,
the rrpa and dhf results are in a good agreement for large carrier masses ( about 4% agreement for @xmath89 ) .
the difference between two approaches starts to grow larger as with decreasing @xmath17 the force starts to probe the atomic shell structure bringing sensitivity to the details of treating electron - electron correlations .
the most drastic difference arises at @xmath90 when the dhf ratio becomes negative while the rrpa ratio remains positive .
we find that for sufficiently large masses , the entire dependence of the ratio @xmath93 on the carrier mass can be well approximated by taking only a single channel contribution in the edm sum over states ( [ eq : edm - angular ] ) .
this is the contribution from the excitation of the outer - most occupied orbital @xmath94 to the excited orbitals @xmath95 .
the @xmath96 orbital is the least bound leading to the smallest energy denominators .
moreover , as @xmath73 increases the electrons tend to reside less in the nuclear region due to increased centrifugal barrier thereby suppressing t , p - odd matrix elements .
this single - channel approximation is fully supported by our numerical experimentation @xcite .
the above observation motivates an analytical approach which consists in evaluating matrix elements of the p , t - odd interactions analytically .
we use the fact that the matrix elements are mostly accumulated in the region close to the nucleus . in this region
, the large and small radial components of atomic orbitals ( [ eq : wavefunction ] ) can be approximated as @xmath97 where @xmath98 is the effective principal quantum number , @xmath99 , with @xmath100 being the quantum defect .
@xmath101 is the effective screened charge felt by the electron , e.g. , for the valence orbital @xmath102 .
@xmath103 is the nuclear charge and @xmath104 .
these formulae were adopted from ref .
@xcite for our parameterization ( [ eq : wavefunction ] ) of atomic orbital bi - spinors .
notice that these expressions were obtained for a point - like nucleus and they are valid for radial distances @xmath105 where the nuclear charge can be considered unscreened .
now the reduced matrix element ( [ eq : t - y - reduced ] ) can be evaluated with the effective nuclear density ( [ eq : effectivedensity ] ) . while forming the ratio ( [ eq : ratio ] ) and limiting the summation to the single channel , we factor out the integrals @xmath106 which depend on the nuclear density .
such integrals do not depend on principal quantum numbers , energies , nor dipole matrix elements and the ratio can be simplified to @xmath107 with @xmath108 , because @xmath109 for both the @xmath110 and @xmath111 orbitals .
notice that this ratio does not depend on specific quantum numbers and it is valid as long as one of the excitation channels from an occupied orbital @xmath66 ( @xmath112 or @xmath113 ) is dominant .
this argument is applicable to all diamagnetic atoms of current experimental interest : xe , yb , hg , rn , and ra .
for the uniform nuclear density distribution we find the following formula ( @xmath114 ) @xmath115 it is expressed in terms of the generalized hypergeometric function @xmath116 and the exponential integral function @xmath117 . for small values of the argument ,
i.e. , in the limit @xmath118 ( yet @xmath91 ) @xmath119 { } \left ( \frac{r}{\lambda_\chi } \right)^{(\alpha z)^2 } \ , , \ ] ] where we also show the non - relativistic limit . in the opposite case , @xmath120 , @xmath121{}1 - 3(\alpha z)^2 \left ( \frac{\lambda_\chi}{r } \right)^2 \,,\ ] ] i.e. ,
as expected , for large carrier masses the interaction becomes increasingly contact . in this case
the mass scaling of the ratio from the above equation is @xmath122 we present the comparison between fully - numerical and analytical results in fig .
[ fig : analyticalvsnumerical ] for mercury atom .
now we would like to specify the region of validity of the formula ( [ eq : analyticratio ] ) .
first of all , the atomic wave functions ( [ eq : wavefunction ] ) were obtained using point - like nuclear charge distribution . in reality , the atomic orbitals are affected by the extended nuclear charge and inside the nucleus the relevant product @xmath123 instead of the `` softer '' dependence @xmath124 ( @xmath125 ) used in the integral ( [ eq : integral ] ) . thus ( [ eq : analyticratio ] ) would tend to de - emphasize the nuclear region .
another limitation comes from the fact that the approximate wave functions are valid only in the region @xmath105 .
this places constraints on the compton wavelength of the force mediator @xmath91 , translating into @xmath126 or @xmath127 for @xmath0hg , consistent with fig .
[ fig : ratio ] . ) results for the yukawa - to - contact - interaction edm ratio @xmath93 for @xmath0hg atom .
the ratio is plotted as a function of the carrier mass .
, scaledwidth=50.0% ] , eq . ( [ eq : analyticratio ] ) for xenon ( @xmath128 , black solid line ) , mercury ( @xmath129 , green dashed line ) , and radon ( @xmath130 , dotted red line ) as a function of the ratio of the nuclear radius @xmath49 to the compton wavelength of the force carrier .
notice that eq .
( [ eq : analyticratio ] ) and the plotted curves hold only for @xmath131 . [
fig : ratio - xe - hg - rn],scaledwidth=45.0% ] fig .
[ fig : ratio - xe - hg - rn ] compares ratios @xmath93 computed with eq .
( [ eq : analyticratio ] ) for xenon ( @xmath128 ) , mercury ( @xmath129 ) , and radon ( @xmath130 ) . for
smaller @xmath103 the ratio tends to start off steeper at smaller masses @xmath17 and saturates earlier for larger values of @xmath17 .
it is worth emphasizing that eq .
( [ eq : analyticratio ] ) and the curves on fig .
[ fig : ratio - xe - hg - rn ] hold only for @xmath131 . to extend the analytical treatment into the region of sub - kev masses , we notice that when the force range is much larger than the atomic size ( @xmath132 ) , the yukawa potential ( [ eq : yukawa ] ) becomes coulomb - like since @xmath134 .
in this case the interaction is no longer resides near the nucleus , and the single - channel approximation introduced in sec .
[ sec : analyticalabovemev ] may break down .
however , we may still find the mass dependence of the ratio analytically . indeed , for @xmath132 , the effective nuclear distribution is simply ( see appendix ) @xmath135 & \frac{1}{3 } \ , \frac { r }
{ r } , \ , r > r \ , .
\label{eq : effectivedensitylimhugelambda } \end{cases}\end{aligned}\ ] ] thereby we may simply factor out the entire mass dependence from the ratio and for very low masses , @xmath136 , @xmath137 where the mass - independent proportionality constant @xmath138 has to be evaluated with atomic - structure techniques , @xmath139 , with the effective density ( [ eq : effectivedensitylimhugelambda ] ) . for @xmath0hg
we find @xmath140 in the dhf approximation and * @xmath141 * in the more accurate rrpa method .
to summarize this section , the entire dependence of @xmath93 on @xmath142 can be described analytically with eq .
( [ eq : analyticratio ] ) for masses above @xmath143 mev and with eq .
( [ eq : rverylowmasses ] ) for @xmath17 below @xmath143 kev .
the values of @xmath93 in the transition region between these two limits depend on the atomic - shell structure of specific atom .
now with the computed yukawa - to - contact - interaction edm ratio @xmath93 , eq .
( [ eq : ratio ] ) , we proceed to placing constraints on the coupling strengths and masses of the light gauge bosons .
essentially , we require @xmath144 to place the limits on the coupling constant @xmath42 we use the ratio @xmath93 computed in the more sophisticated rrpa approach together with the the atomic edm for contact interaction taken from ref.@xcite .
the experimental limit @xcite on hg atom edm reads @xmath145 ( 95% c.l . ) .
thereby , @xmath146 the resulting exclusion region for @xmath147 is shown in fig .
[ fig : limits ] .
the exclusion region can be trivially extended to the lower masses using eq .
( [ eq : rverylowmasses ] ) ( basically continuing the straight line on the log - log plot , fig .
[ fig : limits ] ) . for higher masses
@xmath148 saturates to @xmath149 .
the bounds on @xmath148 become less stringent for lighter carriers due to the fact that as the range of the interaction becomes larger than the atomic size , the effective nuclear density scales down as @xmath150 , eq .
( [ eq : effectivedensitylimhugelambda ] ) , reducing the atomic edm enhancement factor . for a fixed experimental limit on edm
, this translates into larger values of the coupling constant @xmath148 . and
carrier masses @xmath17 .
this exclusion region is derived from computed yukawa - to - contact - interaction edm ratio @xmath93 in the rrpa method and the experimental limit @xcite on edm of @xmath151 atom .
[ fig : limits],scaledwidth=50.0% ]
the effective `` yukawa - weighted '' nuclear density is defined as @xmath152 where the yukawa potential @xmath153 can be expanded as @xmath154 where @xmath155 , @xmath156 . for a spherically - symmetric nuclear distribution the angular part of the integral in eq.([apeq : yukawand1 ] )
is reduced to @xmath157 i.e. , only the monopole contribution remains in eq.([apeq : yukawa_potentioal ] ) .
the bessel and hankel functions of imaginary arguments are the modified bessel functions : @xmath158 therefore , we can rewrite eq.([apeq : yukawand1 ] ) as @xmath159 for a uniform nuclear distribution contained inside a sphere of radius @xmath49 , ( i.e. , @xmath50 ) the integral yields @xmath51 & e^{-\frac{r}{\lambda_\chi}}\left(\frac{r}{\lambda_\chi } \cosh(\frac{r}{\lambda_\chi})-\sinh(\frac{r}{\lambda_\chi } ) \right),\ , r > r. \nonumber \end{cases}\end{aligned}\ ] ] notice that in the limit @xmath118 ( i.e. , for @xmath160 ) , @xmath161 & \frac{1}{3 } \left ( \frac{r}{\lambda_\chi}\right)^3\ , \frac { \lambda_\chi } { r } \,e^{-\frac{r}{\lambda_\chi } } , \ ,
r > r \ , .
\label{eq : effectivedensitylimlargelambda } \end{cases}\end{aligned}\ ] ] finally , in the limit @xmath132 ( i.e. , the range of the potential being much larger than the atomic size ) , we could drop the exponent in the second line of eq .
( [ eq : effectivedensitylimlargelambda ] ) @xmath135 & \frac{1}{3 } \ , \frac { r } { r } , \ , r > r \ , .
\label{eq : app : effectivedensitylimhugelambda } \end{cases}\end{aligned}\ ] ] 28ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) link:\doibase 10.1103/physrevd.79.015014 [ * * , ( ) ] link:\doibase 10.1103/physrevd.75.115017 [ * * , ( ) ] link:\doibase 10.1103/physrevd.80.095002 [ * * , ( ) ] http://arxiv.org/abs/1311.0029 http://arxiv.org/abs/1311.0029\nhttp://www.arxiv.org/pdf/1311.0029.pdf [ ( ) ] , \doibase http://dx.doi.org/10.1016/0370-2693(86)91377-8 [ * * , ( ) ] link:\doibase 10.1103/physrevd.89.095006 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1126/science.1248213 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.101601 [ * * , ( ) ] link:\doibase 10.1140/epjd
/ e2010 - 00044 - 5 [ * * , ( ) ] @noop * * , ( ) link:\doibase
10.1140/epjd / e2004 - 00175 - 2 [ * * , ( ) ] link:\doibase 10.1007/s10751 - 013 - 0898 - 2 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1016/j.nuclphysa.2010.05.013 [ * * , ( ) ] , @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1103/physreva.80.032120 [ * * , ( ) ] @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) http://arxiv.org/abs/1408.4337 [ ( ) ] , @noop _ _ ( , , ) link:\doibase 10.1016/j.cpc.2008.03.004 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) _ _ , @noop , ( )
|
postulating the existence of a finite - mass mediator of t , p - odd coupling between atomic electrons and nucleons we consider its effect on permanent electric dipole moment ( edm ) of diamagnetic atoms .
we present both numerical and analytical analysis for such mediator - induced edms and compare it with edm results for the conventional contact interaction . based on this analysis
we derive limits on coupling strengths and carrier masses from experimental limits on edm of @xmath0hg atom .
|
the last decades have seen the rise and success of the hierarchical paradigm for galaxy formation in a cold dark - matter dominated universe .
although very powerful , the concordance model is still far from providing us with a complete and coherent view of how galaxies form and evolve .
this is mainly because we still do not understand the physics involving the baryonic component .
the current challenge for galaxy formation and evolution studies is thus to improve our knowledge of the astrophysical processes responsible for _ transforming _ simple dark matter halos into the bimodal population of galaxies inhabiting today s universe .
it is in fact well established that , when we look at their integrated optical colours , galaxies constitute a bimodal population ( e.g. , @xcite ) composed of a ` red sequence ' , dominated by old stellar populations , and a ` blue cloud ' where the vast majority of new stars in the universe are formed .
however , the dichotomy in the colour distribution does not automatically reflect a difference in morphological type ( e.g. , light distribution ) and we now know that the red sequence is not only composed of quiescent early - type galaxies ( e.g. , @xcite ) .
for example , while the red sequence accounts for @xmath260 - 85% ( depending on the colour cut used to define star - forming galaxies ) of the total stellar - mass density in the local universe ( e.g. , @xcite ) , stars in late - type galaxies contribute to at least half the local stellar mass budget ( e.g. , @xcite ) .
this automatically implies that a significant fraction of massive late - type galaxies lie in the red sequence , whereas high - mass blue ellipticals are extremely rare .
moreover , not all red galaxies have stopped forming stars , as revealed by recent ultraviolet ( uv ) investigations ( e.g. , @xcite ) .
it thus emerges that , at least at optical wavelength , the red sequence is a heterogeneous family of objects which have likely followed different evolutionary paths .
how galaxies end - up in the red sequence is still partly a mystery , but recently high - redshift surveys have made it possible to start tracing the growth of the star - forming and quiescent galaxy population up to @xmath31 and beyond . despite the observational and theoretical uncertainties ( e.g. , @xcite )
, it seems now commonly accepted that the stellar mass of the blue cloud shows very little growth since @xmath31 , while the red sequence has grown by at least a factor @xmath22 ( e.g. @xcite ) .
the most popular scenario invoked to explain the growth of red galaxies is a migration of a significant fraction of star - forming systems from the blue cloud .
although not always supported @xcite , the possibility of an exchange of galaxies between the two sequences is exciting , and several theoretical and observational studies have started to look for the possible astrophysical processes responsible for such transition .
several mechanisms have been proposed so far , among the most popular are different modes of gas accretion @xcite , feedback from active galactic nuclei ( agn , @xcite ) , and environmental effects ( e.g. , @xcite , hereafter hc09 ) . however , whether a population of migrating galaxies does really exist and what causes the quenching of their star formation is still not clear . in this context , the advent of the _ galaxy evolution explorer _ ( galex ) large - area uv surveys is allowing us to tackle this problem from a different angle .
thanks to its high sensitivity to low - level star formation activity , uv magnitudes can be used to better discriminate between quiescent and still active ( although optically - red ) galaxies .
in fact , contrary to what is observed at optical wavelengths , the uv - optical colour distribution at a given mass is not well fitted by two gaussian distributions @xcite , but it shows a significant excess of objects in the region between the blue and red sequence ( i.e. , the ` so - called ' transition region or ` green - valley ' , @xcite ) .
transition galaxies may thus represent the missing link to understand if and how galaxies move from one population to the other .
however , it is worth reminding that , despite its potential , uv emission is significantly affected by dust and , only after accurate dust corrections , can the uv - optical colour be used to identify transition galaxies .
a significant fraction of galaxies found between the two sequences may in fact be composed of reddened systems @xcite . once transition galaxies are properly identified , a reconstruction of their past evolution is not straightforward .
the correct discrimination between various physical mechanisms able to suppress star formation requires , in theory , a detailed investigation of _ all _ the galactic components , i.e. , stars , gas and dust . of particular importance is the atomic gas content ( hi ) , which represents the fuel for the future star formation activity .
the mechanism responsible for the migration from the blue cloud must in fact inhibit the condensation of atomic into molecular hydrogen and the onset of star formation .
unfortunately , not only is hi astronomy still technically limited to the nearby universe ( e.g. , @xcite ) , but also our knowledge of hi properties of local galaxies is generally restricted to the blue cloud ( e.g. , @xcite ) . the very local universe ( e.g. , up to the distance of the virgo cluster ) is currently the only place where it is possible to investigate the link between hi - content and quenching of star formation in transition galaxies .
for all these reasons , we have collected uv to near - infrared imaging and hi 21 cm line data for a volume - limited sample of nearby galaxies covering different environments . in our previous paper ( hc09 ) , we have highlighted the power of a combination of uv and hi observations to understand the properties of transition galaxies .
our analysis suggested a strong relationship between uv - near - infrared colour and hi content showing that migrating spirals are mainly hi - deficient objects found in high density environments .
this result apparently rules out agn - feedback as the main mechanism responsible for the quenching of the star formation in nearby spirals .
however , a number of important questions still remain to be answered .
are transition galaxies really migrating from the blue to the red sequence ?
what are the time - scales of such migration ? is the quenching followed by a change in morphology ? while in hc09 we have shown that a large fraction of spirals outside the blue cloud is hi - deficient , this is not true for all transition spirals .
how have hi - rich systems reached the transition region ?
the aim of this paper is thus to extend the analysis presented in hc09 , in order to provide important constraints on the recent mass growth of the red sequence .
the paper is arranged as follows . in
2 we briefly describe the sample and discuss possible biases related to the dust extinction correction . in
3 we define the transition region and in 4 discuss the relation between colour and gas content .
the properties of transition galaxies are presented in 5 and their evolutionary histories and implications for galaxy evolution studies are discussed in
6 . finally , our main results are briefly summarized in
7 . throughout the paper we use h@xmath470
km s@xmath5 mpc@xmath5 . in the virgo cluster , where peculiar motions are dominant
, we use distances as determined in @xcite . star formation rates ( sfrs )
are computed from the nuv luminosities , following the conversions by @xcite .
the analysis presented in this paper is based on the k - band selected sample described in hc09 .
briefly , it consists of a volume - limited sample of galaxies having 2mass @xcite k - band magnitude k@xmath6 12 mag and distance between 15 and 25 mpc .
additional selection criteria are a high galactic latitude ( b @xmath7 + 55@xmath8 ) and low galactic extinction , a@xmath9 @xmath10 0.2 @xcite , to minimize galactic cirrus contamination .
the total sample contains 454 galaxies .
observations from the galex @xcite gr2 to gr4 data releases in the near- ( nuv ; @xmath11=2316 : @xmath12=1069 ) and far - ultraviolet ( fuv ; @xmath11=1539 : @xmath12=442 ) band were available for 394 and 325 objects , respectively . in the rest of the paper
we focus our attention mainly on the @xmath13 colour , given the larger number statistics available .
uv magnitudes were obtained by integrating the flux over the galaxy optical size , determined at the surface brightness of @xmath14(b ) = 25 mag arcsec@xmath15 .
the typical uncertainty in the uv photometry is @xmath20.1 mag and @xmath20.15 - 0.20 mag in nuv and fuv respectively .
stellar masses are determined from h - band luminosities using the @xmath16 colour - dependent stellar mass - to - light ratio relation from @xcite , assuming a @xcite initial mass function .
single - dish hi 21 cm line emission data , necessary for quantifying the hi content of galaxies , was mainly taken from @xcite , @xcite and the _ nasa / ipac extragalactic database _ ( ned ) . given the variety of sources from which the hi fluxes are taken , it is impossible to define a 21 cm sensitiveness limit for our sample .
however , as discussed in 4 ,
non detections start to be significant ( @xmath220% ) at @xmath17 - 1.3 and dominate ( @xmath750% ) for @xmath18@xmath1 - 2.1 .
estimates of atomic hydrogen mass or upper limits are available for @xmath283% ( 326/394 ) of the galaxies with nuv photometry .
we estimate the hi deficiency parameter ( @xmath19 ) as defined by @xcite : i.e. , the difference , in logarithmic units , between the observed hi mass and the value expected from an isolated galaxy with the same morphological type @xmath20 and optical linear diameter @xmath21 : @xmath22 .
we used the equation in @xcite for early - type galaxies ( e / s0 and earlier types ) and the four revised values presented by @xcite for late - types ( sa - sab , sb , sbc , sc and later ) to calculate the expected hi mass from the optical diameter .
the regression line coefficients are almost identical from sa to sbc types , whereas they both significantly vary going to e / s0 or to sc and later types @xcite .
the typical uncertainty in the estimate of @xmath19 is @xmath20.3 ( e.g. @xcite ) , but it might slightly increase for dwarf galaxies and early - type systems . given its large uncertainty , in the following , we will mainly use the hi deficiency to select those galaxies which have likely lost a significant amount of atomic hydrogen . in detail
, we use a threshold of @xmath230.5 to discriminate between ` hi - deficient ' and ` hi - normal ' galaxies . hi - deficient systems are thus objects with @xmath2470% less atomic hydrogen than expected for isolated objects of the same optical size and morphological type . the estimate of the internal uv dust attenuation @xmath25 is a crucial step for a correct interpretation of the colour - mass diagram .
this is particularly true outside the blue cloud where a ) evolved stellar populations can contribute to the dust heating @xcite , b ) a significant fraction of the uv emission may come from evolved stars and not from young stellar populations ( in particular in the red sequence , e.g. @xcite ) .
previous works have shown how dramatic the effect of inaccurate dust corrections can be on the transition region and red sequence @xcite . @xcite
have recently calibrated new recipes to estimate the dust attenuation , taking into account the contribution of evolved stellar population to the dust heating .
although this method provides more realistic dust corrections with an average error of @xmath20.5 mag in @xmath26 , the uncertainty dramatically increases up to @xmath21 mag for galaxies lying between the blue and the red sequence .
thus , the transition region can only be investigated using a statistical approach and a comparison between colours and sfrs of transition galaxies could be meaningless given such large uncertainties .
unfortunately , all dust correction recipes developed so far are calibrated on late - type / star - forming galaxies whereas in the case of early - type / quiescent systems ( the dominant population in the red sequence ) it is not a priori appropriate to apply such corrections .
the uv emission coming from old stellar populations should in fact be less affected by dust .
for example , a systematic ( but likely unreal ) shift of @xmath20.5 mag is applied to the red sequence if the extinction corrections calibrated on late - types are applied to early - type galaxies @xcite . in order to avoid any systematic overestimate of the sfr in early - type galaxies ( and also to be consistent with previous works ) here we corrected all galaxies for galactic extinction according to @xcite , but we applied internal extinction corrections only to late - type galaxies . in detail , the internal dust attenuation was determined using the total infrared ( tir ) to uv luminosity ratio method ( e.g. , @xcite ) and the age - dependent relations of @xcite .
the tir luminosity is obtained from iras 60 and 100 @xmath14 m fluxes or , in the few cases when iras observations are not available , using the empirical recipes described in @xcite .
although this technique is likely to overestimate the colour of the few early - type galaxies lying in the blue cloud , we can exclude that it significantly affects the properties of transition region early - type galaxies and peculiar gas - rich ellipticals studied in the following sections . to test this hypothesis
, we applied the dust attenuation corrections described above to these objects finding that , even after correcting for dust , transition early - types still lie in the transition region and peculiar red ellipticals are still in the red sequence .
however , in the following , we will try to discuss as much as possible any systematic error that could be introduced by problems in the extinction corrections and all error - bars shown in our figures take into account the uncertainty in the estimate of @xmath25 .
the morphological type distribution of galaxies in our sample ( dotted histogram ) .
the filled histogram represents galaxies in the transition region , while red and blue sequence galaxies are shown with the empty and dashed histogram , respectively .
, width=321 ]
contrary to what is observed in optical @xcite , at fixed stellar mass , the uv - optical colour distribution of galaxies is not best fitted with the sum of two gaussians .
an excess of galaxies is clearly present in between the two sequences @xcite .
this is likely due to galaxies with suppressed star formation perhaps migrating from the blue to the red sequence .
however , it is not straightforward to clearly define the range of colours characterizing galaxies which do not belong either to the blue cloud or the red sequence . on one side , it is difficult to determine ( both observationally and theoretically ) where the blue cloud ends . in this paper
we decide to use the blue cloud of hi - normal galaxies to determine the range of colours typical of unperturbed galaxies .
this is motivated by the fact , discussed in hc09 , that the transition region is mainly populated by hi - deficient galaxies in high - density environments .
hi - normal late - type galaxies form a blue cloud clearly separated from the red sequence at all masses , as shown in the right panel of fig .
[ cmall ] .
we thus define galaxies with suppressed star formation as those objects with @xmath274.5 mag ( corresponding to the 90th percentile of the colour of hi - normal spirals with m@xmath2810@xmath29 m@xmath30 ) , consistently with @xcite .
this colour - cut roughly corresponds to star formation histories ( sfhs ) having e - folding time - scale of @xmath23 gyr ( assuming a galaxy age of 13.7 gyr , solar metallicities and the models of @xcite ) e.g. , the typical e - folding time dividing local late- and early - type galaxies @xcite .
we note that , adopting this convention , m31 would be classified as ` normal ' blue - cloud galaxy ( @xmath314.1 mag ) .
being calibrated on massive galaxies , our colour - cut must be considered as a conservative upper - limit in the case of dwarf systems . on the other side ,
uv - to - near - infrared colours typical of the red sequence ( @xmath326 ) can either indicate low residual star formation activity or old , evolved stellar populations ( e.g. @xcite ) .
the phenomenon of the uv - upturn @xcite makes colours redder than @xmath336 difficult to interpret so that the @xmath13 colour can not be considered anymore as a good proxy for the specific star formation rate ( ssfr ) . following @xcite ,
we use observations of well known strong uv - upturn galaxies to derive a lower limit on the @xmath13 colour typical of evolved stellar populations . in details , given the typical colour observed in m87 ( @xmath31 6.1 mag ) and ngc4552 ( @xmath31 6.4 mag ) , we assumed @xmath13=6 mag as a conservative lower limit to discriminate between residual star formation and uv - upturn .
the validity of this colour - cut is confirmed by a visual inspection of galex colour images , which indicates that only 6% ( i.e. 6 objects ) of galaxies redder than @xmath13=6 mag show clear evidence of residual sf ( e.g. , blue star - forming knots ) . the morphological type distributions for galaxies belonging to the three groups here considered ( i.e. , blue cloud , red sequence and transition region ) are shown in fig .
[ typedistr ] .
it clearly emerges that red and blue galaxies are two disjoint families not only in colour , but also in shape . of course
, the criteria described above are arbitrary and vary according to the colour adopted and to the stellar mass range investigated .
this can be clearly seen in fig .
[ cmall2 ] , where the position of ` transition galaxies ' in different uv and optical colour - stellar - mass diagrams is highlighted .
although it is indisputable that not all transition galaxies are outside the red sequence in a @xmath34 and @xmath35 colour diagram ( in particular at low stellar - masses ) , it emerges that the definition here adopted is able to select a statistically representative sample of galaxies with suppressed star formation .
moreover , the comparison between the top and bottom row in fig.[cmall2 ] highlights the necessity of uv colours to select fair samples of transition galaxies : e.g. , a simple @xmath36 colour - cut would significantly contaminate our sample with star - forming blue - sequence and quiescent red - sequence systems .
finally , it is interesting to note the wide range in @xmath37 colour spanned by the transition galaxies , suggesting different current star formation rates . as shown in the following sections , the sample of transition galaxies selected using these criteria is not significantly contaminated either by active star - forming or quiescent galaxies erroneously classified as transition systems .
thus , in the rest of this paper we will refer to the colour interval 4.5@xmath06 mag as the ` transition region ' .
finally , it is worth reminding the reader that , in all the figures presented in this paper , the uv - near - infrared colour is directly related to the ssfr only outside the red sequence . for colours redder than @xmath316 mag , uv magnitudes can not be blindly used to quantify current sfrs and the presence of a sequence does not imply that all red galaxies have the same ssfr , as shown in the following sections .
in hc09 we showed that late - type galaxies outside the blue cloud appear to have lost at least @xmath270% of their atomic hydrogen content , when compared with isolated galaxies of similar size and morphology .
we commented this result as a strong evidence of a physical relation between the loss of gas and quenching of the star formation .
this interpretation is confirmed and reinforced in fig .
[ colgfrac ] .
the left panel shows the correlation between colour and hi deficiency : galaxies outside the blue cloud have not only lost a significant fraction of their atomic hydrogen content , but we also find a correlation between hi deficiency and @xmath13 colour , although with large scatter .
at least part of the scatter is due to the large uncertainty in the estimate of hi deficiency ( @xmath20.3 dex ) .
the same relation can be expressed in terms of gas - fraction ( here defined as the ratio of the hi to the stellar mass ) , as discussed by @xcite and shown in the central and right panels of fig .
[ colgfrac ] : lower gas - fractions correspond to redder colours .
the best linear fit to the relation ( excluding upper - limits ) is @xmath18=@xmath38 + 0.19 , with a dispersion of @xmath20.43 dex .
however , from the three panels in fig . [ colgfrac ] it clearly emerges that not all transition region and red - sequence galaxies are hi - deficient , but a number of systems have an amount of hydrogen typically observed in objects lying in the blue cloud .
this is particularly interesting if we look at the colour - gas - fraction relations ( central and right panels ) .
outside the blue cloud , galaxies lie mainly at the two edges of the relation depending on whether they are hi - deficient ( empty circles ) or not and the colour - gas - fraction relation appears more scattered .
this suggests that , for @xmath274.5 mag , the gas - fraction is not a good proxy of the uv - optical colour anymore and vice - versa .
the dispersion in the colour - gas - fraction relation increases from @xmath20.35 dex ( consistent with @xcite and @xcite ) to @xmath20.54 dex when we move from hi - normal blue - cloud galaxies to transition and red - sequence objects .
this is in reality a lower - limit on the real scatter increase since upper - limits are not included in the calculation .
the results shown in fig . [ colgfrac ] are strongly suggestive of a ) a different evolutionary path followed by hi - deficient and hi - normal galaxies outside the blue cloud and b ) of a weaker link between hi - content and colour than the one typically observed in star - forming galaxies .
therefore , in order to gain additional insights on the evolution of galaxies in the transition region , in the following we divide transition galaxies into two families according to their gas content and investigate separately their properties . before investigating the detailed properties of transition galaxies , it is worth adding a few notes about the validity of the classification for hi - normal transition systems .
the low number of objects in this category and the large uncertainties in the estimate of gas fractions and uv dust attenuation might suggest that these are just random outliers , not different from the bulk of the hi - deficient population . although we can not exclude the presence of a few misclassified galaxies in both the hi - deficient and hi - normal population
, it is very unlikely that all ( and only ) the hi - normal galaxies outside the blue sequence are affected by a large ( @xmath70.5 dex ) systematic underestimate of dust attenuation , and/or gas fraction .
more importantly , the analysis presented in the next sections will clearly show that these two families have reached the transition region following different evolutionary paths .
the stellar mass distribution of transition galaxies ( filled circles ) . left : spirals ( empty squares ) , lenticulars ( triangles ) and e+de ( asterisks ) are indicated .
right : galaxies are highlighted according to their hi content .
galaxies for which the estimate of hi deficiency is unsure are indicated with asterisks .
, width=321 ]
in total , 67 galaxies in our sample , corresponding to @xmath217% in both number and total stellar - mass lie in the transition region as defined in 3 . in fig .
[ trdistr ] , we show the stellar - mass distribution of transition galaxies divided according to their morphological type ( left panel ) and gas content ( right panel ) . for @xmath39 , galaxies with 4.5@xmath06 mag are mainly spirals , whereas at lower stellar masses they are preferentially dwarf elliptical systems .
more importantly , the majority of transition galaxies have @xmath4070% less atomic hydrogen content than isolated galaxies of similar optical size and morphological type .
however , as already noted in hc09 , not all galaxies in the transition region are hi - deficient . for @xmath39 , @xmath230% ( 8 galaxies ) of the transition galaxies have hi deficiency lower than 0.5 . for lower stellar masses , it is difficult to quantify the number of gas - rich objects .
hi observations are not available for 9 of our galaxies and for 3 additional objects the lower limits obtained for the hi deficiency are below our threshold of 0.5 . these are mainly de cluster galaxies ( fig .
[ trdistr ] , left panel ) , suggesting that their evolution is related to the cluster environment @xcite .
however , to be conservative , in the following analysis we will focus our attention on the 55 galaxies for which the classification as hi - deficient or hi - normal galaxy is reliable . overall , sure hi - deficient galaxies represent @xmath270% ( 47 galaxies ) in number and @xmath263% in stellar mass of the transition region88% in case all galaxies without hi measurement are hi - deficient galaxies .
we note , however , that these values may not be representative of the local universe , being our sample likely biased towards high - density environments ( see 6 ) . ] .
all except four galaxies lie in the virgo cluster region ( as defined in @xcite ) suggesting that the cluster environment is playing an important role in quenching the star formation .
additional support to this scenario is obtained when we consider the properties of galaxies divided according to their gas content and their position in the colour - mass diagram .
we compared the median projected distance from the cluster center of the different populations .
given the large asymmetry of virgo and the presence of two main sub - clusters ( virgo a and virgo b , at @xmath21 virial radii projected - distance ) , for each galaxy we computed the projected - distance from the center of both clouds and adopted the smallest of the two values .
our results do not qualitatively change if just the distance from m87 is adopted .
the median cluster - centric distance decreases from @xmath20.83 virial radii ( r@xmath41 ) , in case of hi - normal blue - cloud galaxies , to 0.51 , 0.52 and 0.42 r@xmath41 for hi - deficient blue - cloud , hi - deficient transition and red - sequence objects , respectively ( see fig . [ cluster ] , left panel ) .
similarly , the difference between the 25th and 75th percentiles of the line - of - sight velocity distribution ( i.e. a good estimate of the velocity dispersion in case of non gaussian distributions ) increases in the blue cloud from @xmath2760 km s@xmath5 to @xmath21400 km s@xmath5 when we consider hi - normal and hi - deficient galaxies respectively .
then , the typical velocity dispersion gradually decreases to @xmath2915 km s@xmath5 and @xmath2610 km s@xmath5 when we consider the transition region and the red sequence respectively ( see fig .
[ cluster ] , right panel ) .
the gradual variation in projected distance and velocity distribution when moving in the colour magnitude diagram from blue , hi normal systems to red , quiescent objects supports the idea that hi - deficient galaxies represent a population of galaxies recently infalled into the cluster and not yet virialized .
for example , a free - falling population is expected to have a velocity dispersion @xmath42{2}$ ] times larger than the virialized population . moreover ,
the velocity dispersion profile for hi - deficient galaxies decreases with cluster - centric distance consistent with isotropic velocities in the center and radial velocities in the external regions , as expected in the case of galaxy infall onto the cluster @xcite .
the opposite trend ( i.e. , increasing with cluster - centric distance ) is observed for red - sequence galaxies , as expected in a relaxed cluster undergoing two - body relaxation in the dense central region , with circular orbits in the center and more isotropic velocities in the external regions .
finally , a visual investigation of uv images reveals that in at least 50% of star - forming virgo galaxies in the transition region the star formation is only present well within the optical radius , completing the collection of evidence supporting environmental effects behind the quenching of the star formation in hi - deficient virgo galaxies .
less clear is the origin of the hi deficiency in galaxies outside virgo : 4 objects in total , namely ngc4684 , ugc8756 , ugc8032 and ngc5566 .
ngc5566 is the brightest member of a galaxy triplet while ugc8032 lies just @xmath21.1 virial radii from the center of virgo .
thus for these two galaxies it is still possible that environmental effects are playing a role in the gas stripping .
the fact that ugc8032 is not included in our virgo sample despite its small distance from m87 is due to the fact that it just lies outside the virgo boundaries defined by @xcite .
the origin of the hi deficiency in ngc4684 and ugc8756 remains a puzzle .
ugc8756 has in fact no nearby companions or any clear sign of interaction .
ngc4684 is a lenticular galaxy with very strong uv nuclear emission , probably related to the extended h@xmath43 outflow discovered by @xcite .
the outflow has been interpreted as related to bar instability and it is not clear whether such process can be responsible for the hi deficiency observed in this object .
thus , the origin of the hi deficiency in these two objects still remains unclear .
combining the observational evidence presented above with previous works on the virgo cluster ( e.g. @xcite and references therein ) , we favor hydrodynamical interactions like ram - pressure as the main process responsible for the suppression of the star formation .
however , we note that gravitational interactions can not be excluded in at least one case ( ngc4438 , e.g. , @xcite ) .
only recently , has it become possible to accurately quantify the time - scale for the quenching of the star formation after the stripping event @xcite .
luckily , almost all the objects for which a stripping time - scale has been computed are included in our sample . in fig .
[ timescale ] , we highlight the position of these six objects in the colour - mass diagram according to the age of the stripping event : hexagons and triangles indicate galaxies in which the star formation in the outer regions has been suppressed less or more than @xmath2300 myr , respectively .
interestingly , there is a difference in the average colour between very recent quenching ( @xmath44300 myr ) and older events , so that only galaxies with quenching time - scale @xmath2400 - 500 myr are at the edge or have already reached the transition region .
this is consistent with the fact that the virgo transition galaxy population is not virialized , implying a recent ( @xmath45 1.7 gyr , i.e. , the virgo crossing time ; @xcite ) infall into the cluster center .
thus , we can conclude that once the hi has been stripped from the disk , a galaxy moves from the blue cloud to the transition region in a time - scale roughly @xmath20.5 - 1 gyr .
it is more difficult to predict the future evolution of transition galaxies and in particular to estimate how long they will remain in the transition region and whether they will eventually join red - sequence galaxies .
as already pointed out by @xcite and @xcite , one cluster - crossing is not sufficient to completely halt the star formation in massive ( m@xmath46@xmath4710@xmath29 m@xmath30 ) galaxies .
in fact , while the outer disk is completely deprived of its gas content and star formation is quickly stopped , in the central regions the restoring force is too strong , keeping the atomic hydrogen reservoir necessary to sustain continuos star formation . moreover
, @xcite showed that two cluster crossings ( @xmath22 - 3 gyr ) are already necessary to move the brightest de from the blue to the red sequence , suggesting that the transition galaxies described here will take at least the same amount of time to have their star formation completely quenched . assuming that the @xcite formalism for ram - pressure is still valid after the first passage and that the galaxy s orbit does not change significantly
, we can expect that very little additional gas will be stripped during the second passage by ram - pressure .
significant stripping would occur only if the galaxy s restoring force is lowered by gravitational interactions with other members and the cluster potential well .
other intra - cluster medium related environmental effects , like viscous stripping @xcite and thermal evaporation @xcite , may thus play an important role in the complete suppression of the star formation .
same as fig .
[ cmall ] ( left panel ) .
large symbols indicates galaxies in the crowl & kenney ( 2008 ) sample for which a stripping time - scale estimate is available . stripping time - scale shorter than 300 myr and between 300 - 500 myr
are shown with hexagons and triangles respectively . ]
an upper limit to the time spent by hi - deficient galaxies in the transition region can be obtained if we assume that all the remaining gas will be consumed by star formation .
it is in fact plausible that the intra - cluster medium will prevent additional infall of cold gas . in this case
, the ` roberts ' time ( @xcite , defined as the ratio of the gas mass to the current star formation rate : i.e. , m(gas)/sfr ) can be used to obtain a rough estimate of the gas consumption time . assuming that 15% of the total gas is in the molecular state @xcite and @xmath230% is composed by helium and heavy elements @xcite , we find that the ` roberts ' time is already @xmath22.2 gyr1.5 if a salpeter imf is adopted ] .
this is in reality a lower limit to the real value since it does not take into account gas recycling .
as shown by @xcite , the real gas consumption time is 1.5 - 4 times longer then the time scale calculated above .
thus , although they have lost a significant amount of their original gas content , hi - deficient transition galaxies still have enough fuel to sustain star formation at the current rate for at least a couple of gyr .
detailed simulations focused on the effect of the cluster environment after the first passage will thus be extremely interesting to understand the future evolution of these systems . at this stage ,
the main conclusion we can draw from our analysis is that at least @xmath23 gyr seem to be necessary for the complete migration of a galaxy from the blue to the red sequence when gas stripping via the intra - cluster medium is involved .
l@c@c@c@c@c@c@c@c@c@c@c@c@c@c@ + name & type & d & agn & b - v & fuv & nuv & h & f@xmath48 & f@xmath49 & m(hi ) & m(hi)/m@xmath46 & c@xmath50(h ) & merging / accretion ? & ref .
+ & & mpc & & & m@xmath51 & m@xmath51 & m@xmath51 & jy & jy & 10@xmath52 m@xmath30 & & & & + ngc3619 & s0 & 20.7 & - & 0.86 & 16.86 & 16.14 & 7.40 & 0.43 & 1.61 & 7.08 & 0.04 & 6.2 & yes & 1,2 + ngc3898 & sa & 15.7 & lin & 0.79 & - & 15.03 & 6.25 & 0.42 & 2.02 & 26.9 & 0.09 & 3.1 & may be & 3 + ngc4324 & sa & 17.0 & lin / sey & 0.87 & 16.62 & 15.99 & 7.29 & 0.45 & 1.96 & 6.76 & 0.05 & 3.5 & - & - + ngc4370 & sa & 23.0 & nol & 0.70 & - & 18.26 & 8.28 & 0.94 & 3.27 & 4.0 & 0.04 & 3.8 & yes & 4,5 + ngc4378 & sa & 17.0 & sey & 0.81 & 16.00 & 15.50 & 7.23 & 0.36 & 1.45 & 10.0 & 0.07 & 4.9 & - & - + ngc4772 & sa & 17.0 & lin & 0.87 & 17.07 & 16.16 & 7.12 & 0.38 & 1.32 & 6.17 & 0.04 & 5.1 & yes & 6 + ngc5701 & sa & 20.1 & lin & 0.84 & 15.48 & 15.11 & 6.74 & 0.27 & 1.36 & 61.7 & 0.20 & 4.0 & - & - + + name & type & d & agn & b - v & fuv & nuv & h & f@xmath48 & f@xmath49 & m(hi ) & m(hi)/m@xmath46 & c@xmath50(h ) & merging / accretion ? & ref . + & & mpc & & & m@xmath51 & m@xmath51 & m@xmath51 & jy & jy & 10@xmath52 m@xmath30 & & & & + ngc4203 & s0 & 17.0 & lin . & 0.99 & 16.94 & 15.87 & 6.26 & 0.59 & 2.16 & 33.1 & 0.09 & 6.8 & yes & 7 + ngc4262@xmath53 & s0 & 17.0 & nol & 0.83 & - & 16.9 & 7.20 & - & 0.50 & 5.13 & 0.04 & 6.0 & yes & 8 + ngc4698@xmath53 & sa & 17.0 & sey & 0.83 & 16.64 & 15.71 & 6.16 & 0.63 & 1.89 & 17.0 & 0.04 & 4.2 & yes & 9 + ngc4866 & s0 & 17.0 & lin & 0.96 & 17.26 & 16.27 & 6.76 & - & - & 13.5 & 0.06 & 3.7 & - & - + ngc5103 & sab & 17.0 & - & - & 19.49 & 17.76 & 8.29 & - & - & 2.40 & 0.05 & 6.3 & - & - + @xmath53 uv fluxes increase significantly if the uv rings outside the optical radius are included .
+ references : ( 1 ) @xcite ; ( 2 ) @xcite ; ( 3 ) @xcite ; ( 4 ) @xcite ; ( 5 ) @xcite ; ( 6 ) @xcite ; ( 7 ) @xcite ; ( 8) @xcite ; ( 9 ) @xcite + contrary to hi - deficient objects , sure hi - normal transition galaxies are equally distributed between the field and cluster environments .
they are @xmath212 % in number30% in case all galaxies without hi measurement are hi - normal galaxies . ]
( 8 objects in total ) and @xmath234% in stellar mass of the whole transition galaxy population in our sample .
we note that ngc4565 is the only example of perfectly edge - on transition object , in this case we can not exclude that the corrections adopted still underestimate the real uv extinction @xcite making this galaxy possibly an erroneous transition object .
therefore , we will exclude ngc4565 from the following analysis .
the properties of the remaining 7 objects are listed in table [ hirich ] as follows : col .
col.2 : morphological type .
3 : distance in mpc . col.4
agn classification ( following the criteria described in @xcite ) : lin = liner , sey = seyfert , nol= no emission lines .
col.5 : @xmath16 colour corrected for dust extinction .
col.6 - 8 : fuv , nuv and h ab magnitudes .
col.9 - 10 : iras fluxes at 60 and 100 @xmath54 .
11 : hi mass .
12 : hi- to stellar - mass ratio .
13 : concentration index in h band taken from 2mass ( @xmath55 defined as the ratio between the radii containing 75% and 25% of the total h - band light ) .
14 - 15 : note regarding any evidence ( and relative reference ) supporting an external origin for the hi . in fig .
[ fctrans ] , we show sdss optical and galex uv colour images for each galaxy . in appendix
a , we describe the properties of each object in order to investigate the possible origins for the normal hi content and low ssfr . from this analysis
, it emerges that hi - normal transition galaxies are a heterogeneous class of objects going from merger remnants ( e.g. , ngc3619 ) with star - formation activity limited to the center , to satellites of big ellipticals ( ngc4370 ) , with no evident signs of recent star formation .
contrary to the hi - deficient family , gas stripping by environmental effect seems not to be playing any role in their recent evolution . as expected in the case of massive ( @xmath56 m@xmath30 ) galaxies , the majority of these objects host an ` optical ' agn .
more surprisingly , our analysis shows that , although they have likely followed different evolutionary paths , a significant fraction of galaxies ( at least 4 out of 7 ) has recent star formation mainly in form of one or more uv rings . as shown in fig .
[ fctrans ] , the uv rings have different morphologies going from inner rings ( ngc3898 , ngc4324 , ngc4772 ) to hoag - like objects ( ngc5701 ) . resolved hi maps , available for two of our objects ( ngc3898 , ngc4772 ) , reveal that the hi is distributed in extended low surface density disks , exceeding significantly the typical column density of 1 - 2 m@xmath30 pc@xmath15 only in correspondence of the star forming rings .
thus , in these cases , star formation is reduced not because the hi has been stripped but just because the gas is not able to collapse into stars efficiently .
this is likely due to the fact that the gas reservoir has a typical column density well below the critical density necessary to convert hi into molecular hydrogen and onset the star formation @xcite .
the presence of uv rings becomes more intriguing when hi - normal galaxies in the ` nuv red sequence ' are also taken into account ( 5 galaxies in total , see table [ hirich ] , fig . [ fcred ] and appendix a for a description of these objects ) .
the four galaxies with clear star formation activity ( ngc4203 , ngc4262 , ngc4698 , ngc4866 ) have star - forming regions mainly arranged in structures which are suggestive of one or multiple rings ( see fig . [ fcred ] ) . in the case of ngc4203 and ngc4698 , hi maps reveal a morphology similar to the one observed in hi - normal transition objects with extended low surface density hi disks and peaks of column density in correspondence of uv star - forming regions .
the only known exception is represented by ngc4262 , where the hi is mainly segregated in the uv star forming ring .
interestingly , ngc4262 and ngc4698 could be immediately reclassified as transition region galaxies if the outer uv rings are included in the estimate of the uv flux .
all this observational evidence strongly suggests that the evolutionary paths leading these objects to the transition region have been significantly different from the one followed by hi - deficient galaxies . the main piece of information necessary to understand the recent evolutionary history of these unusual ` gas - rich ' systems , is the origin of their gas reservoir . has the gas an external origin ( e.g , accretion , infall , merging , shells , etc . ) or was it always part of the galactic halo but has not been efficiently converted into stars ?
the origin of the gas can in fact not only provide clues on the past history of these objects ( i.e. , whether they are really migrating from the blue to the red sequence ) but also may help in predicting their future evolution in the colour - mass diagram . in the following ,
we combine the information available for hi - normal transition and red sequence galaxies to determine whether these systems have recently left the blue sequence after a quenching episode or are migrating back from the red sequence to the transition region thanks to a recent accretion event .
as discussed in appendix a , many galaxies in our sample show direct or indirect evidence of past gas accretion / infall events ( e.g , warps , counter - rotating or decoupled components , stellar shells ) . among the best candidates for an external origin of
all the hi observed , there are ngc4262 and ngc4203 in the red sequence and ngc3619 in the transition region @xcite .
however , it is interesting to note that the acquisition mechanisms ( and therefore the evolution ) of the three systems is likely to be different . in the case of ngc3619
, we are likely witnessing a minor merger with a gas - rich satellite .
the hi is segregated well within the optical radius , roughly coinciding with the star forming disk observed in uv , suggesting that a satellite has sunk into the center initiating an episode of star formation .
given that the stellar populations have ages and metallicities typical of unperturbed ellipticals @xcite , the most plausible scenario is that ngc3619 has left the red sequence after the merging event .
interestingly , at the current sfr ( @xmath20.1 m@xmath30 yr@xmath5 ) , the amount of atomic hydrogen present within the optical disk ( @xmath27@xmath5710@xmath52 m@xmath30 ) is sufficient to sustain the star formation for several billion years .
thus , ngc3619 will either remain in the transition region for a long time or , in case of a significant increase of the sfr , may be able to temporarily rejoin the blue cloud in a uv - near - infrared colour magnitude diagram . a similar evolutionary path could also have been followed by the dust - lane early - type ngc4370 .
however , the lack of detailed hi maps prevent us from drawing any conclusion . on the contrary , the infall of hi into ngc4203 and ngc4262 has likely followed less ` violent ' paths . in ngc4262 ,
the presence of a ring composed only of hi and newly formed stars is strongly suggestive of recent accretion , apparently ruling out that the ring has been formed from galactic material through bar instability .
an interesting possibility is that the bar could still be responsible for the peculiar configuration of the hi , preventing the newly accreted gas to collapse into the center .
what remains unclear is whether the gas in the ring has been accreted from the intergalactic medium ( as proposed in the case of polar ring galaxies ; e.g. , @xcite ) or during an interaction with another galaxy @xcite . as for ngc3619 , the gas reservoir in the ring is sufficient to keep the galaxy in the transition region for several gyr or to move it back to the blue cloud , building - up a new stellar disk / ring .
to this regard , it is tempting to consider ngc4262 the ancestor of hoag - type objects like ngc5701 , thus implying that these two systems may be on their way back to the uv blue cloud .
however , at this stage it is impossible to determine whether these two systems are at different stages of the same evolutionary path .
a migration back to the transition region appears instead very unlikely in the case of ngc4203 . despite its huge hi reservoir
, this galaxy shows only weak traces of recent star formation activity and at this rate the integrated colour will not be significantly affected , leaving this object in the red sequence .
a similar scenario could also be valid for ngc5107 and ngc4886 which already are in the red sequence . however , additional observations are required to unveil the evolutionary history of these systems . in summary , for at least a few cases ,
observations seem to suggest that hi - normal red galaxies have recently acquired atomic hydrogen and have started a new cycle of star formation activity leaving , at least temporarily , the red sequence . + for other transition galaxies , this scenario appears extremely unlikely .
this is particularly the case of ngc3898 , ngc4772 and ngc4698 . these three systems have very similar properties : i.e. , sa / sab type with a significant bulge component ( bulge - to - total ratio @xmath20.2 - 0.4 ; @xcite ) , hi mainly distributed in two rings , one inside and one outside the optical radius , corresponding to the sites of recent star formation activity .
hst images reveal that all three galaxies harbor a classical bulge @xcite , consistent with a ` violent ' and quick bulge formation in the past through mergers or clump coalescence in primordial disks ( e.g. , @xcite ) .
in addition , the presence of a decoupled core ( ngc4698 ) or a counter - rotating gas disk ( ngc4772 ) is suggestive of a more recent accretion event ( e.g. , minor merger ) supporting an external origin for at least part of the hi in these objects .
the preferred explanation for the properties of ngc4698 is in fact a later formation of the disk through the acquisition of material by a completely formed spheroid .
thus , it would be natural to argue that we are witnessing the build - up of the disk and that these galaxies are gradually moving from the red to the blue cloud , following a path consistent with what expected by hierarchical models @xcite .
however , a rough time - scale argument rules out this hypothesis .
all three galaxies harbor massive stellar disks ( m@xmath581 - 3@xmath5710@xmath29 m@xmath30 depending on the mass - to - light ratio difference between bulge and disk ) and a sfr of @xmath20.7 - 2 m@xmath30 yr@xmath5 during the last hubble time is thus necessary to form the observed disks .
these sfrs are more typical of blue - cloud galaxies and a factor @xmath210 larger than the sfr observed in these systems ( @xmath590.07 - 0.2 m@xmath30 yr@xmath5 ) .
while minor mergers have probably affected these systems , we can exclude that a great part of the stellar disk is composed of ` accreted stars ' , since the whole stellar component of the satellite is supposed to collapse into the center , contributing to the growth of the bulge @xcite .
although the uncertainties in the estimate of stellar masses and sfr are still quite significant , the large discrepancy between the observed and expected sfr suggests that the existing stellar disks are too massive to have been formed at the current sfr . of course
, we can not exclude multiple transitions from the blue to the red sequence and vice - versa @xcite , but given the low amount of observational constraints available , we prefer not to include such scenario in our analysis . we thus propose that the sfr in the disks was higher in the past or , in other words , that these galaxies have probably migrated from the blue cloud .
a reduction in the sfr is likely due to the low hi column density in these systems : on average below the threshold for the onset of star formation . what caused this reduction is still unclear and only more detailed theoretical models and simulations will help us to solve this mystery .
the same mechanism is probably behind the ring - like structures observed in both hi and uv . both internal ( e.g , bar instability ) or external ( e.g. accretion , merging ) processes can be responsible for such features .
however the absence of strong bars , the presence of decoupled / counter rotating components and the size - ratio of the inner and outer rings ( @xmath23.3 - 3.6 , i.e. different from the typical value expected for lindblad resonances @xmath22.2 , although not completely inconsistent ; @xcite ) , favour an external mechanism behind the unusual properties of these systems . since all these galaxies harbour an agn , it is natural to think about agn - feedback
however , these are generally low - energetic agns and no direct evidence ( e.g. , jets ) supporting this scenario has been found so far .
moreover , it is not clear how star - forming rings and low surface density hi disks can be formed via agn - feedback .
finally , it is important to note that , whatever the past evolutionary history of these systems , the hydrogen reservoir available can sustain the current star formation activity in these galaxies for at least 2 - 5 gyrs ( ignoring molecular hydrogen , helium and recycling ) .
thus , if star formation will remain as efficient as it is now , it will take a long time for these galaxies to reach the red sequence in a uv - optical / near - infrared colour - mass diagram .
unfortunately , no speculation can be made about the past evolutionary history of ngc4324 and ngc4378 given the lack of multiwavelength observations .
our analysis provides definitive evidence that galaxies in the transition region of the @xmath13 colour - mass diagram are a heterogeneous population .
this result strongly suggests that galaxies lying between the blue and red sequence have followed different evolutionary paths , not always going towards redder colours .
we remind the reader that the difference in number density between hi - deficient and hi - normal transition galaxies must be taken with a grain of salt .
although our sample is magnitude- and volume - limited , it might be biased towards high - density environments .
nearly half of the galaxies in our sample lies in fact within the virgo cluster , which might not be a fair representation of the local universe .
interestingly , external ( i.e. , environmental ) mechanisms are almost always behind the peculiar properties of transition galaxies , both hi - deficient and hi - normal objects .
although the majority of high - mass ( @xmath60@xmath4710@xmath29 m@xmath30 ) transition galaxies harbour an agn , feedback from accreting super - massive black holes appears not to be necessary to explain their properties . as discussed in hc09 , the presence of agns in transition galaxies
does not automatically imply a physical link between nuclear activity and quenching .
moreover , we do not find any direct observational evidence ( e.g. , jets , radio lobes , etc . ) supporting an interaction between the central black hole and the galaxy s gas reservoir .
hi - deficient transition galaxies constitute the majority of the transition population in our sample . by extending the analysis presented in hc09 , we have shown that these galaxies not only are mainly found in the virgo cluster but they also are the only population which is clearly migrating from the blue towards the red sequence .
while environmental effects are certainly able to strip the gas from the disk , reducing the star formation in just a few hundreds million years and forcing the galaxy to leave the blue cloud , less clear is the last leg of the journey , i.e. the migration from the transition region to the red sequence .
the complete suppression of the star formation requires at least a few billion years .
this ` two - step ' migration is more dramatic , and perhaps only visible , in a uv - near - infrared colour - mass diagram whereas in optical the first stripping event is sufficient to make the colours almost as red as an early - type galaxy . at this stage
, it is very tempting to use our data to quantify the current ` migration ' rate and , consequently , the mass accretion rate of the red sequence .
unfortunately , the large uncertainties in both observables and on the typical time - scale of the migration ( at least a factor 2 ) make this exercise not very useful .
for example , a quenching time of @xmath23 gyr would suggest that , at the current rate , the red sequence in our sample could have been built by the migration of objects from the blue cloud in a hubble time , consistent with previous works ( e.g. , @xcite ) .
however , at the same time , we are not able to reject scenarii in which either the observed rate is able to build - up the red sequence in half a hubble time or the observed migration is not able to explain the growth of the quiescent galaxy population in the last 13 gyr .
thus , no additional constraints on the evolution of the colour - stellar mass diagram are imposed by the estimate of the stellar mass accretion rate observed in our sample .
a more interesting exercise is to look for any morphological transformation during the migration towards the red sequence .
the crucial question here is whether the red sequence is fed with bulge dominated or disk galaxies .
the answer is clear from fig .
[ c31virgo ] , where we compare the distribution of the concentration index in h - band for galaxies in the virgo cluster . in the transition region
, we show only virgo hi - deficient galaxies .
overall ( fig .
[ c31virgo ] , left panel ) , hi - deficient transition galaxies have a concentration index much more similar to blue than red - sequence systems .
however , such difference is only evident at stellar masses higher than @xmath210@xmath29 m@xmath46 ( fig . [ c31virgo ] , central panel ) whereas for smaller galaxies ( right panel ) the distribution of @xmath55 does not significantly vary across the whole range of colours , reflecting the fact that dwarf ellipticals have exponential light profiles like dwarf irregulars ( e.g.,@xcite ) .
this result implies that , while hi - deficient transition galaxies are likely the progenitors of cluster low - mass red objects ( see also @xcite ) , this is not completely true at high stellar masses .
this is additionally supported by the fact that the vast majority of high - mass transition galaxies are early - type spirals and almost no ellipticals are present ( fig [ typedistr ] and [ trdistr ] ) .
the vast majority of galaxies in the process of reaching the red sequence are thus disk systems , significantly different from ellipticals or bulge - dominated galaxies characterizing high - mass , quiescent objects at low redshift .
since a significant fraction of hi - deficient transition galaxies appears to have recently infalled into the center of virgo and will likely spend a few gyr in the transition region ( see [ hidef ] ) , it may be possible that a morphological transformation still takes place before reaching the red sequence . although we can not completely exclude this scenario , we consider it unlikely .
given the long time required to halt the star formation , galaxies with increased bulge component should be present in our sample .
moreover , since gas stripping appears not to significantly increase the bulge component in early type galaxies ( e.g. , @xcite ) , an additional environmental effect ( different from the one responsible for the quenching of the star formation ) must be invoked .
thus , the fact that transition galaxies are not morphologically transformed before reaching the red sequence may have two different implications : either 1 ) galaxies are morphologically transformed once already in the red sequence , or 2 ) the mechanism controlling the accretion of stellar mass into the red sequence at @xmath30 is not the one responsible for the creation of the red sequence in the first place .
although it is possible that the bulge component is enhanced in some red galaxies via gravitational interactions , such scenario seems unlikely to explain the growth of the red sequence .
firstly , the mechanism responsible for the morphological transformation should be efficient only on giants and not on dwarf systems , which seems inconsistent with what known about environmental effects @xcite . secondly , the major mergers required to significantly increase the bulge component are extremely rare in today s clusters of galaxies .
thirdly , the presence of red - sequence galaxies in isolation ( e.g. , hc09 ) implies that star formation has been suppressed also outside clusters of galaxies .
although our sample could be biased against isolated objects , the fact that we do not find a field population which is clearly migrating from the blue sequence may suggest that in low density environments the red sequence has been mainly populated in the past .
finally , if the morphological transformation takes place in the red sequence , the number of bulge - dominated / elliptical galaxies should decrease at increasing redshift , which does not seem to be the case @xcite .
thus , the most favourite scenario emerging from our analysis is that the red sequence is currently accreting mass , at all masses , mainly via disk galaxies .
no significant structural modification takes place during the journey from the blue cloud to the red sequence .
the main process responsible for the suppression of the star formation in nearby galaxies is thus not the same responsible for the formation of the red sequence at high redshift . .
the distribution of the h - band concentration index for virgo hi - deficient ( dashed ) and hi - normal transition galaxies .
all galaxies and galaxies with @xmath56 m@xmath30 are shown in the left and right panel respectively.,width=321 ] while in the case of hi - deficient objects it seems plausible to associate transition galaxies with objects that are migrating from the blue to red sequence , this is not always the case for hi - normal transition galaxies .
the discovery of such systems is probably the most exciting result of this work .
they represents @xmath212% of our transition galaxy population , and are only found at high stellar masses ( @xmath6110@xmath29 m@xmath46 ) . as shown in fig .
[ c31gasrich ] , these galaxies are mainly disks with a significant bulge component * ( they are in fact all sa or s0 galaxies ) * and , despite the low number statistics , it seems clear that , , contrary to the hi - deficient population , in this family we find very few ` disk - only ' galaxies . despite their different properties , a great fraction of these galaxies
show active star - forming regions and hi segregated in one or multiple ring - like structures .
the picture emerging from our analysis is quite complex and exciting at the same time , showing that the transition region can be fed with galaxies from both sequences .
red sequence galaxies can acquire new gas supply and restart their star formation activity , as predicted by cosmological simulations .
however , it is unlikely that the red galaxies in our sample will re - build a significant stellar disk ( see also @xcite ) .
merging , accretion of gas - rich satellites and exchange of material during close encounters are among the likely responsible for this rejuvenation .
such processes are more frequent in low density environments , where red - sequence galaxies are rarer , perhaps reducing the chances to observe such phenomenon . when a suppression of the star formation is the most likely scenario to explain hi - normal transition galaxies , the process behind such migration is still unclear .
in this case , star formation must be reduced by making the hi stable against fragmentation ( e.g. , by decreasing the hi column density below the threshold for star formation ) and not via hi stripping as observed in hi - deficient objects .
starvation @xcite by removing any extended gaseous halo surrounding the galaxy , preventing further infall , could be a possibility .
this would imply a longer time - scale ( several gyr ) for the migration from the blue cloud @xcite than the one observed in the case of hi stripping ( a few hundreds million years ) .
however , the evidence for accretion / interaction in some of these objects may suggest that starvation , if efficient , is not the only mechanism at work .
@xcite have recently proposed a ` morphological quenching ' to explain the origin of gas - rich bulge - dominated objects .
the idea behind this mechanism is that the presence of a bulge could inhibit the collapse of a gas disk .
however , the ` morphological quenching ' appears only to be effective when the disk stellar component is negligible , which is not the case for the majority of the systems in our sample .
finally , the fact that the majority of these objects show some level of nuclear activity , might indicate a link between agn activity and their position in the colour stellar mass diagram .
however , although we can not exclude that recent accretion events may have triggered the agn , we do not find any direct observational evidence suggesting that agn - feedback is playing a significant role in the recent star formation history of these objects ( see also hc09 ) .
thus , the past evolutionary history of hi - normal transition galaxies has still to be unravelled . whatever is the path followed to get to the transition region
, hi - normal systems currently have enough fuel to sustain star formation at the current rate for almost another hubble time .
this is an unexpected result , implying that such systems could remain in the transition region for a very long time and that the colour range 4.5@xmath06 mag does not automatically correspond to a snapshot in the star formation history of nearby galaxies .
in this paper , we have combined uv , hi and near - infrared observations to investigate the properties of local transition galaxies .
our main results are as follows .
+ i ) we confirm that the reddening of the galaxy colour is accompanied by a decrease in the hi gas - fraction .
however we show that , while in the blue cloud colour and gas - fraction are tightly correlated , in the transition region and red sequence such correlation is more scattered .
transition region galaxies can thus be divided into two main families according to their hi - content .
+ ii ) hi - deficient transition objects reside in high - density environments and environmental effects are the likely cause for the loss of gas .
star formation is rapidly suppressed after the gas stripping and the transition region is reached in less than a billion years . however , once in the transition region , massive galaxies are still forming stars and will not immediately reach the red sequence . a subsequent quenching is thus required to reach the red sequence , implying that the total migration time - scale is , at least , a few billion years .
+ iii ) although hi - deficient galaxies represents the bulk of the migrating population in our sample , they * are apparently not responsible * for the formation of the red sequence in the first place .
the gas stripping is in fact ` polluting ' the red sequence mainly with disk galaxies and no morphological transformation is observed during the quenching of the star formation .
+ iv ) contrary to hi - deficient systems , hi - normal transition galaxies ( @xmath212% of the whole transition population in our sample ) represent an heterogeneous population of objects , at least two of which are probably following the inverse path , migrating back from the red sequence .
the high hydrogen content in these systems is , at least in some cases , due to external accretion / interaction events .
the detailed evolution of these objects is still unclear , but it is likely that they will remain in the transition region for several billion years .
thus , a connection between transition region and migration from the blue to the red sequence may not always be true .
+ the discovery of a hi - normal population in the transition region has only been possible thanks to the combination of uv imaging ( necessary to properly separate blue and red sequence at all masses ) and hi single - dish observations ( to quantify the atomic hydrogen content ) .
the results here presented have inevitably lead to several questions we have not been able to answer in the current work .
is accretion really playing a crucial role in the evolution of these objects ? which mechanism suppressed star formation without removing the gas ?
how frequent are such objects ? etc . to be able to answer these questions ,
both detailed theoretical and observational studies of individual objects and statistical investigations of larger samples are mandatory .
luckily , we are currently in an exciting time for uv and 21 cm astronomy . particularly promising for this topic
is the _ galex arecibo sdss survey _ ( gass , @xcite ) which is currently making a census of the hi content in a complete sample of @xmath21000 massive galaxies ( @xmath6210@xmath29 m@xmath30 ) , ideal to investigate the role of gas and accretion on the evolution of transition galaxies .
thus , we are probably not too far from being able to unravel some of the mysteries still surrounding hi - normal galaxies in the transition region .
we are greatly indebted to alessandro boselli for his support and encouragement in carrying out this analysis and for providing part of the data before publication .
we wish to thank alessandro boselli , barbara catinella , jonathan davies and rory smith for useful discussions and helpful comments on the manuscript .
we thank the anonymous referee for useful comments which helped to improve this paper .
lc thanks the hospitality of the max planck institute for astrophysics where part of this paper was written .
lc and tmh are supported by the uk science and technology facilities council .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation , from the galex mission , developed in cooperation with the centre national detudes spatiales of france and the korean ministry of science and technology , from the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration and from the goldmine data base .
this research has made extensive use of nasa s astrophysics data system and of the astro - ph preprint aechive at http://arxiv.org .
123 natexlab#1#1 , s. , walcher , c. j. , le fvre , o. , et al .
2007 , , 476 , 137 , e. , bosma , a. , creze , m. , & schwarz , m. p. 1982 , , 107 , 101 , i. k. , glazebrook , k. , brinkmann , j. , et al .
2004 , , 600 , 681 , c. m. , cole , s. , & frenk , c. s. 1996 , , 283 , 1361 , e. f. , mcintosh , d. h. , katz , n. , & weinberg , m. d. 2003 , , 149 , 289 , e. f. , zheng , x. z. , papovich , c. , et al .
2007 , , 663 , 834 , f. , buson , l. m. , & zeilinger , w. w. 1988 , , 335 , 705 , f. , corsini , e. m. , vega beltrn , j. c. , et al . 1999 , , 519 , l127 , d. , galletta , g. , & sage , l. j. 1993 , , 280 , 121 , b. & cameron , l. m. 1991 , , 252 , 27 , b. , sandage , a. , & tammann , g. a. 1985 , , 90 , 1681 , y. , dekel , a. , & neistein , e. 2007 , , 380 , 339 , m. r. 2006 , , 648 , 268 , a. , meisenheimer , k. , bell , e. f. , et al . 2006 , , 453 , 869 , a. , boissier , s. , cortese , l. , & gavazzi , g. 2008 , , 674 , 742 , a. , boissier , s. , cortese , l. , et al .
2005 , , 623 , l13 , a. , boissier , s. , cortese , l. , et al .
2006 , , 651 , 811 , a. , cortese , l. , deharveng , j. m. , et al .
2005 , , 629 , l29 , a. & gavazzi , g. 2006 , , 118 , 517 , a. , gavazzi , g. , donas , j. , & scodeggio , m. 2001 , , 121 , 753 , a. , lequeux , j. , & gavazzi , g. 2002 , , 384 , 33 , n. 1985 , , 153 , 199 , m. j. i. , dey , a. , jannuzi , b. t. , et al .
2007 , , 654 , 858 , g. & charlot , s. 2003 , , 344 , 1000 , v. , boissier , s. , burgarella , d. , et al . 2008 , , 483 , 107 , r. 1995 , , 96 , 39 , g. , freeman , t. , & howard , s. 1994 , , 108 , 2078 , g. , rautiainen , p. , salo , h. , buta , r. , & crocher , d. a. 1994 , , 108 , 476 , b. , haynes , m. p. , giovanelli , r. , gardner , j. p. , & connolly , a. j. 2008 , , 685 , l13 , b. , schiminovich , d. , & kauffmann , g. 2008 , in american institute of physics conference series , vol . 1035 , the evolution of galaxies through the neutral hydrogen window , ed .
r. minchin & e. momjian , 252255 , a. , daddi , e. , & renzini , a. 2006 , , 453 , l29 , c. , gunn , j. e. , & white , m. 2009 , , 699 , 486 , l. , boselli , a. , buat , v. , et al .
2006 , , 637 , 242 , l. , boselli , a. , franzetti , p. , et al .
2008 , , 386 , 1157 , l. , gavazzi , g. , boselli , a. , iglesias - paramo , j. , & carrasco , l. 2004 , , 425 , 429 , l. , minchin , r. f. , auld , r. r. , et al .
2008 , , 383 , 1519 , l. l. & barger , a. j. 2008 , , 686 , 72 , l. l. & songaila , a. 1977 , , 266 , 501 , h. h. & kenney , j. d. p. 2008
, , 136 , 1623 , r. , gavazzi , g. , arosio , i. , et al .
2007 , , 381 , 136 , a. & birnboim , y. 2006 , , 368 , 2 , v. , dalcanton , j. j. , aragn - salamanca , a. , et al .
2007 , , 660 , 1151 , s. p. , allen , p. d. , graham , a. w. , et al . 2006 , , 368 , 414 , s. p. , allen , p. d. , liske , j. , & graham , a. w. 2007 , , 657 , l85 , s. p. , popescu , c. c. , tuffs , r. j. , et al .
2007 , , 379 , 1022 , n. & fisher , d. b. 2007 , , 664 , 640 , k. & schneider , s. e. 1996 , , 112 , 937 , b. g. , bournaud , f. , & elmegreen , d. m. 2008 , , 688 , 67 , s. m. , willmer , c. n. a. , wolf , c. , et al .
2007 , , 665 , 265 , p. , scodeggio , m. , garilli , b. , et al .
2007 , , 465 , 711 , m. , krumholz , m. r. , prochaska , j. x. , gavazzi , g. , & boselli , a. 2009 , , 697 , 1811 , d. a. & de souza , r. e. 2003 , , 583 , l75 , a. m. 1993 , , 100 , 47 , g. , bonfanti , c. , sanvito , g. , boselli , a. , & scodeggio , m. 2002 , , 576 , 135 , g. , boselli , a. , donati , a. , franzetti , p. , & scodeggio , m. 2003 , , 400 , 451 , g. , boselli , a. , scodeggio , m. , pierini , d. , & belsole , e. 1999 , , 304 , 595 , m. , giuricin , g. , mardirossian , f. , mezzetti , m. , & boschin , w. 1998 , , 505 , 74 , j. e. & gott , j. r. i. 1972 , , 176 , 1 , c. p. , gargiulo , a. , & merluzzi , p. 2008
, , 385 , 1201 , g. k. t. , bower , r. g. , kilborn , v. , et al .
2008 , , 385 , 1965 , m. p. & giovanelli , r. 1984 , , 89 , 758 , m. p. , jore , k. p. , barrett , e. a. , broeils , a. h. , & murray , b. m. 2000 , , 120 , 703 , a. a. 1950 , , 55 , 170 , p. f. , cox , t. j. , younger , j. d. , & hernquist , l. 2009 , , 691 , 1168 , j. h. 2006 , , 131 , 2469 , t. m. & cortese , l. 2009 , , 396 , l41 , s. j. 2004 , , 611 , l89 , g. 1996 , , 281 , 487 , s. , schawinski , k. , devriendt , j. e. g. , et al .
2007 , , 173 , 619 , j. d. p. , tal , t. , crowl , h. h. , feldmeier , j. , & jacoby , g. h. 2008 , , 687 , l69 , r. c. , tamblyn , p. , & congdon , c. e. 1994 , , 435 , 22 , d. , katz , n. , weinberg , d. h. , & dav , r. 2005 , , 363 , 2 , c. s. , pahre , m. a. , falco , e. e. , et al .
2001 , , 560 , 566 , m. & noguchi , m. 1997 , , 481 , 132 , j. & kennicutt , jr . , r. c. 2004 , , 42 , 603 , p. , tout , c. a. , & gilmore , g. 1993 , , 262 , 545 , m. r. , mckee , c. f. , & tumlinson , j. 2009 , , 693 , 216 , n. , van driel , w. , & van woerden , h. 1985 , , 144 , 202 , r. b. , tinsley , b. m. , & caldwell , c. n. 1980 , , 237 , 692 , a. v. , moore , b. , & stadel , j. 2006 , , 636 , l25 , m. , bournaud , f. , teyssier , r. , & dekel , a. 2009 , arxiv e - prints:0905.4669 , d. c. , fanson , j. , schiminovich , d. , et al .
2005 , , 619 , l1 , d. c. , small , t. , schiminovich , d. , et al .
2007 , , 173 , 415 , d. c. , wyder , t. k. , schiminovich , d. , et al .
2007 , , 173 , 342 , e. d. , knezek , p. m. , & bregman , j. n. 1999 , , 510 , l95 , r. , davies , j. , disney , m. , et al .
2007 , , 670 , 1056 , m. 1999 , , 514 , 77 , e. , van der hulst , j. m. , sancisi , r. , swaters , r. a. , & van albada , t. s. 2005 , , 442 , 137 , e. , van der hulst , j. m. , sancisi , r. , swaters , r. s. , & van albada , t. s. 2007 , , 376 , 1513 , p. e. j. 1982 , , 198 , 1007 , r. w. 1999 , , 37 , 603 , t. a. , morganti , r. , sadler , e. m. , vergani , d. , & caldwell , n. 2002 , , 123 , 729 , m. a. , ashby , m. l. n. , fazio , g. g. , & willner , s. p. 2004
, , 154 , 229 , p. , bressan , a. , granato , g. l. , silva , l. , & danese , l. 2003 , , 409 , 99 , m. k. , pandey , s. k. , kembhavi , a. , & sahu , d. k. 2009 , arxiv e - prints:0901.1747 , p. g. , rieke , g. h. , villar , v. , et al .
2008 , , 675 , 234 , e. , corsini , e. m. , vega beltrn , j. c. , et al .
2001 , , 323 , 188 , m. , franx , m. , cross , n. j. g. , et al .
2005 , , 623 , 721 , p. j. , hernquist , l. , & fullagar , d. p. 1993
, , 403 , 74 , m. s. 1963 , , 1 , 149 , a. & bedke , j. 1994 , the carnegie atlas of galaxies ( washington , dc : carnegie institution of washington with the flintridge foundation , , k. , virani , s. , simmons , b. , et al .
2009 , , 692 , l19 , d. , wyder , t. k. , martin , d. c. , et al .
2007 , , 173 , 315 , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 , m. p. 1984 , , 209 , 93 , f. , ford , w. k. j. , jederzejewski , r. , & giovanelli , r. 1987 , , 320 , 454 , m. , gavazzi , g. , franzetti , p. , et al .
2002 , , 384 , 812 , m. s. 2005 , , 361 , l20 , m. f. , cutri , r. m. , stiening , r. , et al .
2006 , , 131 , 1163 , j. m. , giovanelli , r. , & haynes , m. p. 1996
, , 461 , 609 , c. m. , haynes , m. p. , & giovanelli , r. 2005 , , 621 , 215 , r. b. , mould , j. r. , & aaronson , m. 1982 , , 257 , 527 , s. 1976 , , 206 , 883 , w. , balkowski , c. , & van woerden , h. 1989 , , 218 , 49 , w. & van woerden , h. 1991 , , 243 , 71 , w. , van woerden , h. , schwarz , u. j. , & gallagher , iii , j. s. 1988 , , 191 , 201 , j. c. , pizzella , a. , corsini , e. m. , et al . 2001 , , 374 , 394 , b. , braine , j. , combes , f. , & sofue , y. 2005 , , 441 , 473 , b. , huchtmeier , w. , & van driel , w. 2005 , , 439 , 921 , b. c. , lucas , r. a. , mcelroy , d. b. , et al . 1990 , , 100 , 1489 , t. k. , martin , d. c. , schiminovich , d. , et al . 2007 , , 173 , 293 , c. & buat , v. 1995 , , 293 , l65 , w. , li , c. , kauffmann , g. , et al . 2009 , , 397 , 1243
* ngc3619 * is a s0 galaxy , member of a group composed of at least five systems @xcite .
the hi distribution is asymmetric and much concentrated in the center @xcite , where star formation is taking place and dust lanes are clearly visible .
the extent of the hi roughly matches the uv star - forming disk , i.e. , @xmath2half the optical radius . in optical , the galaxy shows prominent outer stellar shells , suggesting a recent minor - merging / accretion event .
numerical simulations predict that the stars from a satellite make a system of shells several 10@xmath52 yr after the end of the merging event which can last for more than 1 gyr @xcite .
the accretion scenario is also supported by the spectroscopic analysis performed by @xcite .
he finds that the age , metallicity , and @xmath43-enhancement ratios of ngc3619 are consistent with those of a representative sample of unperturbed nearby elliptical galaxies , contrary to what expected in the case of a recent major merger and supporting a minor merger with a gas - rich dwarf satellite .
thus , all the observational evidence is consistent with an external origin for the atomic hydrogen and residual star formation observed in ngc3619 .
* ngc3898 * is a sa galaxy in the ursa major group . in the hubble atlas it is described as the prototype of sa galaxy :
i.e. , with a prominent bulge component and multiple spiral arms .
hst observations reveal that the bulge is a ` classical bulge ' @xcite , consistent with a major merger / accretion event in the past .
uv observations show that active star formation is mainly segregated in a ring well within the optical radius , although star - forming regions are also found at the edge of the optical disk .
the atomic hydrogen is distributed in a low column density ( @xmath21 - 2 m@xmath30 pc@xmath15 ) disk extending up to @xmath24 times the radius containing 80% of the b - band light @xcite .
the highest hi column density is observed in correspondence with the uv star - forming regions , suggesting that only in these few regions the hi can condensate into molecular hydrogen and initiate the star formation cycle @xcite .
the hi disk appears to be warped in the inner parts and shows some ` wiggles ' with an amplitude of 30 - 50 km s@xmath5 @xcite .
although h@xmath43 long - slit spectroscopy @xcite and 21 cm hi line interferometry @xcite suggest that both ionized and atomic hydrogen have a regular velocity field and a smooth distribution , @xcite suggest that this galaxy may have experienced a recent interaction / accretion event
. however , the basis for this interpretation is not completely clear .
* ngc4324 * is a sa galaxy belonging to the virgo s cloud @xcite .
star formation is segregated in a ring , also visible in sdss images , distinct from the central bulge component .
although hi maps are not available , @xcite detect hi up to @xmath22 optical diameters suggesting that the atomic hydrogen is not only concentrated in the star - forming ring .
* ngc4370 * is an early type galaxy member of the virgo b cloud @xcite forming a non - interacting pair with ngc4365 .
its detailed morphological classification is rather uncertain , varying from sa @xcite to elliptical @xcite .
this is mainly due to the presence of a prominent dust lane along the equatorial plane of the galaxy .
although active star - forming regions are not clearly evident from uv and h@xmath43 images , we can not exclude that we are observing an analogous of the sombrero galaxy ( ngc4594 ) completely edge - on . excluding the prominent dust lane , no other evidence for interaction / accretion
is found and the gas and stars in the galaxy are co - rotating @xcite .
surprisingly , @xcite interpret the co - rotation in ngc4370 as a sign of possible accretion , suggesting that the gas present in this galaxy has an external origin .
although the basis for this interpretation may be debatable , the same conclusion has been recently reached by @xcite , who investigated the dust content in ngc4370 .
they propose that the amount of dust observed in the galaxy can not have been formed in situ and that at least part of the material must have been accreted during an interaction event .
an external origin for the atomic hydrogen content in this galaxy is thus not excluded , as observed in other hi - rich dust - lane elliptical galaxies @xcite .
* ngc4378 * is a sa galaxy belonging to the virgo s cloud .
it has a prominent nuclear bulge surrounded by a low surface brightness disk which contains a single , tightly - wound , spiral arm .
given the low surface brightness of the disk component , @xcite classified this object as an early - type anemic spiral , although we now know that its atomic hydrogen content is not significantly different from the one in high surface brightness disks .
uv star formation is clearly visible across all the disk .
numerical simulations @xcite suggest that the unusual one - armed spiral structure in ngc4378 is in reality an impulsive trailing arm created by the recent passage ( @xmath2200 myr ) of a small companion ( @xmath21/30 the galaxy s mass ) . since the satellite is supposed to survive the interaction , it seems unlikely that a significant fraction of the hi present in ngc4378 has been recently accreted .
the low ssfr of ngc4378 may therefore have an internal origin . by studying a sample of nearby galaxies
, @xcite recently found a correlation between the average shear rate in disks and their ssfr , suggesting the existence of a shear - threshold above which star formation is inhibited .
interestingly , ngc4378 has a shear rate ( [email protected] ) just below the threshold value ( [email protected] ) suggested by @xcite .
if this is the case , the internal dynamical properties of ngc4378 may be responsible for the low star formation efficiency observed in this object .
* ngc4772 * is a sa galaxy in the virgo s cloud .
recently , @xcite carried out a detailed multiwavelength analysis of this object , showing that it is probably the result of a minor merger event .
the main observational evidence supporting such scenario is : 1 ) the presence of dynamically decoupled central gas and stellar components , in particular in correspondence of the star forming ring visible in uv ( fig .
[ fctrans ] ) and h@xmath43 , 2 ) the unusual hi distribution , which is segregated in two concentric rings , the inner one where star formation is taking place and 3 ) the presence of a warp in the outer hi ring .
the most likely scenario is that ngc4772 is at the end stage of a prograde merger in which the transfer of angular momentum leads to an outward spread of the disk @xcite .
* ngc5701 * is the bluest object in the transition region .
we classified this galaxy as an sa ( ned gives s0/a ) .
however its peculiar morphology , characterized by a central bulge surrounded by a faint and apparently detached star - forming ring , makes ngc5701 a possible hoag - type object @xcite . in this case
the central condensation is a bar , not a ` bulge or elliptical ' @xcite and @xcite dubbed this object ` a barred galaxy without a disk ' .
the origin of hoag - type objects is still a puzzle , but two different scenarii seem to be the most commonly accepted : a ) accretion of small satellites @xcite , or b ) a strong bar instability @xcite , which destroys the disk and builds up a ring ( e.g. @xcite ) .
the lack of detailed dynamical and spectroscopic data prevent us from driving a conclusion on the origin of the ring in ngc5701 , but it is quite likely that the bar is playing an important role either destroying the disk or preventing the accreted gas from infalling into the center . * ngc4203 * is an , apparently isolated , barred s0 galaxy .
uv images show low level star - formation at least in the south - east part of the galaxy , corresponding to the outer edge of the warm - dust spiral structure detected at 8@xmath14 m in the galaxy center @xcite . despite its low star formation activity
, ngc4203 has a large reservoir of atomic hydrogen , distributed in a peculiar , filamentary structure and extended up to @xmath22.2 optical radii @xcite .
the atomic gas has a general sense of motion around the center , but it is clearly not on regular circular orbits @xcite .
again , the hi is mainly found in two ring structures : an inner ring ( in correspondence with the uv features ) and an outer ring which is likely rotating on a different plane @xcite .
the peculiar dynamics of ngc4203 is thus suggestive of an accretion event @xcite .
the hi disk of ngc4203 is also one of the nearest damped lyman alpha systems .
@xcite found that the metallicity of the gas is significantly lower than the one observed in galaxies of the same luminosity .
if confirmed , this would represent a strong observational evidence supporting the external origin for the hi .
* ngc4262 * is a barred s0 galaxy belonging to the virgo main cloud a. no star formation is observed within the optical radius .
however , uv images reveal the presence of a uv star - forming ring extended up to @xmath22.5 optical radii .
although faint uv emission is marginally visible all across the ring , great part of the star formation is taking place in the south - west .
this roughly coincides with the peak in hi distribution as revealed in the maps presented by @xcite and @xcite .
moreover , all the hi is segregated into a ring and no low surface density hi disk is detected .
@xcite and @xcite suggested that the ring has probably an external origin and more recently @xcite tentatively proposed that the hi could have been accreted during an interaction between ngc4262 and ngc4254 .
being a barred system plus a ring , ngc4262 shows some similarities with the hoag - type object ngc5701 . if this is the case , ngc4262 is probably at an earlier evolutionary stage , since its ring is only visible in uv whereas the ring in ngc5701 is evident also in optical , suggesting an older age for the stellar populations .
we note that , if we integrate the uv light up to the extent of the ring , ngc4262 would move from the red sequence to the transition region .
* ngc4698 * is a bulge - dominated sa galaxy in the virgo e cloud .
the carnegie atlas @xcite describes it as an elliptical - like bulge in which there is no evidence of recent star formation or spiral structure , plus a low - surface - brightness disk with spiral arms become prominent only in the outer part of the disk .
hst observations reveal that the bulge is a ` classical bulge ' @xcite , consistent with a major merger / accretion event in the past .
once again , uv images reveal that star formation is segregated into two ` almost perfect ' rings : one within and the other outside the optical radius .
interestingly , the two rings are not concentric : the inner ring is centered on the nucleus , whereas the outer one is offset @xmath2 20 arcsec to the north .
the peculiarity of ngc4698 becomes more evident when the dynamical properties of the bulge and disk component are compared .
@xcite showed that the rotation axis of the disk and of the bulge are almost perpendicular .
they interpret this unusual geometry as an evidence of a later formation of the disk through the acquisition of material by a completely formed spheroid .
also in this case , the galaxy moves back to the transition region if the outer ring is included in the estimate of the uv flux .
* ngc4866 * is an isolated , highly inclined , s0 galaxy ( although @xcite classify it as sa ) .
very few data are available in the literature , and the only unusual feature is the presence , once again , of a uv star - forming ring at the edge of the optical disk . *
ngc5103 * is an isolated sab galaxy . as for ngc3414 ,
no recent star formation activity is clearly visible in the uv image and the galaxy resembles an edge - on polar ring / disk system : @xcite classifies ngc5103 as possible candidate for polar - ring galaxy . no high - resolution hi maps are available for this object .
finally , a visual inspection of galaxies in the red sequence reveals the presence of two other objects with uv rings : ngc3945 and ngc4643 .
unfortunately , no hi observations are available and they are thus excluded from the discussion in the text .
|
we investigate the properties of galaxies between the blue and the red sequence ( i.e. , the transition region , 4.5@xmath06 mag ) by combining ultraviolet ( uv ) and near - infrared imaging to 21 cm hi line observations for a volume - limited sample of nearby galaxies .
we confirm the existence of a tight relation between colour and hi - fraction across all the range of colours , although outside the blue cloud this trend becomes gradually weaker .
transition galaxies are divided into two different families , according to their atomic hydrogen content . `
hi - deficient ' galaxies are the majority of transition galaxies in our sample .
they are found in high density environments and all their properties are consistent with a quenching of the star formation via gas stripping .
however , while the migration from the blue cloud is relatively quick ( i.e. , @xmath11 gyr ) , a longer amount of time ( a few gyr at least ) seems required to completely suppress the star formation and reach the red sequence . at all masses , migrating ` hi - deficient ' galaxies
are mainly disks , implying that the mechanism responsible for today s migration in clusters can not have played a significant role in the creation of the red sequence at high - redshift .
conversely , ` hi - normal ' transition galaxies are a more heterogeneous population .
a significant fraction of these objects show star formation in ring - like structures and evidence for accretion / minor - merging events suggesting that at least part of the hi reservoir has an external origin .
the detailed evolution of such systems is still unclear , but our analysis suggests that , in at least two cases , galaxies might have migrated back from the red sequence after accretion events .
interestingly , the hi available may be sufficient to sustain star formation at the current rate for several billion years .
our study clearly shows the variety of evolutionary paths leading to the transition region and suggests that the transition galaxies may not be always associated with systems quickly migrating from the blue to the red sequence .
[ firstpage ] galaxies : evolution galaxies : fundamental parameters galaxies : clusters : individual : virgo ultraviolet : galaxies radio lines : galaxies
|
grammatical acquisition proceeds on the basis of a partial genotypic specification of ( universal ) grammar ( ug ) complemented with a learning procedure enabling the child to complete this specification appropriately . the parameter setting framework of chomsky ( 1981 ) claims that learning involves fixing the values of a finite set of finite - valued parameters to select a single fully - specified grammar from within the space defined by the genotypic specification of ug .
formal accounts of parameter setting have been developed for small fragments but even in these search spaces contain local maxima and subset - superset relations which may cause a learner to converge to an incorrect grammar ( clark , 1992 ; gibson and wexler , 1994 ; niyogi and berwick , 1995 ) .
the solution to these problems involves defining default , unmarked initial values for ( some ) parameters and/or ordering the setting of parameters during learning .
bickerton ( 1984 ) argues for the bioprogram hypothesis as an explanation for universal similarities between historically unrelated creoles , and for the rapid increase in grammatical complexity accompanying the transition from pidgin to creole languages . from the perspective of the parameters framework ,
the bioprogram hypothesis claims that children are endowed genetically with a ug which , by default , specifies the stereotypical core creole grammar , with right - branching syntax and subject - verb - object order , as in saramaccan .
others working within the parameters framework have proposed unmarked , default parameters ( e.g. lightfoot , 1991 ) , but the bioprogram hypothesis can be interpreted as towards one end of a continuum of proposals ranging from all parameters initially unset to all set to default values .
a model of the language acquisition device ( lad ) incorporates a ug with associated parameters , a parser , and an algorithm for updating initial parameter settings on parse failure during learning .
basic categorial grammar ( cg ) uses one rule of application which combines a functor category ( containing a slash ) with an argument category to form a derived category ( with one less slashed argument category ) .
grammatical constraints of order and agreement are captured by only allowing directed application to adjacent matching categories . generalized categorial grammar ( gcg )
extends cg with further rule schemata .
the rules of fa , ba , generalized weak permutation ( p ) and backward and forward composition ( fc , bc ) are given in figure [ rules ] ( where x , y and z are category variables , @xmath0 is a variable over slash and backslash , and @xmath1 denotes zero or more further functor arguments ) .
once permutation is included , several semantically equivalent derivations for _ kim loves sandy _ become available , figure [ p - ba ] shows the non - conventional left - branching one .
composition also allows alternative non - conventional semantically equivalent ( left - branching ) derivations .
x / y y @xmath2 x & @xmath3 y [ x(y ) ] ( y ) @xmath2 x(y ) + + y x@xmath4y @xmath2 x & @xmath3 y [ x(y ) ] ( y ) @xmath2 x(y ) + + x / y y / z @xmath2 x / z & @xmath3 y [ x(y ) ] @xmath3 z [ y(z ) ] @xmath2 @xmath3 z [ x(y(z ) ) ] + + y@xmath4z x@xmath4y @xmath2 x@xmath4z & @xmath3 z [ y(z ) ] @xmath3 y [ x(y ) ] @xmath2 @xmath3 z [ x(y(z ) ) ] + + ( x@xmath0y@xmath5)@xmath6y@xmath7 @xmath2(x@xmath0y@xmath7)@xmath0y@xmath8 & @xmath3 y@xmath9,y@xmath5 [ x(y@xmath8,y@xmath7 ) ] @xmath2
@xmath3 y@xmath5,y@xmath9 [ x(y@xmath8,y@xmath7 ) ] + lll kim & loves & sandy + np & ( s@xmath4np)/np & np + kim & @xmath3 y , x [ love@xmath10(x y ) ] & sandy@xmath10 + & p & + & ( s / np)@xmath4np & + & @xmath3 x , y [ love@xmath10(x y ) ] & + & + s / np & & + @xmath3 y [ love@xmath10(kim@xmath10 y ) ] & & + + s & & + love@xmath10(kim@xmath10 sandy@xmath10 ) & & gcg as presented is inadequate as an account of ug or of any individual grammar .
in particular , the definition of atomic categories needs extending to deal with featural variation ( e.g. bouma and van noord , 1994 ) , and the rule schemata , especially composition and weak permutation , must be restricted in various parametric ways so that overgeneration is prevented for specific languages .
nevertheless , gcg does represent a plausible kernel of ug ; hoffman ( 1995 , 1996 ) explores the descriptive power of a very similar system , in which generalized weak permutation is not required because functor arguments are interpreted as multisets .
she demonstrates that this system can handle ( long - distance ) scrambling elegantly and generates mildly context - sensitive languages ( joshi _ et al _ , 1991 ) .
the relationship between gcg as a theory of ug ( gcug ) and as a the specification of a particular grammar is captured by embedding the theory in a default inheritance hierarchy .
this is represented as a lattice of typed default feature structures ( tdfss ) representing subsumption and default inheritance relationships ( lascarides _ et al _ , 1996 ; lascarides and copestake , 1996 ) .
the lattice defines intensionally the set of possible categories and rule schemata via type declarations on nodes .
for example , an intransitive verb might be treated as a subtype of verb , inheriting subject directionality by default from a type * gendir * ( for general direction ) . for english , * gendir * is default * right * but the node of the ( intransitive ) functor category , where the directionality of subject arguments is specified , overrides this to * left * , reflecting the fact that english is predominantly right - branching , though subjects appear to the left of the verb .
a transitive verb would inherit structure from the type for intransitive verbs and an extra np argument with default directionality specified by * gendir * , and so forth .
for the purposes of the evolutionary simulation described in 3 , gc(u)gs are represented as a sequence of _ p - settings _ ( where _ p _ denotes principles or parameters ) based on a flat ( ternary ) sequential encoding of such default inheritance lattices .
the inheritance hierarchy provides a partial ordering on parameters , which is exploited in the learning procedure .
for example , the atomic categories , * n * , * np * and * s * are each represented by a parameter encoding the presence / absence or lack of specification ( t / f/ ? ) of the category in the ( u)g . since they will be unordered in the lattice their ordering in the sequential coding
is arbitrary .
however , the ordering of the directional types * gendir * and * subjdir * ( with values l / r ) is significant as the latter is a more specific type
. the distinctions between absolute , default or unset specifications also form part of the encoding ( a / d/ ? ) .
figure [ binencoding ] shows several equivalent and equally correct sequential encodings of the fragment of the english type system outlined above .
[ cols= " < , < , < , < , < , < " , ] to test whether it was memory limitations during learning or during parsing which were affecting the results , another series of runs for `` english '' was performed with either memory limitations during learning but not parsing enabled , or vice versa .
memory limitations during learning are creating the bulk of the preference for a default learner , though there appears to be an additive effect . in seven of the ten
runs with memory limitations only in learning , a clear preference for default learners emerged . in five of the runs with memory limitations only in parsing there appeared to be a slight preference for defaults emerging .
default learners may have a fitness advantage when the number of interactions required to learn successfully is greater because they will tend to converge faster , at least to a subset language .
this will tend to increase their fitness over unset learners who do not speak any language until further into the learning period .
the precise linguistic environment of adaptation determines the initial values of default parameters which evolve .
for example , in the runs initialized with 16 unset learning `` malagasy '' vos adults and 16 default ( svo ) learning vos adults , the learning procedure which dominated the population was a variant vos default learner in which the value for * subjdir * was reversed to reflect the position of the subject in this language . in some of these runs , the entire population evolved a default * subjdir * ` right ' setting , though some lagts always retained unset settings for the other two ordering parameters , * gendir * and * argo * , as is illustrated in figure [ vos - subjdir ] .
this suggests that if the human language faculty has evolved to be a right - branching svo default learner , then the environment of linguistic adaptation must have contained a dominant language fully compatible with this ( minimal ) grammar . to explore the emergence and persistence of structured language , and consequently the emergence of effective learners , ( pseudo ) random
initialization was used .
a series of simulation runs of 500 cycles were performed with random initialization of 32 lagts p - settings for any combination of p - setting values , with a probability of 0.25 that a setting would be an absolute principle , and 0.75 a parameter with unbiased allocation for default or unset parameters and for values of all settings .
all lagts were initialized to be age 1 with a critical period of 3 interaction cycles of 2000 random interactions for learning , a maximum age of 10 , and the ability to reproduce by crossover ( 0.9 probability ) and mutation ( 0.01 probability ) from 4 - 10 . in around 5% of the runs ,
language(s ) emerged and persisted to the end of the run .
languages with close to optimal wml scores typically came to dominate the population quite rapidly .
however , sometimes sub - optimal languages were initially selected and occasionally these persisted despite the later appearance of a more optimal language , but with few speakers . typically , a minimal subset language dominated although full and intermediate languages did appear briefly , they did not survive against less expressive subset languages with a lower mean wml .
figure [ emerg - lgs ] is a typical plot of the emergence ( and extinction ) of languages in one of these runs . in this run ,
around 10 of the initial population converged on a minimal ovs language and 3 others on a vos language .
the latter is more optimal with respect to wml and both are of equal expressivity so , as expected , the vos language acquired more speakers over the next few cycles .
a few speakers also converged on vos - n , a more expressive but higher wml extension of vso - n - gwp - comp .
neither this nor the ovs language survived beyond cycle 14 . instead a vso language emerged at cycle 10 , which has the same minimal expressivity of the vos language but a lower wml ( by virtue of placing the subject before the object ) and this language dominated rapidly and eclipsed all others by cycle 40 . in all these runs ,
the population settled on subset languages of low expressivity , whilst the percentage of absolute principles and default parameters increased relative to that of unset parameters ( mean % change from beginning to end of runs : + 4.7 , + 1.5 and -6.2 , respectively ) .
so a second identical set of ten was undertaken , except that the initial population now contained two sov - v2 `` german '' speaking unset learner lagts . in seven of these runs ,
the population fixed on a full sov - v2 language , in two on the intermediate subset language sov - v2-n , and in one on the minimal subset language sov - v2-n - gwp - comp .
these runs suggest that if a full language defines the environment of adaptation then a population of randomly initialized lagts is more likely to converge on a ( related ) full language .
thus , although the simulation does not model the development of expressivity well , it does appear that it can model the emergence of effective learning procedures for ( some ) full languages .
the pattern of language emergence and extinction followed that of the previous series of runs : lower mean wml languages were selected from those that emerged during the run .
however , often the initial optimal svo - v2 itself was lost before enough lagts evolved capable of learning this language . in these runs ,
changes in the percentages of absolute , default or unset p - settings in the population show a marked difference : the mean number of absolute principles declined by 6.1% and unset parameters by 17.8% , so the number of default parameters rose by 23.9% on average between the beginning and end of the 10 runs .
this may reflect the more complex linguistic environment in which ( incorrect ) absolute settings are more likely to handicap , rather than simply be irrelevant to , the performance of the lagt .
partially ordering the updating of parameters can result in ( experimentally ) effective learners with a more complex parameter system than that studied previously
. experimental comparison of the default ( svo ) learner and the unset learner suggests that the default learner is more efficient on typologically more common constituent orders .
evolutionary simulation predicts that a learner with default parameters is likely to emerge , though this is dependent both on the type of language spoken and the presence of memory limitations during learning and parsing .
moreover , a svo bioprogram learner is only likely to evolve if the environment contains a dominant svo language .
the evolution of a bioprogram learner is a manifestation of the baldwin effect ( baldwin , 1896 ) genetic assimilation of aspects of the linguistic environment during the period of evolutionary adaptation of the language learning procedure . in the case of grammar
learning this is a co - evolutionary process in which languages ( and their associated grammars ) are also undergoing selection .
the wml account of parsing complexity predicts that a right - branching svo language would be a near optimal selection at a stage in grammatical development when complex rules of reordering such as extraposition , scrambling or mixed order strategies such as v1 and v2 had not evolved .
briscoe ( 1997a ) reports further experiments which demonstrate language selection in the model .
though , simulation can expose likely evolutionary pathways under varying conditions , these might have been blocked by accidental factors , such as genetic drift or bottlenecks , causing premature fixation of alleles in the genotype ( roughly corresponding to certain p - setting values ) .
the value of the simulation is to , firstly , show that a bioprogram learner could have emerged via adaptation , and secondly , to clarify experimentally the precise conditions required for its emergence . since in many cases these conditions will include the presence of constraints ( working memory limitations , expressivity , the learning algorithm etc . ) which will remain causally manifest , further testing of any conclusions drawn must concentrate on demonstrating the accuracy of the assumptions made about such constraints .
briscoe ( 1997b ) evaluates the psychological plausibility of the account of parsing and working memory .
1.6em - plus 1pt plus 1pt joshi , a. , vijay - shanker , k. and weir , d. ( 1991 ) ` the convergence of mildly context - sensitive grammar formalisms ' in sells , p. , shieber , s. and wasow , t. ( ed . ) , _ foundational issues in natural language processing , _ mit press , pp . 3182 .
niyogi , p. and berwick , r.c .
( 1995 ) ` a markov language learning model for finite parameter spaces ' , _ proceedings of the 33rd annual meeting of the association for computational linguistics , _ mit , cambridge , ma .. rambow , o. and joshi , a. ( 1994 ) ` a processing model of free word order languages ' in c. clifton , l. frazier and k. rayner ( ed . ) , _ perspectives on sentence processing , _ lawrence erlbaum , hillsdale , nj . , pp . 267301 .
|
a new account of parameter setting during grammatical acquisition is presented in terms of generalized categorial grammar embedded in a default inheritance hierarchy , providing a natural partial ordering on the setting of parameters .
experiments show that several experimentally effective learners can be defined in this framework .
evolutionary simulations suggest that a learner with default initial settings for parameters will emerge , provided that learning is memory limited and the environment of linguistic adaptation contains an appropriate language .
|
the einstein tensor is the only symmetric and conserved tensor depending on the metric and its derivatives up to the second order , which is linear in the second derivatives of the metric . dropping the last condition , lovelock @xcite found the most general tensor satisfying the other ones .
the obtained tensor is non linear in the riemann tensor and differs from the einstein tensor only if the space - time has more than 4 dimensions .
therefore the lovelock theory is the most natural extension of general relativity in higher dimensional space - times .
the lovelock theory for a particular choice of the coefficients of the action could be thought as the gravitational analogue of born - infeld electrodynamics @xcite . in the last decades
a renewed interest in both lovelock gravity and born - infeld electrodynamics has appeared because they emerge in the low energy limit of string theory @xcite .
since the lovelock tensor contains derivatives of the metric of order not higher than the second , the quantization of the linearized lovelock theory is free of ghosts .
for this reason the lovelock lagrangian appears in the low energy limit of string theory . in particular , the gauss - bonnet terms ( quadratic in the riemann tensor ) were studied in ref . @xcite and the quartic terms in refs .
the lovelock theory of gravity was also discussed in refs .
@xcite .
hoffmann was the first one in relating general relativity and the born - infeld electromagnetic field @xcite .
he obtained a solution of the einstein equations for a point - like born - infeld charge , which is devoid of the divergence of the metric at the origin that characterizes the reissner - nordstrm solution .
however , a conical singularity remained there , as it was later objected by einstein and rosen .
the einstein - born - infeld black hole has been revisited in refs .
@xcite the aim of this paper is to study the charged black hole solutions in five - dimensional lovelock gravity coupled to born - infeld electrodynamics . in section
ii , we discuss the five - dimensional lovelock gravity and study the black hole solutions as a preliminary step toward the calculation of the charged solution which we present in a separated section . in section iii , we discuss the born - infeld electrodynamics which will provide us the necessary tools in order to eventually find , in section iv , the five- dimensional charged black hole solution in lovelock - born - infeld field theory .
we study the geometrical properties of the solution and discuss the similarities and distinctions existing with respect to the reissner - nordstrm black hole .
section v is dedicated to the analysis of the thermodynamics of the lovelock black holes and the conclusions are contained in section vi .
the lovelock lagrangian density in @xmath0 dimensions is @xcite @xmath1 where @xmath2 ( for even @xmath0 ) and @xmath3 ( for odd @xmath0 ) . in ( [ lov1 ] ) , @xmath4 and @xmath5
are constants which represent the coupling of the terms in the whole lagrangian and give the proper dimensions . in eq .
( [ lov1 ] ) @xmath6 is @xmath7 where @xmath8 is the riemann tensor in @xmath0 dimensions , @xmath9 , @xmath10 is the determinant of the metric @xmath11 and @xmath12 is the generalized kronecker delta of order @xmath13 .
the lagrangians up to order 2 are given by @xcite @xmath14 @xmath15 @xmath16 @xmath17 where we recognize the usual lagrangian for the cosmological term , the einstein - hilbert lagrangian and the lanczos lagrangian @xcite respectively .
@xmath18 and @xmath19 the lovelock lagrangian is a linear combination of the einstein - hilbert and lanczos lagrangians .
the spaces of dimensions @xmath20 and @xmath21 includes the lagrangian @xmath22 , which was first obtained by mller - hoissen @xcite . in general , if the dimension is even and the manifold is compact with positive definite metric , then @xmath23 is the euler characteristic classes generator @xcite @xmath24 where @xmath25 are the euler - poincar topological invariants , which are zero in odd dimensions . hence , the geometric action is written as @xmath26 from the variational principle we obtain the lovelock tensor ( @xmath27 ) , @xmath28 where the general expression of @xmath27 is @xcite @xmath29 @xmath30 where @xmath31 for even @xmath0 or @xmath3 for odd @xmath0 .
when the dimension is @xmath32 , @xmath27 differs from the einstein tensor in a divergence term .
the lovelock tensors for @xmath33 and @xmath34 are @xmath35 @xmath36 in five dimensions the lagrangian is a linear combination of the einstein - hilbert and the lanczos ones , and the lovelock tensor results @xmath37 @xmath38 @xmath39 to be precise , let us notice that in the third term of eq .
( [ love ] ) we are abusing of the notation when including the cosmological constant @xmath40 explicitely .
we do that with the intention to indicate the nature of each term in a clear way , even though the constant @xmath40 is actually related to the couplings @xmath41 and @xmath42 in eq .
( [ tensorl ] ) by the relation @xmath43 .
the gauss - bonnet constant @xmath44 will allow us to track the changes in the equations , when we compare with the respective ones of general relativity .
the coupling constant @xmath45 introduces a length scale @xmath46 in the theory which physically represents a short - distance range where the einstein gravity turns out to be corrected .
the vacuum field equations are given by @xmath47 and accept spherically symmetric solutions in five dimensions , which in terms of a suitable schwarzschild - like _ ansatz _ , can be written as @xmath48 @xmath49 in this case , the solution of eqs .
( [ ecuvacio ] ) is @xmath50 where @xmath51 is an integration constant . by requesting the proper newtonian potential in the weak field region @xmath52 ( @xmath53 ) , it results that the adm mass is @xmath54 @xcite with @xmath55 for @xmath56 and @xmath57 for @xmath58 .
then @xmath59 asymptotically , this solution goes to the general relativity solution in five dimensions when @xmath60 , as it is expected .
namely @xmath61 let us notice that in the case of non - vanishing cosmological constant , besides the leading term in the expansion ( [ sol1 ] ) , we find finite-@xmath45 corrections to the black hole parameters .
namely @xmath62 where the _ dressed _ parameters @xmath63 and @xmath64 are given by @xmath65 being @xmath66 furthermore , in the case of the charged solution we will discuss in section iv , the ( charge ) parameter @xmath67 receives similar corrections due to these finite-@xmath45 effects , resulting @xmath68 notice that the parameter @xmath69 controls the _
dressing _ of the whole set of black hole parameters .
the above power expansion converges for values such that @xmath70 .
besides , for the case @xmath71 we find a different expansion , leading to the following _ dressed _ parameters in the large @xmath72 regime @xmath73 thus , we note that the newtonian term @xmath74 vanishes in the limit @xmath75 . the particular case @xmath76 is discussed below .
moreover , it is possible to see that , if one considers the case @xmath77 , the effective cosmological constant in the large @xmath69 limit turns out to be @xmath78 on the other hand , for the case of vanishing cosmological constant ( @xmath53 ) , the solution ( [ solnueva2 ] ) displays an event horizon located at @xmath79 when @xmath80 .
then , the horizon can reach the point @xmath81 for a massive object with @xmath82 ; in this case @xmath81 is a naked singularity . if @xmath83 there is no horizon .
one of the relevant differences existing between the black hole solutions in einstein and lovelock theories is the fact that @xmath84 is not singular at the origin .
instead , the metric is regular everywhere . from ( [ sol ] )
we obtain @xmath85 and , in fact , we will find a similar aspect for the case of the lovelock - born - infeld charged black hole . if the object has no mass ( @xmath86 ) , one gets de sitter ( @xmath87 ) and anti - de sitter ( ads ) solutions ( @xmath88 ) as particular cases ; namely @xmath89 where @xmath90 another interesting geometry is found in the particular case @xmath91 . at this point of the space of parameters ,
the solution ( [ sol ] ) becomes @xmath92 where we have considered @xmath93 , and introduced the notation @xmath94 .
certainly , we could refer to this particular black hole solution as the _ btz branch _
, due to its reminiscence of btz black hole @xcite ; this is an aspect that was already pointed out in ref .
actually , the solution ( [ gaston ] ) shares several properties with the three - dimensional black hole geometry , as the thermodynamical properties which will discussed in section v. indeed , the parameter @xmath95 in eq .
( [ gaston ] ) plays the role of the mass @xmath96 in the btz solution .
for instance , as well as @xmath97 spacetime is obtained as a particular case of the btz geometry by setting the negative mass @xmath98 , also the five - dimensional anti - de sitter space corresponds to setting @xmath99 .
on the other hand , let us notice that , in a consistent way , if the large @xmath45 limit is taken while fixing the condition @xmath100 one finds that the solution becomes the metric which represents the near boundary limit of @xmath101 , like it happens with the _ massless btz _ ( @xmath102 ) which is obtained by making the three - dimensional black hole to disappear .
then , the parallelism with the solutions in @xmath103 turns out to be exact since the five - dimensional metric obtained by keeping only the leading terms in the near boundary limit of @xmath101 corresponds to @xmath104 in ( [ gaston ] ) as well , which is precisely the lovelock black hole solution ( [ sol ] ) with _ minimal _ mass @xmath105 .
besides , a conical singularity is found in the range @xmath106 ( corresponding to @xmath107 ) in a complete analogy . the @xmath108-symmetry invariance of this particular lovelock solution was also discussed in ref .
@xcite . here
, we have shown how the solution ( [ gaston ] ) appears as a particular case of the geometry ( [ sol ] ) .
the digression above relies in the fact that , besides the parameter @xmath109 , which classifies the black hole geometry , we have additional parameters characterizing the field theory by means of the couplings @xmath4 of the different contributions @xmath110 in the lovelock lagrangian @xmath111 .
this is the case of the cosmological constant @xmath112 and the gauss - bonnet constant @xmath45 . as an example
, we observed that if one constrains the theory on a particular curve in the space of parameters defined by @xmath113 one finds the particular solution ( [ gaston ] ) .
thus , different regions of the space of parameters ( @xmath114 ) lead to quite different qualitative behaviours of the theory and , consequently , of its solutions .
in 1934 born and infeld @xcite proposed a non - linear electrodynamics with the aim of obtaining a finite value for the self - energy of a point - like charge .
the born - infeld lagrangian leads to field equations whose spherically symmetric static solution gives a finite value @xmath115 for the electrostatic field at the origin
. the constant @xmath115 appears in the born - infeld lagrangian as a new universal constant . following einstein , born and infeld considered the metric tensor @xmath11 and the electromagnetic field tensor @xmath116 as the symmetric and anti - symmetric parts of a unique field @xmath117 .
then they postulated the lagrangian density @xmath118 where the second term is chosen so that the born - infeld lagrangian tends to the maxwell lagrangian when @xmath119 . in four dimensions ,
this lagrangian results to be @xmath120 where @xmath121 and @xmath122 are the scalar and pseudoscalar field invariants @xmath123 @xmath124 the born - infeld lagrangian is usually mentioned as an exceptional lagrangian because its properties of being the unique structural function which : 1- assures that the theory has a single characteristic surface equation ; 2- fulfills the positive energy density and the non - space like energy current character conditions ; 3- fulfills the strong correspondence principle . as a consequence of these conditions ,
the lagrangian has time - like or null characteristic surfaces @xcite . in order to obtain the static spherically symmetric solution in five dimensions
, we will replace @xmath125 and the metric ( [ intervalo ] ) in the born - infeld lagragian ( [ bil ] ) ; then we will vary the action ( this procedure is valid due to the high symmetry of the solution we are looking for ) .
therefore @xmath126 the field equation derived from this lagrangian ( [ lagrang ] ) is @xmath127=0\ ] ] where @xmath128 .
so the born - infeld point charge field in five dimensions is @xmath129 the energy - momentum tensor is @xmath130 in the static isotropic case it results to be diagonal : @xmath131\ ] ] @xmath132 \label{energia}\ ] ] the energy of this field is finite in contrast to the energy of the maxwell field : @xmath133
we will study the exact solutions of lovelock gravity for a born - infeld isotropic electrostatic source .
the field equations to be solved are @xmath134 , where @xmath135 is the lovelock tensor ( [ love ] ) and @xmath136 is the born - infeld energy - momentum tensor corresponding to a point charge located in the origin ( [ energia ] ) . because of the symmetry of the source we repit the _ ansatz _ ( [ intervalo ] ) for the metric . in this case
only the diagonal components of the lovelock tensor survive : the components @xmath137 are equal , and they are integrals of the components @xmath138 .
therefore , it is enough to solve @xmath139 , which amounts to the equation @xmath140 @xmath141 the left hand side can be written as a total radial derivative , to be easily integrated .
the solution is @xmath142 @xmath143 where @xmath51 is an integration constant .
the integral inside the square involves an incomplete elliptic integral of the first kind @xmath144 @xcite , namely @xmath145 , where @xmath146 @xmath147,\frac{2+\sqrt{3}}{4}\bigg)$ ] thus , we obtain two solutions for the metric , but the sign of @xmath148 is determined requiring that in the limit @xmath149 we must recover the newtonian potential in five dimensions @xmath150 , so we obtain @xmath55 for @xmath56 and @xmath57 for @xmath58 . in that limit the solution is @xmath151 with @xmath152 where @xmath153 . in terms of the adm mass @xmath109 , @xmath84 becomes @xmath154 @xmath155 this class of solutions was also studied in reference @xcite . in the limit @xmath156 the solution tends to @xmath157 @xmath158
this limit agrees with the quoted four - dimensional hoffmann solution @xcite with a conical singularity in the origin of the black hole .
consequently , by performing the limit @xmath159 in eq .
( [ limitealfa ] ) we recover the reissner - nordstrm solution in five dimensions with cosmological constant , @xmath160 on the other hand , by taking the limit @xmath161 in ( [ solteo ] ) , we also recover the lovelock - maxwell solution found by wiltshire in @xcite ; namely @xmath162^{\frac{1}{2}}\ ] ] notice that , because to the existence of a birkhoff - like theorem ( see appendix a in ref . @xcite and references therein )
, this limit turns out to be more than a simple heuristical argument to check the solution ( [ solteo ] ) , representing a necessary condition which required to be proved .
we also observe that the uncharged solution ( 11 ) is recovered in the limit @xmath163 .
coming back to the general solution ( [ solteo ] ) , we can see that , differing from schwarzschild and reissner - nordstrm solutions , the metric has no singularities ( fig .
[ comparacion ] ) @xmath164 ( lbi ) [ @xmath165 , @xmath166 , @xmath53 , @xmath167 , @xmath168.,width=264 ] depending on the values of the parameters of the black hole ( charge @xmath67 , mass @xmath109 , born - infeld constant @xmath115 and gauss - bonnet constant @xmath148 ) the square root could be imaginary . if @xmath169 and @xmath170 the solution is valid for all @xmath171 .
if @xmath172 then @xmath173 , so the metric has a conical singularity at the origin : whereas the equator measures @xmath174 , the radius at the equator is @xmath175 . in the critical case @xmath176 one finds @xmath177 and the conical singularity disappears from the metric since in this case @xmath178 when @xmath179 .
conversely , we can think in the critical value @xmath180 as follows : we can define a critical value @xmath181 which represents an upper bound on @xmath67 for the metric ( 24 ) to be well defined in the whole spacetime .
thus , a critical value for the black hole charge appears in this context as a direct consequence of the finite-@xmath115 effects .
the role payed by @xmath115 is setting the critical value @xmath182 in the maxwellian regime @xmath183 , which is consistent with the fact that the lovelock - maxwell black hole geometry is not regular if @xmath184 . besides , the finite-@xmath115 effects act in ( [ solteo ] ) as a kind of effective cosmological constant @xmath185 , with @xmath186 and @xmath187 .
thus , this fact could lead one to infer that the _ dressing _ of the black hole parameters discussed in section ii can also receive contribution due to the presence of @xmath115 . however , this is not the case , as it can be verified by noting that no @xmath115-dependent quadratic terms in @xmath72 arise when expanding the right hand side of ( [ solteo ] ) . in this geometry ,
the position of the horizon @xmath188 is defined by @xmath189 , then @xmath190 @xmath191 and thus , for @xmath53 and @xmath192 there is only one horizon . if @xmath193 then the solution is similar to reissner - nordstrm in the sense that there could be two horizons .
when the equality holds one of the horizons is at the origin ( see eq .
( [ cerosdepsi ] ) ) .
figure [ mvsh ] shows the position of the horizon ( @xmath188 ) as a function of the mass @xmath109 for the lovelock - born - infeld black hole and the reissner - nordstrm case ( @xmath161 ) .
and in the reissner - nordstrm case ( rn ) [ @xmath194 , @xmath195 , @xmath53].,width=264 ] these graphics exhibit the crucial difference existing between the lovelock - born - infeld black hole and the reissner - nordstrm black hole ; namely the fact that for a given charge @xmath67 there exists a finite value for the black hole mass @xmath109 such that the black hole geometry presents only one horizon , because the _ internal _ one reached the origin .
this enhancement of the region bounded by both internal and external horizons is also @xmath115-dependent and represents , by itself , one of the principal distinctions between the black hole geometries in both theories .
moreover , fig .
[ mvsh ] shows how the extremal configuration @xmath197 , which is translated into a complicated expression in terms of the parameters @xmath198 and @xmath115 , experiments a displacement for finite values of @xmath115 with respect to the reissner - nordstrm configuration @xmath199 .
figure [ b1 ] shows different behaviors of the solution @xmath84 for different values of b. this parameter controls the cualitative behaviour close to the origin .
the metric for the subcritical case ( @xmath200 ) is not well defined in the whole spacetime . for different values of @xmath115 , for @xmath200 : [ @xmath201 , @xmath194 , @xmath166 , @xmath202 , @xmath53 ] , for @xmath203 [ @xmath204 , @xmath194 , @xmath166 , @xmath202 , @xmath53 ] and for @xmath172 [ @xmath205 , @xmath194 , @xmath166 , @xmath53].,width=264 ]
now , let us move to the thermodynamics of the vacuum solution .
the hawking temperature of the black hole is proportional to the surface gravity @xmath206 and it is given by the formula @xmath207 where @xmath208 is the boltzmann constant . in order to calculate @xmath206 we must evaluate the derivative of @xmath209 in the exterior horizon radius @xmath188 . from the lovelock vacuum solution ( [ solnueva2 ] ) with @xmath53
we obtain @xmath210 .
hence , the surface gravity of the corresponding geometry results @xmath211 and the hawking temperature is given by @xmath212 notice that the temperature goes to zero for bigger black holes ( @xmath213 ) .
this behaviour is the same of the schwarzschild black hole .
a remarkable results is that , unlike the schwarzschild case , for small black holes ( @xmath214 or @xmath215 ) the temperature also goes to zero and there would not be hawking radiation ( fig .
[ hawtem ] ) . on the other hand ,
when @xmath156 in equation ( [ temp1 ] ) we recover the general relativity result since , in that case , the temperature is proportional to the inverse of the radius .
, @xmath53].,width=264 ] the specific heat @xmath216 is given by @xmath217 and , from this , we observe that the position @xmath218 @xmath219 , where the specific heat changes the sign ( fig . [ caloresp ] ) , represents a transition point ( in the fig .
[ hawtem ] this point is the maximum ) .
when @xmath220 the specific heat is positive and the black hole is stable ; while in the range @xmath221 the specific heat is negative and the black hole turns out to be , in this sense , unstable .
the temperature of this transition point is @xmath222 . ,
@xmath53].,width=264 ] regarding the entropy of this black hole solution , we make use of the fact that the first law for this non - rotating black hole can be written as usual , @xmath223 where @xmath224 .
thus the entropy results @xmath225 in the limit @xmath156 the entropy is proportional to the area of the event horizon , in agreement with the bekenstein - hawking formula . on the other hand , the first term in the equation ( [ entrdeh ] ) is a consequence of the gauss - bonnet term .
it is possible to see that this correction term in the entropy formula is proportional to the integral of the scalar curvature invariant of the horizon @xcite .
this computation confirms , by means of a concise example , the observation derived from hamiltonian methods in @xcite about the fact that the entropy of lovelock black holes is not simply the area formula , but topological invariants also contributes to the whole entropy .
this aspect was also discussed in @xcite . besides
, by using the stefan - boltzmann law in five dimensions to compute the flux of thermal radiation we can approximately obtain the black hole evaporation time @xmath226 ; namely @xmath227 with the expression for the hawking temperature ( [ temp1 ] ) and integrating ( [ tiempo ] ) one finds that the lifetime of the lovelock black holes turns out to be infinite .
this is due to the linear dependence between the temperature and the horizon radius for small black holes .
hence , these are eternal black holes .
furthermore , we find that the themodynamics of the lovelock black holes in presence of a negative cosmological constant is also consistent with the known results of einstein gravity solutions within the appropriate range .
certainly , the lanczos lagrangian represents short - distance corrections to einstein gravity which start to be relevant at scales comparable with @xmath228 , which we could denominate _ lanczos - lovelock scale_.
conversely , the value of the cosmological constant @xmath112 introduces another length scale , given by @xmath229 , which marks the scales where the cosmological term starts to dominate .
hence , we can identify the range bounded by the maximum and the minimum of fig .
[ temp6 ] as the scales where the results coming from general relativity fit the thermodynamic behaviour of the lovelock solution .
consequently , for very large scales ( large black holes ) the thermodynamic behaviour is well approximated by the ads - schwarzschild black hole solution while the short distance corrections start to dominate at the scale @xmath230 . ,
@xmath231.,width=264 ] within this context , we can also notice that the thermodynamical properties of the black hole geometry that appears in the case @xmath100 ( with @xmath93 ) exhibit particular features .
indeed , as we early mentioned , this case is analogous to the btz three - dimensional black hole since the phase diagram involving the temperature and the mass is a monotonic function ( see fig .
[ temp34 ] ) describing a phase such that the specific heat is positive everywhere . in some sense , this case can be considered as a transition and , thus , its stability could be an interesting point to be studied with particular attention .
the temperature of these black holes is given by @xmath232 ; this goes to zero for large @xmath45 and identically vanishes for the critical mass @xmath233 . as a function of the mass [ @xmath165 , @xmath234.,width=264 ]
in this paper we studied a solution ( [ solteo ] ) representing five - dimensional charged black holes in lovelock gravity coupled to born - infeld electrodynamics .
the corrections induced by the quadratic terms in the lagrangian ( gauss - bonnet terms ) correspond to short - distance modifications to general relativity and , therefore , the relevant differences between both theories appear for small radius .
the lovelock - born - infeld black holes are characterized by the mass ( @xmath109 ) , the gauss - bonnet constant ( @xmath148 ) , the charge ( @xmath67 ) and the born - infeld constant ( @xmath115 ) .
the constant @xmath148 must be positive in order to have a well behaved solution for all value of @xmath72 .
the metric does not diverge at @xmath81 ; for the critical mass @xmath203 the conical singularity , which is characteristic of the hoffmann - born - infeld solution , is removed ( nevertheless , the origin is a curvature singularity ) .
we commented the differences existing with respect to the reissner - nordstrm black holes and , from this analysis , we observed that , unlike the general relativity , the lovelock - born - infeld theory admits charged black hole solutions with only one horizon .
this is due to the fact that for a given charge @xmath67 , there exist values of mass that force the internal black hole radius to reach the origin .
there is another important distinction between the solutions of both theories .
also in contrast to the schwarzschild solution , the temperature of the black hole remains finite ; in particular , we showed that the temperature goes to zero when the horizon radius approximates to the origin , and there is not hawking radiation .
this leads to find an infinite lifetime for lovelock solutions because the short - distance effects render the small black holes stable .
the temperature of the transition point , where the short - distance corrections start to be relevant , is @xmath235 , which corresponds to black holes with a size comparable to the lovelock - lanczos scale , @xmath236 .
we discussed different limits of the solutions in terms of the coupling constant of lanczos lagrangian @xmath45 and born - infeld lagrangian @xmath115 , and we proved that these limits correspond to the expected geometries .
hence , the solution we present here represents a geometry interpolating between the quoted hoffmann metric for einstein - born - infeld theory and the solution found by wiltshire for the case of lovelock - maxwell field theory .
furthermore , we showed how other solutions studied in the literature are included as particular cases , representing a _ btz phase _ which arises on the curve @xmath237 in the space of parameters .
we discussed the similar features of this phase and the anti - de sitter black holes .
baados m. , teitelboim c. and zanelli j. , _ lovelock - born - infeld theory of gravity _ in j.j .
giambiagi festschrift , la plata ( edited by falomir h. , gamboa r. re . , leal p. and schaposnik f. , world scientific , singapore ) ( 1990 ) .
izaurrieta f. , rodrguez e. and salgado p. , _ phys .
b _ 586 , 397 , ( 2004 ) .
navarro i. and santiago j. , _ jhep _ 0404 , 62 ( 2004 ) .
rong - gen cai and kwang - sup soh , _ phys .
_ d59 , 044013 ( 1999 ) .
lemos j.p . , _ a profusion of black holes from two to ten dimensions _ , proceedings of the xviith brazilian meeting of particle physics and fields , ( edited by adilson j. da silva _ et al _ ) ( 1997 ) . allemandi g. , francaviglia m. and raiteri m. , _ class . and
quant . grav .
_ 20 , 5130 ( 2003 ) .
ilha a. , kleber a. and lemos j.p . ,
_ j. math .
_ 40 , 3509 ( 1999 ) .
baados m. , henneaux m. , teiltelboim c. and zanelli j. , _ phys .
d _ 48 , 1506 ( 1993 ) .
horowitz g. and welch d. , _ phys .
_ 71 , 328 ( 1993 ) .
coussaert o. , henneaux m. and van driel p. , _ class .
_ 12 , 2961 ( 1995 ) .
coussaert o. and henneaux m. , _ phys .
_ 72 , 183 ( 1994 ) .
baados m. , _ black holes in einstein - lovelock gravity _
, talk given at the viii latin american symposium on relativity and gravitation , silarg , sao paulo , july 1993 , [ arxiv : hep - th/9309011 ] .
baados m. , teitelboim c. and zanelli j. , _ phys .
_ d49 , 975 ( 1994 ) .
crisostomo j. , troncoso r.and zanelli j. , _ phys.rev.d _ 62 , 084013 ( 2000 ) .
|
the exact five - dimensional charged black hole solution in lovelock gravity coupled to born - infeld electrodynamics is presented .
this solution interpolates between the hoffmann black hole for the einstein - born - infeld theory and other solutions in the lovelock theory previously studied in the literature .
the conical singularity of the metric around the origin can be removed by a proper choice of the black hole parameters .
the thermodynamical properties of the solution are also analyzed and , in particular , it is shown that the behaviour of the specific heat indicates the existence of a stability transition point in the vacuum solutions .
we discuss the similarities existing between this five - dimensional geometry and the three - dimensional black hole . like btz black hole
, the lovelock black hole has an infinite lifetime .
|
[ i ] the magnetoresistance ( mr ) in a transverse field @xmath6 is one of the most frequently studied characteristics of the two - dimensional ( 2d ) electron gas @xcite . within the drude - boltzmann theory , the longitudinal resistivity of an isotropic degenerate system is @xmath6independent , _
xx(b)=_0=(e^2v_f^2 ) ^-1 , [ drude ] where @xmath7 is the density of states per spin direction , @xmath8 the fermi velocity , and @xmath3 the transport scattering time .
deviations from the constant @xmath9 are customarily called a positive or negative mr , depending on the sign of the deviation .
there are several distinct sources of a non - trivial mr , which reflect the rich physics of the magnetotransport in 2d systems .
first of all , it has been recognized recently that even within the quasiclassical theory memory effects may lead to strong mr @xcite .
the essence of such effects is that a particle `` keeps memory '' about the presence ( or absence ) of a scatterer in a spatial region which it has already visited . as a result ,
if the particle returns back , the new scattering event is correlated with the original one , yielding a correction to the resistivity ( [ drude ] ) .
since the magnetic field enhances the return probability , the correction turns out to be @xmath6-dependent . as a prominent example ,
memory effects in magnetotransport of composite fermions subject to an effective smooth random magnetic field explain a positive mr around half - filling of the lowest landau level @xcite .
another type of memory effects taking place in systems with rare strong scatterers is responsible for a negative mr in disordered antidot arrays @xcite .
however , such effects turn out to be of a relatively minor importance for the low
field quasiclassical magnetotransport in semiconductor heterostructures with typical experimental parameters , while at higher @xmath6 they are obscured by the development of the shubnikov - de haas oscillation ( sdho ) .
second , the negative mr induced by the suppression of the quantum interference by the magnetic field is a famous manifestation of weak localization @xcite .
while the weak - localization correction to conductivity is also related to the return probability , it has ( contrary to the quasiclassical memory effects ) an intrinsically quantum character , since it is governed by quantum interference of time - reversed paths . as a result
it is suppressed already by a classically negligible magnetic field , which changes relative phases of the two paths .
consequently , the corresponding correction to @xmath10 in high - mobility structures is very small and restricted to the range of very weak magnetic fields .
finally , another quantum correction to mr is induced by the electron electron interaction . while this effect is similar to those discussed above in its connection with the return probability ( see sec .
[ iv ] below ) , it is distinctly different in several crucial aspects .
in contrast to the memory effects , this contribution is of quantum nature and is therefore strongly @xmath2-dependent at low temperatures . on the other hand , contrary to the weak localization ,
the interaction correction to conductivity is not destroyed by a strong magnetic field . as a result
, it induces an appreciable mr in the range of classically strong magnetic fields .
this effect will be the subject of the present paper .
it was discovered by altshuler and aronov @xcite that the coulomb interaction enhanced by the diffusive motion of electrons gives rise to a quantum correction to conductivity , which has in 2d the form ( we set @xmath11 ) _ xx ( 1 - 32 f ) t , t1 .
[ aadiff ] the first term in the factor @xmath12 originates from the exchange contribution , and the second one from the hartree contribution . in the weak - interaction regime , @xmath13 , where @xmath14 is the inverse screening length
, the hartree contribution is small , @xmath15 .
the conductivity correction ( [ aadiff ] ) is then dominated by the exchange term and is negative . the condition @xmath4 under which eq .
( [ aadiff ] ) is derived @xcite implies that electrons move diffusively on the time scale @xmath16 and is termed the `` diffusive regime '' .
subsequent works @xcite showed that eq .
( [ aadiff ] ) remains valid in a strong magnetic field , leading ( in combination with @xmath17 ) to a parabolic interaction induced quantum mr , ( 1 - 32 f ) ( _ c)^2 - 1k_f l t , t1 , [ mraa ] where @xmath18 is the cyclotron frequency and @xmath19 the transport mean free path .
indeed , a @xmath2dependent negative mr was observed in experiments @xcite and attributed to the interaction effect .
however , the majority of experiments @xcite can not be directly compared with the theory @xcite since they were performed at higher temperatures , @xmath20 .
( in high - mobility gaas heterostructures conventionally used in mr experiments , @xmath21 is typically @xmath22 and becomes even smaller with improving quality of samples . ) in order to explain the experimentally observed @xmath2-dependent negative mr in this temperature range the authors of refs .
conjectured various _
ad hoc _ extensions of eq .
( [ mraa ] ) to higher @xmath2 .
specifically , ref .
conjectures that the logarithmic behavior ( [ mraa ] ) with @xmath3 replaced by the quantum time @xmath23 is valid up to @xmath24 , while ref
. proposes to replace @xmath25 by @xmath26 .
these proposals , however , were not supported by theoretical calculations .
there is thus a clear need for a theory of the mr in the ballistic regime , @xmath27 .
in fact , the effect of interaction on the conductivity at @xmath28 has been already considered in the literature @xcite .
gold and dolgopolov @xcite analyzed the correction to conductivity arising from the @xmath2-dependent screening of the impurity potential .
they obtained a linear - in-@xmath2 correction @xmath29 . in the last few years , this effect attracted a great deal of interest in a context of low - density 2d systems showing a seemingly metallic behavior @xcite , @xmath30 .
recently , zala , narozhny , and aleiner @xcite developed a systematic theory of the interaction corrections valid for arbitrary @xmath1 .
they showed that the temperature - dependent screening of ref .
has in fact a common physical origin with the altshuler - aronov effect but that the calculation of ref . took only the hartree term into account and missed the exchange contribution . in the ballistic range of temperatures ,
the theory of refs . , predicts , in addition to the linear - in-@xmath2 correction to conductivity @xmath31 , a @xmath16 correction to the hall coefficient @xcite @xmath32 at @xmath33 , and describes the mr in a _ parallel _
field @xcite .
the consideration of ref .
is restricted , however , to _ classically weak _ transverse fields , @xmath34 , and to the _ white - noise _ disorder .
the latter assumption is believed to be justified for si - based and some ( those with a very large spacer ) gaas structures , and the results of refs .
have been by and large confirmed by most recent experiments @xcite on such systems . on the other hand ,
the random potential in typical gaas heterostructures is due to remote donors and has a long range character .
thus , the impurity scattering is predominantly of a small angle nature and is characterized by two relaxation times , the transport time @xmath3 and the single - particle ( quantum ) time @xmath23 governing damping of sdho , with @xmath35 .
therefore , a description of the mr in such systems requires a more general theory valid also in the range of strong magnetic fields and for smooth disorder .
[ a related problem of the tunneling density of states in this situation was studied in ref . .
] in this paper , we develop a general theory of the interaction induced corrections to the conductivity tensor of 2d electrons valid for arbitrary temperatures , transverse magnetic fields , and range of random potential .
we further apply it to the problem of magnetotransport in a smooth disorder at @xmath36 . in the ballistic limit , @xmath37 ( where the character of disorder is crucially important ) ,
we show that while the correction to @xmath10 is exponentially suppressed for @xmath38 , a mr arises at stronger @xmath6 where it scales as @xmath39 .
we also study the temperature - dependent correction to the hall resistivity and show that it scales as @xmath40 in the ballistic regime and for strong @xmath6 .
we further investigate a `` mixed - disorder '' model , with both short - range and long - range impurities present .
we find that a sufficient concentration of short - range scatterers strongly enhances the mr in the ballistic regime .
the outline of the paper is as follows . in sec .
[ ii ] we present our formalism and derive a general formula for the conductivity correction .
we further demonstrate ( sec .
[ iic ] ) that in the corresponding limiting cases our theory reproduces all previously known results for the interaction correction . in sec .
[ iii ] we apply our formalism to the problem of interaction - induced mr in strong magnetic fields and smooth disorder .
section [ iv ] is devoted to a physical interpretation of our results in terms of a classical return probability . in sections
[ v ] and [ vi ] we present several further applications of our theory .
specifically , we analyze the interaction effects in systems with short - range scatterers and in magnetotransport in modulated systems ( lateral superlattices ) .
a summary of our results , a comparison with experiment , and a discussion of possible further developments is presented in sec .
some of the results of the paper have been published in a brief form in the letter @xcite .
we consider a 2d electron gas ( charge @xmath41 , mass @xmath42 , density @xmath43 ) subject to a transverse magnetic field @xmath6 and to a random potential @xmath44 characterized by a correlation function [ corr - function ] u(*r * ) u(*r*)= w(|*r*-*r*| ) with a spatial range @xmath45 .
the total ( @xmath46 ) and the transport ( @xmath47 ) scattering rates induced by the random potential are given by & = & 2_0 ^ 2d2 w ( ) , [ taus ] + 1 & = & 2_0 ^ 2d2 w ( ) ( 1- ) , [ tautr ] where @xmath48 is the scattering cross - section .
we begin by considering the case of smooth disorder , @xmath49 , when @xmath50 ; generalization onto systems with arbitrary @xmath51 will be presented in sec .
we assume that the magnetic field is not too strong , @xmath52 , so that the landau quantization is destroyed by disorder .
note that this assumption is not in conflict with a condition of classically strong magnetic fields ( @xmath53 ) , which is a range of our main interest in the present paper .
we consider two types of the electron - electron interaction potential @xmath54 : ( i ) point - like interaction , @xmath55 , and ( ii ) coulomb interaction , @xmath56 . in order to find the interaction - induced correction @xmath0 to the conductivity tensor
, we make use of the `` ballistic '' generalization of the diffuson diagram technique of ref . .
we consider the exchange contribution first and will discuss the hartree term later on . within the matsubara formalism , the conductivity is expressed via the kubo formula through the current - current correlation function , @xmath57 where @xmath58 is the bosonic matsubara frequency .
diagrams for the leading - order interaction correction are shown in fig .
[ fig1 ] and can be generated in the following way .
first , there are two essentially different ways to insert an interaction line into the bubble formed by two electronic green s function .
second , one puts signs of electronic matsubara frequencies in all possible ways . on the third step ,
one connects lines with opposite signs of frequencies @xmath59 by impurity line ladders ( which are not allowed to cross each other ) . finally , in the case of the diagram _ a _ , where four electronic lines form a `` box '' , one should include two additional diagrams , _ b _ and _ c _ , with an extra impurity line ( `` hikami box '' ) . .
the wavy ( dashed ) lines denote the interaction ( impurity scattering ) , the shaded blocks are impurity ladders , and the @xmath60 symbols denote the signs of the matsubara frequencies .
the diagrams obtained by a flip and/or by an exchange @xmath61 should also be included .
`` inelastic '' part of the diagrams _ f _ , _ g _ is canceled by a contribution of the coulomb - drag type , see appendix [ a1 ] .
, width=302 ] the impurity line ladders are denoted by shaded blocks in fig .
[ fig1 ] ; we term them `` ballistic diffusons '' . formally , the ballistic diffuson is defined as an impurity average ( denoted below as @xmath62 ) of a product of a retarded and advanced green s functions , & & d(i_m , i_n;_1,_2;_3,_4 ) + & & = ( -_m_n)g(_1,_2;i_m ) g(_3,_4;i_n ) _ imp .
+ [ balldiffrr ] following the standard route of the quasiclassical formalism @xcite , we perform the wigner transformation , & & d(i_m - i_n;*r*_1,_1;*r*_2,_2)= dd + & & e^-i[_1-(e / c)*a(*r_1 ) ] * * e^-i[_2-(e / c)*a(*r_2)] * * + & & ( i_m , i_n;*r*_1 , ; * r*_2, ) , [ balldiff ] where @xmath63 , @xmath64 , @xmath65 , and @xmath66 . note that the factors depending on the vector potential make the ballistic diffuson ( [ balldiff ] ) gauge - invariant . finally , we integrate out the absolute values of momenta @xmath67 and get the final form of the ballistic diffuson & & d(i_l;*r*_1,_1;*r*_2,_2 ) + & & = 12 ( i_l;*r*_1,_1;*r*_2,_2 ) , + [ calbrn ] which describes the quasiclassical propagation of an electron in the phase space from the point @xmath68 to @xmath69 . here
@xmath70 is the unit vector characterizing the direction of velocity on the fermi surface .
the ballistic diffuson satisfies the quasiclassical liouville - boltzmann equation ( i_l,;, ) + & = & 2(- ) , [ lb ] where @xmath71 is the polar angle of @xmath72 and @xmath73 is the collision integral , determined by the scattering cross - section @xmath74 for the case of a smooth disorder , the collision integral is given by = -1 ^ 2 ^ 2 .
[ csmooth ] in contrast to the diffusive regime , where @xmath75 has a universal and simple structure @xmath76 determined by the diffusion constant @xmath77 only , its form in the ballistic regime is much more complicated .
we are able , however , to get a general expression for @xmath0 in terms of the ballistic propagator @xmath78 .
the temperature range of main interest in the present paper is restricted by @xmath79 , since at higher @xmath2 the mr will be small in the whole range of the quasiclassical transport @xmath80 ( see below ) . in this case
the ladders are dominated by contributions with many ( @xmath81 ) impurity lines .
we will assume this situation when evaluating diagrams in the present subsection .
a general case of arbitrary @xmath82 and @xmath83 will be addressed in sec .
[ iib ] . from the diagram d.,width=302 ] we start with the diagrams _ d _ and _ e _ that give rise to the logarithmic correction in the diffusive regime @xcite .
let us fix the sign of the external frequency , @xmath84 each of the diagrams _ d _ and _ e _ generates four diagrams by a flip with respect to the horizontal line or by exchange @xmath61 , see fig . [ figd ] .
consider first the diagram @xmath85 .
there are two triangular boxes containing each a current vertex and three electron green s functions ( fig .
[ kroko ] ) . in the quasiclassical regime @xmath80
one may neglect the effect of magnetic field on the green s functions ( keeping @xmath86 in the ballistic propagators only ) . furthermore , using @xmath87 , we neglect the difference in momenta and frequencies in the green s functions , since typical values of frequencies @xmath88 , @xmath89 and momenta @xmath90 carried by the ballistic diffusons are set by the temperature .
each triangle then reads _ ( ) & = & em + & & i
2_s^2 e v_f n _ , where @xmath91 . combining the triangles with the three ballistic propagators separated by the impurity lines ( see fig . [ kroko ] ) , we obtain the following expression for the electronic part of the diagram @xmath85 , & & ( 2)^3_i=1 ^ 6d_i2 d(i_l,;,_1)w(_1-_2 ) _ ( _ 2 ) + & & w(_2-_3)d(i_l - i_k,;_3,_4 ) w(_4-_5 ) + & & _ ( _ 5)w(_5-_6 ) d(i_l,;_6, ) + & & ^d_(i_l
,- i_k,;, ) [ bdef ] . in
what follows we will use for brevity a short - hand notation @xmath92 for the l.h.s . of ( [ bdef ] ) and
analogous notations for other structures of this type . making use of the small - angle nature of scattering in a smooth random potential
, we can replace the @xmath93 factors in ( [ bdef ] ) by @xmath94 , yielding & & b^d_(q , i_l , i_k;, ) + & & -d(i_l , ) n_(i_l+i_k , ) n_(i_l , ) . in the exchange term ( calculated in the present subsection )
this structure is further integrated over the angles @xmath95 and @xmath96 , b^d_(i_l , i_k,)= ^d_(i_l , i_k,;, ) . [
15 ] the angular brackets @xmath97 denote averaging over velocity directions , e.g. @xmath98 the fermionic frequencies obey the inequalities @xmath99 , @xmath100 , and @xmath101 , which implies @xmath102 , so that the summation over @xmath103 gives the factor @xmath104 . drawn in a different way in order to visualize the structure of eq .
( [ bdef ] ) . the dashed frame encloses the electronic part @xmath105.,width=302 ] the diagram @xmath106 has the same structure ( both triangles have opposite signs , thus the total sign remains unchanged ) , but the frequency summation is restricted by @xmath107 , @xmath108 , and @xmath109 , yielding the factor @xmath110 in the conductivity correction .
the diagrams @xmath111 and @xmath112 obtained from @xmath85 and @xmath106 by a flip ( or , equivalently , by reversing all arrows ) double the result . combining the four contributions and changing sign of the summation variable , @xmath113 in @xmath106 and @xmath112 terms
, we have & & _
^d(i_k ) = -8_0t^2_k + & & , [ sigmad ] where @xmath114 is the interaction potential equal to a constant @xmath115 for point - like interaction and to u(i_l,)= [ screen ] for screened coulomb interaction . in ( [ sigmad ] ) we used the fact that @xmath116 and @xmath117 .
equation ( [ screen ] ) is a statement of the random - phase approximation ( rpa ) , with the polarization operator given by [ polarization ] ( i_l,)=2 .
the first term ( unity ) in square brackets in ( [ polarization ] ) comes from the @xmath118 and @xmath119 contributions to the polarization bubble , while the second term is generated by the @xmath120 contribution ( ballistic diffuson ) .
the diagrams _ e _ are evaluated in a similar way . in all four diagrams of this type one of the electron triangles
is the same as in diagrams _
d _ while another one has an opposite sign .
the structures arising after integrating out fast momenta in electron bubbles coincide with those of _
d_-type ( @xmath121 ) .
summation over the fermionic frequency @xmath103 is constrained by the condition @xmath102 for all the @xmath122-type diagrams .
the correction due to the diagrams _ e _ therefore reads & & _
^e(i_k ) = 8_0t^2_k _ _ l>_k u(i_l , ) + & & .
[ sigmae ] we see that the first term in square brackets in ( [ sigmae ] ) cancels the first term in ( [ sigmad ] ) .
thus , the sum of the contributions of diagrams _ d _ and _ e _ takes the form _
^d+e(i_k ) & = & -4_0t_k , [ sigmae+d ] where we introduced a notation ^_(i_l , i_k)= u(i_l , ) b^_(i_l , i_k , ) , [ f - def ] with the index @xmath123 labeling the diagram . similarly , we obtain for the diagram _ h _ _ ^h(i_k ) & = & -4_0t_k , [ sigmah ] with b^h_(i_l , i_k,)=-2 t _
n _ d(i_l+i_k , ) n _ d(i_l , ) .
+ [ bh ] the tensor @xmath124 appearing in ( [ bh ] ) describes the renormalization of a current vertex connecting two electronic lines with opposite signs of frequencies , [ t_alphabeta ] t_&=&2.n_n_|_q=0,0= _ e^2v_f^2 + & = & 1+_c^2 ^ 2 ( cc 1 & -_c + _ c & 1 ) .
we turn now to diagrams _ f _ and _ g_. the expressions for the corresponding contributions read @xmath125,\label{sigmaf } \\ & & \delta\sigma_{\alpha\beta}^{g}(i\omega_k ) = { 4\sigma_0\over \tau}{t\over \omega_k } \left[\sum_{\omega_l=0}^{\omega_k}(\omega_k-\omega_l ) \phi^{f}_{\alpha\beta}(i\omega_l , i\omega_k ) \right .
\nonumber \\ & & + \left .
\sum_{-\omega_k<\omega_l<0 } \!\!(\omega_k+\omega_l ) \phi^{f}_{\alpha\beta}(-i\omega_l , i\omega_k ) \right ] , \label{sigmag}\end{aligned}\ ] ] with b^f_(i_l , i_k,)= t
d(i_l+i_k , ) n_t_. + [ bf ] the sum of the contributions _
f _ and _ g _ is therefore given by _
^f+g(i_k ) & = & -4_0t_k . [ sigmaf+g ]
we see that when the diagrams @xmath126 and @xmath127 are combined , the same matsubara structure as for other diagrams [ eqs .
( [ sigmae+d ] ) , ( [ sigmah ] ) ] arises . in other words ,
the role of the diagrams _
g _ is to cancel the extra contribution of diagrams _ f _ , which has a different matsubara structure .
a word of caution is in order here . in our calculation
we have set the value of velocity coming from current vertices to be equal @xmath8 , thus neglecting a particle - hole asymmetry .
if one goes beyond this approximation and takes into account the momentum - dependence of velocity ( violating the particle - hole symmetry ) , the above cancellation ceases to be exact and an additional term with a different matsubara structure arises in @xmath128 . after the analytical continuation is performed , the corresponding correction to the conductivity has a form ^inel_&= & -2_0_-^ + & & b^f_(,q ) i m u(,q ) , [ sigma - inel ] characteristic for effects governed by inelastic scattering .
this contribution is determined by real inelastic scattering processes with an energy transfer @xmath129 and behaves ( in zero magnetic field ) as @xmath130 .
this implies that the corresponding resistivity correction , @xmath131 is independent of disorder .
however , such a correction should not exist because of total momentum conservation .
indeed , an explicit calculation ( see appendix [ a1 ] ) shows that this term is canceled by the aslamazov - larkin - type diagrams analogous to those describing the coulomb drag .
finally , we consider the diagrams _ a , b _ , and _ c_. already taken separately , each of them has the expected matsubara structure ( contrary to the diagrams _ d , e _ and _ f , g _ , which should be combined to get this structure ) .
however , another peculiarity should be taken into account .
the diagrams _ a , b _ , and _ c _ form together the hikami box , so that their sum is smaller by a factor @xmath132 than separate terms .
therefore , some care is required : subleading terms of order @xmath83 should be retained when contributions of individual diagrams are calculated .
the result reads @xmath133 .
\label{sigmaabc } \end{aligned}\ ] ] here the contributions of individual diagrams _ a , b , _ and _ c _ have the form [ ba ] & & b^a_(i_l , i_k , ) + & & = 12 t _ t _
( i_l , ) d(i_l , ) + & & + 12 t _
t_(i_l , ) , where the matrix @xmath134 has the same form as @xmath135 with a replacement @xmath136 , [ bb ] b^b_(i_l , i_k,)= -12_st _ t _
( i_l , ) d(q , i_l ) , and & & b^c_(i_l , i_k,)= + & & -12(1_s-1 ) t _ t _
( i_l , ) d(i_l , ) .
+ [ bc ] we see that although each of the expressions ( [ ba ] ) , ( [ bb ] ) , and ( [ bc ] ) depends on @xmath23 , the single - particle time disappears from the total contribution of the hikami - box , b^a+b+c_(i_l , i_k,)&= & 12 t _ ( i_l , ) d(i_l , ) + & + & 12 t _ ( i_l,)t_. [ babc ] the total correction to the conductivity tensor is obtained by collecting the contributions ( [ sigmae+d ] ) , ( [ sigmah ] ) , ( [ sigmaf+g ] ) , and ( [ sigmaabc ] ) . carrying out the analytical continuation to real frequencies ,
we get _ ( ) & = & _ 0i_-^ d + & & [ _ ( + , ) - _ ( , ) ] , [ e34 ] where _ ( , ) & = & ^a+b+c _ ( , ) + ^d _ ( , ) + & + & ^f_(,)+^h _ ( , ) . [ f - total ] we are interested in the case of zero external frequency , @xmath137 , when eq .
( [ e34 ] ) can be rewritten as _ & = & -_0i
_ -^d + & & _ ( , 0 ) . [ anal - co ] recalling the definition ( [ f - def ] ) of @xmath138 , we finally arrive at the following result @xmath139\nonumber \\ & \times&\int \frac{d^2{\bf q}}{(2\pi)^2}\ { \rm i m } \left[\ u(\omega,{\bf q})\ b_{\alpha\beta}(\omega,{\bf q})\ \right ] , \nonumber \\ \label{sigma } \end{aligned}\ ] ] where the tensor @xmath140 is given by @xmath141 the first term in ( [ bwq ] ) originates from the diagrams _ a , b , c _ , the second term from _ a , f , g _ , the third term from _
h _ , and the last one from _ d _ and _
e_. we remind the reader that this result has been obtained under the assumption @xmath142 ; generalization to arbitrary @xmath83 and @xmath143 will be considered in sec .
it will be shown there that the conductivity correction retains the form ( [ sigma ] ) in the general case but the expression ( [ bwq ] ) for @xmath144 is slightly modified . in the previous subsection
we have derived the formula for the correction to the conductivity tensor for the case of a smooth disorder ( with @xmath145 ) assuming @xmath146 . since characteristic momenta @xmath147 and frequencies @xmath148
are set by the temperature , this assumption implies @xmath149 and @xmath150 .
this allowed us to simplify the calculation by neglecting the @xmath147 and @xmath148 dependence of green s functions connecting ballistic diffusons and by considering only the ladders with many impurity lines .
furthermore , we have used the small - angle nature of scattering when calculating the hikami box contribution ( [ babc ] ) .
we are now going to discuss the general case of arbitrary @xmath83 and @xmath82 .
it turns out that the expressions ( [ sigmae+d ] ) , ( [ sigmah ] ) , and ( [ sigmaf+g ] ) for the contribution of the diagrams @xmath151 derived in the case of a smooth disorder remain valid in the general situation .
the simplest way to show this is to use the following technical trick ( cf
. refs . ) .
one can add to the system an auxiliary weak smooth random potential with a long transport scattering time @xmath152 but short single - particle @xmath153 , such that @xmath154 .
this potential will not affect the quasiclassical dynamics and thus should not change the result .
on the other hand , it allows us ( in view of the condition @xmath154 ) to perform the gradient and frequency expansion in green s functions as was done in sec .
adding such an auxiliary disorder amounts to a re - distribution between quantum and quasiclassical degrees of freedom : all the information about the real disorder is now contained in the ballistic propagators .
it can be verified by a direct calculation ( without using the additional disorder ) that the above procedure yields the correct result .
it remains to consider the hikami - box contribution ( [ sigmaabc ] ) . when calculating it in sec .
[ iia ] , we used the small - angle nature of scattering implying that a single scattering line inserted between two ballistic propagators approximately preserves the direction of velocity , @xmath155 and @xmath156 . in the more general situation , when the scattering is at least partly of the large
angle character , this is no longer valid and eq .
( [ babc ] ) acquires a slightly more complicated form , [ hikami ] & & b^a+b+c_(,i_l , i_k ) = + & & t _ [ _ d- 2n _ w n_d ] t _ + & & + 12 t _
, where @xmath157 .
summarizing the consideration in this subsection , in the general situation the interaction correction retains the form ( [ sigma ] ) with the tensor @xmath140 given by [ bwq_general ] & & b_(,*q * ) = t _ t _ + & & + t_(_2 - n _ d n _ ) t _ + & & -2 t _
d n _ d - n _ d n _ d .
the correction @xmath158 to the resistivity tensor is then immediately obtained by using @xmath159 . this yields @xmath160\nonumber \\ & \times&\int \frac{d^2{\bf q}}{(2\pi)^2}\ { \rm i m } \left[\ u(\omega,{\bf q})\ b^{(\rho)}_{\alpha\beta}(\omega,{\bf q})\ \right ] , \nonumber \\
\label{rho } \end{aligned}\ ] ] where the tensor @xmath161 is related to @xmath162 , eq .
( [ bwq_general ] ) , via b^ ( ) _ = ( t^-1 ) _ b _ ( t^-1)_. [ brho ] explicitly , corrections to the components of the resistivity tensor are expressed through @xmath163 and @xmath164 as follows [ rhoxx - sigma ] _
xx&=&_0 ^ 2 [ ( _ c^2 ^ 2 - 1)_xx+2_c_xy ] , + _ xy&=&_0 ^ 2 [ ( _ c^2 ^ 2 - 1)_xy-2_c_xx ] . [ rhoxy - sigma ] note that the results ( [ bwq ] ) , ( [ bwq_general ] ) for @xmath140 satisfy the requirement b_(,0)=0 , [ b(q=0 ) ] as follows from @xmath165 and @xmath166 the condition ( [ b(q=0 ) ] ) implies that spatially homogeneous fluctuations in the potential do not change the conductivity , see refs . , for discussion .
having obtained the general formula , we will now demonstrate that it reproduces , in the appropriate limits , the previously known results for the interaction correction . specifically , in sec .
[ iic1 ] we will consider the diffusive limit @xmath4 studied in refs . , while sec .
[ iic2 ] is devoted to the @xmath167 case with a white - noise disorder addressed in refs . .
[ iic3 ] we will analyze how the linear - in-@xmath2 asymptotics of @xmath168 in the ballistic regime obtained in ref
. for a white - noise disorder depends on the character of the random potential .
we begin by considering the diffusive limit @xmath4 in which we reproduce ( for arbitrary @xmath6 and disorder range ) the logarithmic correction ( [ aadiff ] ) , ( [ mraa ] ) determined by the diagrams _ a - e_. let us briefly outline the corresponding calculation .
the propagator for @xmath169 can be decomposed as @xmath170 , where @xmath171 is singular , while @xmath172 is finite ( regular ) at @xmath173 , see e.g. refs . .
the singular contribution is governed by the diffusion mode and has the form [ see eq .
( [ diff - prop ] ) ] @xmath174 where @xmath175 is the diffusion constant in the presence of a magnetic field and @xmath176 the leading - order contribution of the diagrams @xmath177 and @xmath178 ( that containing two singular diffusons @xmath179 ) is exactly canceled by the part of the diagrams @xmath45 and @xmath122 with the structure @xmath180 , i.e. with one regular part of the propagator inserted between two singular diffusons , @xmath181 .
indeed , in view of @xmath182 , the latter contribution reduces to @xmath183 , while the diagrams @xmath177 and @xmath178 yield & & 12^s^2 t _ t _ + & & = 12^s ^2 t _ , [ cancelabc ] where @xmath184 is the antisymmetric tensor , @xmath185 .
it remains thus to calculate only the contribution of the diagrams @xmath186 with three singular diffusons , @xmath187 \nonumber \\ & \times & \int{d^2q\over ( 2\pi)^2}{\rm i m } \frac{\langle { \cal d}^{\rm s}n_\alpha { \cal d}^{\rm s } n_\beta { \cal d}^{\rm s}\rangle } { 1+i\omega\langle{\cal d}^{\rm s}\rangle}\nonumber \\ & \simeq & { 2e^2v_f^2\over \pi(1+\omega_c^2\tau^2)^2}\int_t^{1/\tau } d\omega \nonumber \\ & \times & \int{d^2q\over ( 2\pi)^2}{\rm im}\frac{(-iq_\alpha l)(-iq_\beta l)}{dq^2(dq^2-i\omega)^2 } \nonumber \\ & = & { e^2\over 2\pi^2}\ln(t\tau)\delta_{\alpha\beta } , \label{gmdiff}\end{aligned}\ ] ] in agreement with @xcite .
the result for a point - like interaction differs only by a factor @xmath188 .
we allow now for arbitrary @xmath1 but consider the limit of zero magnetic field assuming a white - noise disorder ( @xmath189 and @xmath190 ) , which is the limit studied in refs . .
the contribution ( [ hikami ] ) of the diagrams _ a , b , c _ takes for the white - noise disorder the form & & b^a+b+c _ = 12 t_t_. [ wnhik ] using now the explicit form of the ballistic propagator for the case of white - noise disorder and @xmath167 [ eqs .
( [ prop - wn ] ) , ( [ g0-def ] ) , ( [ d0b=0 ] ) , ( [ g0b=0 ] ) , and ( [ g_0(b ) ] ) ] we recover the results for @xmath191 and @xmath192 obtained in a different way in refs . and , see appendix [ a2 ] . in the ballistic limit @xmath195 and for white - noise disorder the result of ref .
( recovered in sec .
[ iic2 ] and appendix [ a2 ] ) yields a linear - in-@xmath2 conductivity correction , @xmath196 for the point - like interaction and @xmath197 for the coulomb interaction .
the question we address in this subsection is how this behavior depends on the nature of disorder [ i.e. on the scattering cross - section @xmath198 . in order to get the @xmath1 ballistic asymptotics ,
it is sufficient to keep contributions to ( [ bwq_general ] ) with a minimal number of scattering processes .
specifically , the propagator @xmath75 in the first and the third terms of ( [ bwq_general ] ) can be replaced by the free propagator , _
f(,;,)=2(- ) -i(+i0)+iqv_f(-_q ) , [ d - free ] while in the second term it should be expanded up to the linear term in the scattering cross - section @xmath199 [ the second term produces then the same contribution as the first term in ( [ bwq_general ] ) ] .
the last ( fourth ) term in ( [ bwq_general ] ) does not contribute to the @xmath1 asymptotics .
we get therefore b_xx & & 2 ^ 2(_f w d_f- 2_f n_x w n_x d_f ) + & - & 2n_xd_f n_xd_f .
[ bwq_ball ] let us consider first the case of a short - range interaction , @xmath200 .
the structure of eqs .
( [ sigma ] ) , ( [ bwq_ball ] ) implies that the interaction correction is governed by returns of a particle to the original point in a time @xmath201 after a single scattering event .
it follows that the coefficient in front of the linear - in-@xmath2 term is proportional to the backscattering probability @xmath202 , [ sigma_ball ] _
xx=2v_0 e^2 2w ( ) t^2 .
as shown in appendix [ a3 ] , this result remains valid in the case of coulomb interaction , with the factor @xmath203 replaced by unity .
this shows that in the ballistic limit the coulomb interaction is effectively reduced to the statically screened form , @xmath204 when the leading contribution to @xmath191 is calculated . according to ( [ sigma_ball ] ) , in a smooth disorder with a correlation length @xmath205 the @xmath1 contribution
is suppressed by an exponentially small factor @xmath206 .
in fact , for a smooth disorder the linear term represents the leading contribution for @xmath207 only . in the intermediate range @xmath208 the dominant return processes
are due to many small - angle scattering events .
however , the corresponding return probability is also exponentially suppressed @xmath209 for relevant ( ballistic ) times @xmath210 , yielding a contribution @xmath211 $ ] .
thus , the interaction correction in the ballistic regime is exponentially small at @xmath193 for the case of smooth disorder .
moreover , the same argument applies to the case of a non - zero @xmath6 , as long as@xcite @xmath212 . in any realistic system
there will be a finite concentration of residual impurities located close to the electron gas plane and inducing large - angle scattering processes .
in other words , a realistic random potential can be thought as a superposition of a smooth disorder with a transport time @xmath213 and a white - noise disorder characterized by a time @xmath214 .
neglecting the exponentially small contribution of the smooth disorder to the linear term , we then find that the ballistic asymptotics ( [ sigma_ball ] ) of the interaction correction takes the form [ sigma_ball_mixed ] = e^2 _ wn t\ { ll 2v_0 , & point - like , + 1 , & coulomb , .
where @xmath215 is the total transport scattering rate .
if the transport is dominated by the smooth disorder , @xmath216 , the coefficient of the @xmath1 term is thus strongly reduced as compared to the white - noise result of ref . .
finally , it is worth mentioning that in addition to the @xmath1 term corresponding to the lower limit @xmath217 of the frequency integral in ( [ sigma ] ) , there is a much larger but @xmath2-independent contribution @xmath218 governed by the upper limit @xmath219 .
this contribution is just an interaction - induced fermi - liquid - type renormalization of the bare ( noninteracting ) drude conductivity .
we have shown in sec .
[ iic ] that due to small - angle nature of scattering in a smooth disorder the interaction correction is suppressed in the ballistic regime @xmath194 in zero ( or weak ) magnetic field .
the situation changes qualitatively in a strong magnetic field , @xmath53 and @xmath220 .
the particle experiences then within the time @xmath221 multiple cyclotron returns to the region close to the starting point .
the corresponding ballistic propagator satisfies the equation ( [ lb ] ) with the collision term ( [ csmooth ] ) .
the solution of this equation in the limit of a strong magnetic field , @xmath53 , is presented in appendix [ a4 ] . for calculation of the leading order contribution to @xmath191 and @xmath222 ,
the following approximate form is sufficient : & & d(,;, ) = + & & , + & & ^s(,;, ) + d^reg(,;, ) , [ diffuson ] where @xmath223 and @xmath224 , and the polar angles of velocities are counted from the angle of @xmath225 .
equation ( [ diffuson ] ) is valid under the assumption @xmath226 .
we will see below that the characteristic momenta @xmath147 are determined by the condition @xmath227 , so that the above assumption is justified in view of @xmath220 .
furthermore , this condition allows us to keep only the first ( singular ) term @xmath171 in square brackets in ( [ diffuson ] ) when calculating @xmath228 , = j_0 ^ 2(qr_c)dq^2-i , [ < d > ] where @xmath229 is the bessel function .
moreover , the formula ( [ bwq ] ) for @xmath230 can be cast in a form linear in @xmath75 by using & = & -i , [ dd ] + n _ n_&= & i v_f
n _ , [ ndnd ] + n_x d n_x d&= & -1 2 v_f^2 .
[ dndnd ] therefore , it is again sufficient to take into account only the first term in ( [ diffuson ] ) for calculation of @xmath230 if the identities ( [ dd ] ) , ( [ ndnd ] ) , and ( [ dndnd ] ) are used .
( of course , @xmath230 can also be evaluated directly from eq .
( [ bwq ] ) , but then the second ( regular ) term @xmath172 in ( [ diffuson ] ) has to be included . )
combining all four terms in ( [ bwq ] ) , we get b_xx(,q)&=&j_0 ^ 2(qr_c)(_c)^2 + & = & .
[ bxx ] in the second line we introduced dimensionless variables @xmath231 .
note that eqs .
( [ < d > ] ) , ( [ bxx ] ) differ from those obtained in the diffusive regime by the factor @xmath232 only .
this is related to the fact that the motion of the guiding center is diffusive even on the ballistic time scale @xmath233 ( provided @xmath234 ) , while the additional factor corresponds to the averaging over the cyclotron orbit ( see sec .
[ iv ] below ) .
we turn now to the calculation of @xmath235 . substituting ( [ diffuson ] ) in ( [ bwq ] )
, we classify the obtained contributions according to powers of the small parameter @xmath236 .
the leading contributions are generated by the first and the last terms in ( [ bwq ] ) and are of order @xmath236 , i.e. larger by factor @xmath237 as compared to @xmath230 , eq .
( [ bxx ] ) .
( this extra factor of @xmath237 is simply related to @xmath238 . ) however , these leading contributions cancel , & & .|_order 1/ + & = & 2^sd^s- ^s n_x d^reg n_y d^s [ 1-over - beta ] + & = & -2 ^3 + 2 ^3 = 0 , as in the diffusive limit , see the text above eq .
( [ cancelabc ] ) . to evaluate terms of higher order in @xmath236 ,
we need a more accurate form of the propagator ( [ diffuson ] ) .
since the contributions of order @xmath239 to @xmath235 turn out to cancel as well , we have to know the propagator with the accuracy allowing to evaluate the terms of order @xmath240 .
to simplify the calculation , we use again the identities ( [ dd ] ) and ( [ ndnd ] ) . as to eq .
( [ dndnd ] ) , it can not be generalized onto @xmath241-component of the tensor , and we use instead n_x d n_y d=
i v_f n_yd .
[ dxdyd ] it is then sufficient to calculate the propagator @xmath75 up to the @xmath239 order .
this is done in appendix [ a4 ] , see eqs .
( [ dsing-2order])-([psirl - n-2 ] ) . substituting this result for @xmath75 in eq .
( [ bxx ] ) and combining all terms , we get after some algebra b_xy(,q ) & = & - .
[ bxy ] we see that similarly to ( [ bxx ] ) the kernel @xmath242 has a diffusive - type structure with @xmath243 in denominator reflecting the diffusion of the guiding center , while the bessel functions describe the averaging over the cyclotron orbit .
clearly , both kernels ( [ bxx ] ) and ( [ bxy ] ) vanish at @xmath244 , as required by ( [ b(q=0 ) ] ) .
to find the interaction correction to the conductivity , we have to substitute eqs .
( [ bxx ] ) and ( [ bxy ] ) in the formula ( [ sigma ] ) .
we consider first the simplest situation , when the interaction @xmath245 in ( [ sigma ] ) is of point - like character , @xmath246 .
using @xmath247 , we see that all the @xmath6-dependence drops out from @xmath191 , and the exchange contribution reads
_ xx&=&-8e^2 v_0 _ 0^ + & & _
0^ i m .
[ sxxqw ] to simplify the result ( [ sxxqw ] ) , it is convenient to perform a fourier transformation with respect to @xmath248 ( which corresponds to switching to the time representation ) & & im_0^ f ( ) + & & = _ 0^dt ( t ) .
[ fourier ] the integral over @xmath249 is then easily evaluated , yielding _
xx&=&-e^22 ^ 2v_0 g_0(t ) , [ deltaint ] + g_0(x)&=&^2 x^2 _ 0^ [ g0(x ) ] + & & , where @xmath250 and @xmath251 are modified bessel functions .
the hartree term in this case is of the opposite sign and twice larger due to the spin summation ( we neglect here the zeeman splitting and will return to it later ) .
( a ) and @xmath252 ( b ) determining the @xmath2-dependence of the exchange term for point - like , eq .
( [ deltaint ] ) , and coulomb , eq .
( [ exchange ] ) , interaction , respectively .
diffusive and ballistic asymptotics , eq . ( [ g0 ] ) and eq .
( [ gf ] ) , are also shown .
, title="fig:",width=302 ] ( a ) and @xmath252 ( b ) determining the @xmath2-dependence of the exchange term for point - like , eq .
( [ deltaint ] ) , and coulomb , eq .
( [ exchange ] ) , interaction , respectively .
diffusive and ballistic asymptotics , eq . ( [ g0 ] ) and eq .
( [ gf ] ) , are also shown .
, title="fig:",width=302 ] it follows from eqs .
( [ bxx ] ) and ( [ bxy ] ) that the correction to the hall conductivity is smaller by the factor @xmath253 as compared to ( [ deltaint ] )
. this implies , according to ( [ rhoxx - sigma ] ) that in a strong magnetic field the correction to the longitudinal resistivity is governed by @xmath191 , = ( _ c)^2_xx_0 , [ parabola ] similarly to the diffusive limit ( [ mraa ] ) .
in fact , it turns out that the relation ( [ parabola ] ) holds in a strong magnetic field , @xmath254 for arbitrary disorder and interaction , see below . on the other hand , as is seen from ( [ rhoxy - sigma ] ) , contributions of both @xmath191 and @xmath255 to @xmath192 are of the same order in @xmath253 .
we will return to the calculation of @xmath192 in sec .
[ iiig ] .
the mr @xmath9 is thus quadratic in @xmath6 , with the temperature dependence determined by the function @xmath256 , which is shown in fig .
[ fig2]a . in the diffusive ( @xmath257 ) and ballistic ( @xmath258 ) limits the function @xmath259 has the following asymptotics [ g0 ] g_0(x)\ { ll -x + const , & x1 , + c_0x^-1/2 , & x 1 , . with c_0=3(3/2)160.276 [ c0 ] (
here @xmath260 is the riemann zeta - function ) .
let us note that the crossover between the two limits takes place at numerically small values @xmath261 ( a similar observation was made in refs . ) .
this can be traced back to the fact that the natural dimensionless variable in ( [ deltaint ] ) is @xmath262 .
for the case of the coulomb interaction the result turns out to be qualitatively similar .
substituting ( [ < d > ] ) in ( [ screen ] ) and neglecting the first term @xmath263 in the denominator of ( [ screen ] ) , we obtain the effective interaction u(,)=12 .
[ u - rpa - b ] inserting ( [ u - rpa - b ] ) and ( [ bxx ] ) into ( [ sigma ] ) , we get the exchange ( fock ) contribution ^f_xx&=&-e^2 ^2 _ 0^ d + & & i m _ 0^ qdq .
+ [ coulsxxqw ] using ( [ parabola ] ) we find the mr & & ^f_xx(b)_0 = -(_c)^2k_fl g_f(t ) , [ exchange ] + & & g_f(x ) = 32 ^2 x^2 _ 0^dq
q^3 j_0 ^ 2(q ) [ gfx ] + & & _ n=1^ .
note that in contrast to the case of a point - like interaction , a transformation to the time representation does not allow us to simplify ( [ coulsxxqw ] ) , since the resulting @xmath249-integral can not be evaluated analytically .
we have chosen therefore to perform the @xmath248-integration , which results in an infinite sum ( [ gfx ] ) .
this amounts to returning to the matsubara ( imaginary frequency ) representation and is convenient for the purpose of numerical evaluation of @xmath264 . in the diffusive ( @xmath257 ) and ballistic ( @xmath258 ) limits this function has the asymptotics [ gf ] g_f(x)\ { ll -x + const , & x1 , + x^-1/2 , & x 1 , . and
is shown in fig .
[ fig2]b .
we turn now to the hartree contribution .
the corresponding diagrams can be generated in a way similar to exchange diagrams ( sec .
[ iia ] ) but in this case one should start from two electron bubbles connected by an interaction line .
there are again two distinct ways to generate a skeleton diagram : two current vertices can be inserted either both in the same bubble or in two different bubbles .
then one puts signs of matsubara frequencies in all possible ways and insert ballistic diffusons correspondingly .
the obtained set of diagrams is shown in fig .
[ hartree - dia ] there is one - to - one correspondence between these hartree diagrams and the exchange diagrams of fig .
[ fig1 ] , which is reflected in the labeling of diagrams . .
the diagrams are labeled in the way as their exchange counterparts in fig .
the diagrams obtained by a flip and/or by an exchange @xmath61 should also be included.,width=302 ] as seen from comparison of figs .
[ fig1 ] and [ hartree - dia ] , the electronic part @xmath265 of each hartree diagram is identical to that of its exchange counterpart .
the only difference is in the arguments of the interaction propagator , @xmath266 , where @xmath95 and @xmath96 are polar angles of the initial and final velocities [ cf .
( [ bdef ] ) , ( [ 15 ] ) ] .
therefore , in the first order in the interaction , the hartree correction to conductivity has a form very similar to the exchange correction ( [ sigma ] ) , @xmath267 \nonumber \\ & \times & \int \frac{d^2{\bf q}}{(2\pi)^2 } \int{d \phi \over 2\pi } { d \phi ' \over 2\pi}\ \label{sigma - hartree } \\ & \times & \ \ { \rm im}\ \left[\ u_{\rm h}(\phi,\phi ' ) { \cal b}_{\alpha\beta}(\omega,{\bf q};\phi,\phi')\ \right ] , \nonumber\end{aligned}\ ] ] where u_h(,)=u(0,2k_f ) [ uh - def ] is the hartree interaction and @xmath268 is given by eqs .
( [ bwq ] ) , ( [ bwq_general ] ) without angular brackets ( denoting integration over @xmath95 and @xmath96 ) , see eq .
( [ 15 ] ) .
clearly , for a point - like interaction @xmath269 this yields ^h_=-2^f _ , [ sigma - hartree - point ] as has already been mentioned in sec .
[ iiib ] . in the case of the coulomb interaction
the situation is , however , more delicate @xcite . to analyze this case , it is convenient to split the interaction into the singlet and triplet parts @xcite . for the weak interaction , @xmath270 , the conductivity correction in the triplet channel
is then given by eq .
( [ sigma - hartree ] ) with an extra factor @xmath271 . as to the singlet part
, it is renormalized by mixing with the exchange term .
the effective interaction @xmath272 in the singlet channel is therefore determined by the equation u^s(, ) & = & u_0 - 12u_h(,)- + & & [ u_0 - 12u_h(,_1 ) ] p(_1,_2)u^s(_2, ) , + [ singlet - inter ] where @xmath273 is the bare coulomb interaction , and ( , ; _ 1,_2)= 2 + [ calp ] describes the electronic bubble . solving ( [ singlet - inter ] ) to the first order in @xmath274 , we get u^s(,;,)=u(,)- u^s_h(,;, ) , [ 1st - order - h ] where @xmath275 is the rpa - screened coulomb interaction ( [ screen ] ) which has already been considered in sec .
[ iiic ] , while the second term describes the renormalized hartree interaction in the singlet channel , u_h^s(, ) & = & 12u_h(,)- 12 + & & [ u_h(,_1)p(_1,_2 ) + p(_1,_2)u_h(_2, ) ] + & + & 12 ^ 2 + & & p(_1,_2)u_h(_2,_3)p(_3,_4 ) .
[ singlet - hartree ] here @xmath276 is the polarization operator ( [ polarization ] ) , and we have used the singular nature of the bare coulomb interaction implying @xmath277 for all relevant momenta . taking into account that the angular dependence of leading contributions to @xmath278 and @xmath279 is of the form @xmath280,$ ]
we find that the singlet hartree correction to @xmath31 is given by eq .
( [ sigma - hartree ] ) with a replacement u_h(, ) .
[ replacement ] note that in the diffusive limit @xmath281 is independent of @xmath282 , so that only the zero angular harmonic of the interaction contributes . on the other hand ,
the zero angular harmonic is suppressed in the effective singlet - channel interaction ( [ replacement ] ) .
therefore , the singlet channel does not contribute to the hartree correction in the diffusive limit , in agreement with refs . .
the situation changes , however , in the ballistic regime , when @xmath281 becomes angle - dependent .
after the angular integration , the triplet hartree conductivity correction takes the form ( [ sxxqw ] ) with the replacement @xmath283 and j_0 ^ 2(q ) -3y_0^ , [ repl - triph ] where @xmath284 .
for the singlet part we have a result similar to ( [ coulsxxqw ] ) with a slightly different @xmath249-integral , _ 0^q dq , where [ repl - singlh ] j(y , q ) = - y_0^. this yields for the total hartree contribution = ( _ c)^2k_fl [ g_h^s(t , y)+ 3g_h^t(t , y ) ] , [ rhohar ] where @xmath285 and @xmath286 governing the temperature dependence of the singlet and triplet contributions have the form g_h^s(x , y)&=&32 ^2 x^2 _ 0^dq q^3 j(y , q ) + & & _ n=1^ , + [ ghs ] + g_h^t(x , y)&=&x^2 y4 _ 0^ + & & _ 0^ d y+2 ( u-2 ^ 2 ) . + [ ght ] figure [ fig3]a shows @xmath287 as a function of @xmath288 for several values of @xmath289 .
the asymptotic behavior of @xmath290 is as follows : & = & ( _ c)^2k_fl + & & \ {
ll yy [ 34(t)+y ] , & t1 , + y^2[y(t)^1/2 ] , & 1 t1/y^2 , + c_0 ( t)^-1/2 , & t1/y^2 , . + [ rhohasympt ] we see that at @xmath291 a new energy scale @xmath292 arises where the mr changes sign .
specifically , at @xmath293 the mr , @xmath294 , is dominated by the exchange term and is therefore negative , while at @xmath295 the interaction becomes effectively point - like and the hartree term wins , @xmath296 , leading to a positive mr with the same @xmath297 temperature - dependence , see fig .
[ fig3]a .
, for ( a ) weak interaction , @xmath298 , and ( b ) strong interaction , @xmath299 ( from bottom to top ) .
dashed curves represent the exchange contribution.,title="fig:",width=302 ] , for ( a ) weak interaction , @xmath298 , and ( b ) strong interaction , @xmath299 ( from bottom to top ) .
dashed curves represent the exchange contribution.,title="fig:",width=302 ] in sec .
[ iiid ] we have assumed that @xmath300 or , in other words , the interaction parameter @xmath301 ( where @xmath302 is the static dielectric constant of the material ) is small .
this condition is , however , typically not met in experiments on semiconductor structures .
if @xmath303 is not small , the exchange contribution ( [ exchange ] ) remains unchanged , while the hartree term is subject to strong fermi - liquid renormalization @xcite and is determined by angular harmonics @xmath304 of the fermi - liquid interaction @xmath305 in the triplet ( @xmath306 ) and singlet ( @xmath307 ) channels .
the effective interaction @xmath308 replacing @xmath309 in ( [ sigma - hartree ] ) is then given by an equation of the type ( [ singlet - inter ] ) but with @xmath310 substituted for @xmath309 ( and without @xmath311 in the triplet channel ) , u^_eff(, ) & = & u_0+f^(-)2- + & & p(_1,_2)u^_eff(_2, ) , + [ strong - singlet ] + u^_eff(, ) & = & f^(-)2- + & & f^(-_1)2 p(_1,_2)u^_eff(_2, ) , [ strong - triplet ] a general solution of these equations requires inversion of integral operators with the kernels @xmath312 and @xmath313 and is of little use for practical purposes .
the situation simplifies , however , in both diffusive and ballistic limits . in the diffusive regime , @xmath314 , the second term in the polarization bubble
( [ calp ] ) and @xmath281 are independent of angles @xmath282 . as discussed in sec .
[ iiid ] , this leads to the suppression of the hartree contribution in the singlet channel , while in the triplet channel only the zero angular harmonic survives , u^_eff(,)= 1 2f_0^(dq^2-i ) ( 1+f_0^)dq^2-i + [ ueff - trip - diff ] we then reproduce the known result @xcite @xmath315 with g^t_h(t)=t .
[ diff - hartree ] in the ballistic limit , @xmath316 , the first term is dominant in ( [ calp ] ) , since @xmath317 is suppressed by a factor @xmath318 , according to ( [ < d > ] ) .
the angular harmonics then simply decouple in eqs .
( [ strong - singlet ] ) , ( [ strong - triplet ] ) , yielding effective hartree interaction constants @xmath319 @xmath320 and @xmath321 .
therefore , the hartree contribution reads [ fermi - hartree - ball ] g_h(t)=-c_0 2 . from a practical point of view
, it is rather inconvenient to describe the interaction by an infinite set of unknown parameters @xmath322 for this reason , one often assumes that the interaction is isotropic and thus characterized by two coupling constants @xmath323 and @xmath324 only . within this frequently used ( though parametrically uncontrolled ) approximation , the singlet part of the hartree term is completely suppressed .
the hartree contribution is then determined solely by the triplet channel with the effective interaction u^_eff(,)= 12f_0 ^ 1+f_0^ .
[ ueff - trip ] the hartree correction to the resistivity takes the form of eq .
( [ exchange ] ) with an additional overall factor of @xmath325 and with @xmath326 multiplied by @xmath327 , j_0 ^ 2(q)j_0 ^ 2(q ) f_0 ^ 1+f_0^ 1-j^(q ) , [ caljsigmaq ] everywhere in ( [ gfx ] ) ; the result is shown in fig . [ fig3]b for several values of @xmath328 .
until now we assumed that the temperature is much larger than the zeeman splitting @xmath329 , @xmath330 . in typical semiconductor structures
this condition is usually met in non - quantizing magnetic fields in the ballistic range of temperatures , allowing one to neglect the zeeman term .
if , however , this condition is violated , @xmath331 the zeeman splitting suppresses the triplet contributions with the @xmath332-projection of the total spin @xmath333 , while the triplet with @xmath334 and singlet parts remain unchanged . in the case of a weak interaction , @xmath335 , the triplet contribution @xmath336 in eq.([rhohar ] ) is modified in the following manner , 3 g_h^t(x , y ) g^t_h(x , y)+2re^t_h(x , y;_z ) , [ rhohar - weak - zee ] where @xmath337 , and the function @xmath338 describing the temperature dependence of the contribution with @xmath339 projection of the total spin is given by ^t_h(x , y;_z)&=&x^2 y4 _
0^ + & & _ 0^ d y+2 ( u-2 ^ 2 ) .
+ [ tildegth ] we see that at @xmath340 the contributions of @xmath339-components of the triplet saturate at the value given by ( [ ght ] ) with a replacement @xmath341 , i.e. at @xmath342 . in the opposite limit , @xmath343 , we have @xmath344 , and the result ( [ rhohar ] ) is restored . , eq .
( [ zeeman - strong ] ) , describing the temperature dependence of the triplet contribution is shown for @xmath345 and different values of zeeman splitting , @xmath346 ( from top to bottom ) .
dashed curve represents the case @xmath347.,width=302 ] the triplet contribution for strong isotropic interaction ( i.e. determined by @xmath348 only ) in the presence of zeeman splitting reads = -(_c)^2k_fl [ g_h^(t,0)+2re g_h^(t,_z ) ] .
+ [ tripzee ] the function @xmath349 is given by a formula similar to ( [ gfx ] ) , & & g_h^(x,_z ) = 32 ^2 x^2 _ 0^dq q^3 [ 1-j^(q ) ] + & & _ n=1^ , + [ zeeman - strong ] with @xmath350 as defined in ( [ caljsigmaq ] ) . again ,
for high temperatures @xmath351 all the triplet components contribute , so that the overall factor of 3 ( as in the absence of the zeeman splitting ) restores . on the other hand ,
for @xmath340 the contributions with @xmath339 projection of the spin saturate at low temperatures , and therefore the triplet contribution is partly suppressed , see fig .
[ zeeman ] . as discussed in sec .
[ iiib ] , calculation of the correction @xmath192 to the hall resistivity requires evaluation of both @xmath191 and @xmath255 .
in fact , as we show below , the temperature dependence of @xmath192 in a strong magnetic field is governed by @xmath191 in the diffusive limit and by @xmath255 in the ballistic limit .
since @xmath191 has been studied above , it remains to calculate @xmath255 . using the result ( [ bxy ] ) for the corresponding kernel @xmath235
, we get the exchange contribution for the case of a point - like interaction [ sigmaxy - point ] _ xy =- e^22 ^ 2 v_0_c , where the temperature dependence of the correction is governed by the function g_0^(xy)(x ) & = & -^2 _ 0^ ( -1/u ) + & & [ gxy ] + & & .
when writing ( [ gxy ] ) , we subtracted a temperature independent but ultraviolet - divergent ( i.e. determined by the upper limit in frequency integral ) contribution @xmath352 ; we will return to it in the end of this subsection . the function @xmath353 has the following asymptotics : [ g0xy ] g_0^(xy)(x)\ { ll 9 x , & x1 , + 11 c_1 x^1/2 , & x 1 , . with c_1=-4(1/2)0.647 . [ c1 ] combining ( [ deltaint ] ) and ( [ sigmaxy - point ] ) and using ( [ rhoxy - sigma ] )
, we find the correction to the hall resistivity , = g_0^_xy(t ) , [ deltarhoxy - point ] where g_0^_xy(x)&=&2g_0(x)-g_0^(xy)(x ) + & & \ { ll -2 x + const , & x1 , + - 11 c_1 x^1/2 , & x 1 . .
[ g0rhoxy ] the function @xmath354 is shown in fig .
[ grhoxy ] . as usual , the hartree term in the case of point - like interaction has an opposite sign and is twice larger in magnitude , if the zeeman splitting can be neglected .
an analogous consideration for the coulomb interaction yields a similar result for the exchange correction & = & , [ gfrhoxy - coul ] + g_f^_xy(x)&=&2g_f(x)-g_f^(xy)(x ) + & & \ { ll -2 x + const , & x1 , + - 11 2 c_1 x^1/2 , & x 1 . .
[ rhoxy - exchange ] the function @xmath355 is obtained by substituting ( [ u - rpa - b ] ) and ( [ bxy ] ) in ( [ sigma ] ) [ cf .
similar calculation for @xmath356 leading to eqs .
( [ coulsxxqw ] ) and ( [ gfx ] ) . ] the function @xmath357 describing the temperature dependence of the exchange correction to the hall resistivity is shown in fig .
[ grhoxy ] . in the ballistic regime ,
where @xmath355 dominates , the interaction becomes effectively point - like with @xmath358 , so that one can simplify the calculation using @xmath359 ( lower curve)and @xmath360 ( upper curve ) describing the temperature dependence of the hall resistivity for point - like and coulomb interaction , respectively .
diffusive ( @xmath361 ) and ballistic ( @xmath362 ) asymptotics , eqs .
( [ g0rhoxy ] ) and ( [ rhoxy - exchange ] ) , are also shown .
, width=302 ] to analyze the hartree contribution , we restrict ourselves to the isotropic - interaction approximation . then , similarly to the consideration in the end of sec .
[ iiie ] , only the triplet part contributes , and , in order to calculate @xmath363 , one should use eqs .
( [ bxy ] ) and ( [ ueff - trip ] ) . in the diffusive limit the hartree correction to the hall resistivity
is determined by ( [ diff - hartree ] ) , while in the ballistic limit we have again effectively point - like interaction with @xmath364 implying that @xmath365 this yields & = & - , [ rhoxy - hartree ] + g_h^_xy(x ) & & 3 \ { ll 2 x , & x1 , + c_1 x^1/2 , & x 1 . .
+ [ grhoxyh - asympt ] we return now to the @xmath2-independent contribution @xmath366 that was subtracted in eq .
( [ gxy ] ) . in view of the divergency of this term at @xmath367
, it is determined by the short - time cut - off @xmath368 g_uv^(xy)_u_min ~u_min^-1/2 .
[ guv ] since the correction we are discussing is governed by cyclotron returns , the cut - off @xmath369 corresponds to a single cyclotron revolution , @xmath370 .
[ on a more formal level , this is related to the assumption @xmath371 used for derivation of ( [ gxy ] ) ; see the text below eq .
( [ diffuson ] ) .
] we have , therefore , @xmath372 , with a constant @xmath373 of order unity@xcite . for the point - like interaction ,
the considered term produces a temperature - independent correction to the hall resistivity of the form = ( _ c ) ^1/2 .
[ rhouv ] in the case of coulomb interaction , this correction ( with both , exchange and hartree , terms included ) has the same form with @xmath374 $ ] . finally , let us discuss the expected experimental manifestation of the results of this subsection .
equations ( [ rhoxy - exchange ] ) , ( [ rhoxy - hartree ] ) predict that in the presence of interaction the temperature - dependent part of the hall resistivity @xmath375 in a strong magnetic field @xmath376 is linear in @xmath6 at arbitrary @xmath2 , with the @xmath2-dependence crossing over from @xmath377 in the diffusive regime to @xmath378 in the ballistic regime .
more specifically , if the interaction is not too strong , the exchange contribution ( [ rhoxy - exchange ] ) wins and the slope decreases with increasing temperature , while in the limit of strong interaction the slope increases due to the hartree term ( [ rhoxy - hartree ] ) . in an intermediate range of @xmath348 the slope is a non - monotonous function of temperature .
surprisingly , this behavior of the slope of the hall resistivity is similar to the behavior of @xmath31 obtained in ref . for @xmath193 and white - noise disorder .
this is a very non - trivial similarity , since the correction to @xmath379 at weak fields@xcite shows a completely different behavior , vanishing as @xmath380 in the ballistic regime .
in addition to the temperature - dependent linear - in-@xmath6 contribution , the interaction gives rise to a @xmath2-independent correction ( [ rhouv ] ) , which scales as @xmath381 ( assuming again that @xmath376 ) let us recall that these results are governed by multiple cyclotron returns and thus are valid under the assumption @xmath382 . in the opposite case , @xmath383 the correction is suppressed in the ballistic regime ( similarly to @xmath222 , see secs .
[ iib]3 and [ iiia ] ) , and the hall resistance takes its drude value .
it was argued in ref . by using the gutzwiller trace formula and hartree - fock approximation that the interaction correction to conductivity is related to a classical return probability .
the aim of this section is to demonstrate how this relation follows from the explicit formulas for @xmath31 .
we begin by considering the case of smooth disorder , when the kernel @xmath384 is given by eq .
( [ bwq ] ) . for simplicity
, we will further assume a point - like interaction , when only the first two terms in ( [ bwq ] ) give non - zero contributions .
we know that the result for the coulomb interaction is qualitatively the same [ cf . eqs .
( [ g0 ] ) and ( [ gf ] ) ] .
we will concentrate on the first term in ( [ bwq ] ) ; the second one yields a contribution of the same order in the ballistic regime and is negligible in the diffusive limit .
therefore , for the purpose of a qualitative discussion it is sufficient to consider the first term . using ( [ dd ] ) , the corresponding contribution can be estimated as & ~ & v_0_-^ d ( dq ) re + & ~&v_0_0^dt t ( r=0,t ) + & ~ & v_0 _ 0^t^-1dt ( r=0,t ) , [ qual - sigma ] where @xmath31 is the drude conductivity in magnetic field and we performed in the second line the fourier transformation of @xmath75 to the coordinate - time representation ( [ fourier ] ) . in a strong magnetic field and smooth disorder . in the ballistic regime ,
the peaks are separated by the cyclotron period , @xmath385 .
dashed curve represents the smoothened return probability @xmath386,width=302 ] the return probability in a strong magnetic field , @xmath53 , r(t)(r=0,t ) , [ return - prob ] is shown schematically in fig .
[ peaks ] . in the diffusive time range , @xmath387 , it is given by @xmath388 ( where @xmath77 is the diffusion constant in the magnetic field , @xmath389 ) .
equation ( [ qual - sigma ] ) thus yields in the diffusive regime , @xmath4 , ~v_0d(t ) , [ qual - diff ] in agreement with ( [ aadiff ] ) , ( [ gmdiff ] ) . at short ( ballistic ) time ,
@xmath210 , the return probability is governed by multiple cyclotron returns after @xmath390 revolutions , [ r(t ) ] r(t ) = _ n ( - ) . since @xmath391 , the conductivity correction ( [ qual - sigma ] ) is in fact determined by the smoothened return probability , ( t)=1 ( 2)^3/2 1 r_c^2 ( t)^1/2 .
[ rsmooth ] substituting ( [ rsmooth ] ) in ( [ qual - sigma ] ) we find that in the ballistic limit , @xmath194 , the conductivity correction scales as ~v_0d(t)^-1/2 , [ qual - ball ] in agreement with the exact results ( [ deltaint ] ) , ( [ g0 ] ) . as to the diffusive regime , @xmath4 , the contribution of short times @xmath392 to the integrand in ( [ qual - sigma ] )
yields a subleading @xmath2-independent correction @xmath393 to ( [ qual - diff ] ) .
it is worth emphasizing that the ballistic behavior ( [ rsmooth ] ) of the return probability @xmath386 corresponds to a _
one - dimensional _ diffusion .
consequently the _ ballistic _ result ( [ qual - ball ] ) has the same form as the _ diffusive _ altshuler - aronov correction in the quasi - one - dimensional geometry . to clarify the reason for emergence of the one - dimensional diffusion , we illustrate the dynamics of a particle subject to a strong magnetic field and smooth disorder in fig .
[ qual ] let us assume that the velocity is in @xmath394 direction at @xmath395 . as is clear from fig .
[ qual ] , the return probability @xmath396 after the first cyclotron revolution ( the integral of the first peak in fig . [ peaks ] )
is determined by the shift @xmath397 of the guiding center in the cyclotron period @xmath385 . in view of the diffusive dynamics of the guiding center , this shift has a gaussian distribution with _ 1 ^ 2_x^2 = 2dt_1=2r_c^2 _ c , [ x - shift ] yielding r_1=(_c)^1/22v_f r_c .
[ r1 ] furthermore , we have @xmath398 after @xmath399 revolutions , yielding the return probability @xmath400 . as to the @xmath394-component @xmath401 of the guiding center shift , it only governs the width of the peaks in eq .
( [ r(t ) ] ) and fig . [ peaks ] without affecting @xmath386 .
therefore , the smoothened return probability is [ r - sum - over - n ] |r(t)= .
r_n t_1|_n = t / t_1 , which reproduces eq .
( [ rsmooth ] ) . as mentioned in sec .
[ iiia ] , the emergence of the one - dimensional diffusion in the ballistic regime is reflected by the factor @xmath402 in the formula ( [ bxx ] ) for the kernel @xmath384 .
this factor effectively reduces the dimensionality of the @xmath147-integral , @xmath403 in the above we considered a system with smooth disorder , for which @xmath191 at @xmath193 vanishes exponentially in the ballistic limit .
now we turn to the opposite case of a white - noise disorder .
we will show that the linear - in-@xmath2 correction @xcite ( sec .
[ iic]3 ) is again related to the return probability but the relation is different from ( [ qual - sigma ] ) . indeed , according to ( [ wnhik ] ) , we have now the structure @xmath404 instead of @xmath405 that was relevant for smooth disorder .
on the other hand , the return probability at ballistic times @xmath210 is clearly dominated by processes with a single back - scattering event , implying ( r=0,t ) ~1(dq)d_f(,)^2 e^it .
[ wnrt ] therefore , the contribution of the first term in ( [ wnhik ] ) can be cast in the form & ~ & v_0 _
e_f^-1^t^-1dt t ( r=0,t ) , + & ~ & v_0 .
[ wn - qual - sigma ] it is easy to see that the probability of a ballistic return after a single scattering event is [ wn - ret - prob ] ( r=0,t)~1 d^2 r_1 ( t-2 r_1/v_f ) ( v_f r_1)^2 ~1 v_f^2 t. substituting ( [ wn - ret - prob ] ) in ( [ wn - qual - sigma ] ) , we reproduce the linear - in-@xmath2 correction ( [ sigma_ball ] ) , _
xx(t ) ~e^2 v_0 t[qual - lin - t ] the constant term in ( [ wn - qual - sigma ] ) comes from the lower limit of the time integral , which is of the order of @xmath406 .
this constant merely renormalizes the bare value of the drude conductivity . on the diffusive time scale @xmath407 so that there is no difference between white - noise and smooth disorder .
therefore , in the diffusive limit the result ( [ qual - sigma ] ) applies , yielding the usual logarithmic correction ( [ qual - diff ] ) .
in fact the contribution of the type ( [ qual - sigma ] ) arises also in the ballistic regime when all terms in ( [ bwq_general ] ) are taken into account . according to ( [ wn - ret - prob ] ) , it has the form ~ v_0d [ ( t)-const ] , which is a subleading correction to the linear - in-@xmath2 term ( [ sigma_ball ] ) , ( [ qual - lin - t ] ) . in the ballistic regime , @xmath194 ,
the above qualitative arguments for a white - noise disorder can be re - formulated in terms of the interaction - induced renormalization of the differential scattering cross - section on a single impurity .
specifically , the renormalization occurs due to the interference of two waves , one scattered off the impurity and another scattered off the friedel oscillations created by the impurity@xcite .
the interference contribution is proportional to the probability @xmath408 of backscattering off the impurity ( see appendix [ a3 ] ) and hence , to the return probability after a single - scattering event , as discussed above . on the other hand
, this implies that the scattering cross - section around @xmath409 is itself modified by the friedel oscillations ( in other words , the impurities are seen by electrons as composite scatterers with an anisotropic cross - section ) .
the renormalization of the bare impurity depends on the energy of the scattered waves , which after the thermal averaging translates into the @xmath2-dependence of the effective transport scattering time , @xcite @xmath410 [ this corresponds to setting @xmath221 in the return probability , see eq .
( [ wn - qual - sigma ] ) ] .
this mechanism provides a systematic microscopic justification of the concept of temperature - dependent screening @xcite .
we recall that , in addition to the linear - in-@xmath2 term , the conductivity correction contains a @xmath2-independent contribution determined by the ultraviolet frequency cut - off @xmath411 . in the case of strong interaction this term
can be of the same order as the bare ( non - interacting ) drude conductivity .
the coefficient in front of this term can not be calculated within the quasiclassical approach because it is governed by short - distance physics at scales of the order of @xmath412 . at the same time , according to the above picture , this @xmath2-independent correction also modifies the impurity scattering cross - section around @xmath413 .
the corresponding correction @xmath414 may thus be comparable to the bare isotropic scattering probability @xmath415 .
an interesting consequence of this fact is a possible situation when the total relaxation rate @xmath416 $ ] is _ smaller _ than the transport relaxation rate @xmath417 ( 1-\cos\phi)$ ] . in smooth disorder ( small - angle scattering ) , the backscattering amplitude vanishes exponentially with @xmath418 , and so does the amplitude of friedel oscillations .
this leads to the suppression of the @xmath1-contribution to the conductivity [ see sec .
[ iic]3 ; this fact was realized within the @xmath2-dependent screening picture already in ref . for the case of scattering on long - range interface roughness ] .
we note , however , that the understanding of the interaction effects in terms of scattering off friedel oscillations is only possible in the ballistic regime . indeed , the diffusive correction in a _ smooth _ random potential is _ not _ exponentially small and is related to random ( having no @xmath419-oscillating structure ) fluctuations of the electron density , as was pointed out in refs . .
the correlations of these fluctuations ( which reduce to the friedel oscillations on the ballistic scales ) are described by the return probability at arbitrary scales .
finally , we use the interpretation of the interaction correction in terms of return probability to estimate the mr in the white - noise random potential and at sufficiently weak magnetic fields , @xmath38 .
note that the zero-@xmath6 ballistic correction ( [ qual - lin - t ] ) does not imply any dependence of resistivity on magnetic field .
indeed , as follows from ( [ drude ] ) , a temperature dependence of the transport time @xmath410 is not sufficient to induce any non - trivial mr , @xmath420 if @xmath3 is @xmath6-independent . in order to obtain the @xmath6-dependence of the resistivity
, we thus have to consider the influence of the magnetic field on the return probability determining the correction to the transport time . since in the ballistic regime the characteristic length of relevant trajectories is @xmath421 , their bending by the magnetic field modifies only slightly the return probability for @xmath38 .
the relative correction to the return probability is thus of the order of @xmath422 independently of the relation between @xmath86 and @xmath47 .
therefore , to estimate the mr in the white - noise potential for @xmath423 , one can simply multiply the result ( [ qual - lin - t ] ) for @xmath193 by a factor @xmath424 , yielding ~(_c)^2 k_f l1 , _ ct .
[ rho - wn - qual ] a formal derivation of this result is presented in sec .
[ vb ] . in a stronger magnetic field , @xmath382 ,
the situation changes dramatically due to multiple cyclotron returns , see above .
this regime is considered in sec .
[ va ] below .
in sec . [ iii ] , we studied the interaction correction for a system with a small - angle scattering induced by smooth disorder with correlation length @xmath205 .
this is a typical situation for high - mobility gaas structures with sufficiently large spacer @xmath45 .
it is known , however , that with further increasing width of the spacer the large - angle scattering on residual impurities and interface roughness becomes important and limits the mobility .
furthermore , in si - based structures the transport relaxation rate is usually governed by scattering on short - range impurities .
this motivates us to analyze the situation when resistivity is predominantly due to large - angle scattering .
we thus consider the following two - component model of disorder ( `` mixed disorder '' ) : white - noise random potential with a mean free time @xmath425 and a smooth random potential with a transport relaxation time @xmath213 and a single particle relaxation time @xmath426 [ @xmath427 .
we will further assume that while the transport relaxation rate @xmath428 is governed by short - range disorder , @xmath429 , the damping of sdho is dominated by smooth random potential , @xmath430 this allows us to consider the range of classically strong magnetic fields , @xmath431 neglecting at the same time landau quantization ( which is justified provided @xmath432 ) . to calculate the interaction corrections
, we have to find the corresponding kernel @xmath144 determined by the classical dynamics .
naively , one could think that under the assumed condition @xmath429 the smooth disorder can simply be neglected in the expression for the classical propagator . while this is true in diffusive limit ,
the situation in the ballistic regime is much more nontrivial . to demonstrate the problem ,
let us consider the kernel @xmath433 in the ballistic limit @xmath434 and in a strong magnetic field @xmath435 .
if the smooth random potential is completely neglected in classical propagators , we have [ see appendix [ a2 ] ; the second term in eq .
( [ brhoxx - wn - ball ] ) can be neglected for @xmath382 ] b_xx^ ( ) , [ brho - ball - wn ] where @xmath437 is the angle - averaged propagator with only out - scattering processes included , g_0(,)=i_c , [ g0-wn ] and @xmath438 . if characteristic frequencies satisfy @xmath371 ( which is the case for @xmath391 ) , eq .
( [ g0-wn ] ) can be further simplified , g_0=. [ g0-wn-1 ] substituting ( [ brho - ball - wn ] ) and ( [ g0-wn-1 ] ) in ( [ rho ] ) , we see that momentum- and frequency - integrations decouple and that the first term in ( [ brho - ball - wn ] ) generates a strongly ultraviolet - divergent @xmath147-integral @xmath439 .
the physical meaning of this divergency is quite transparent .
the contribution of the first term in ( [ brho - ball - wn ] ) to @xmath222 is proportional to the time - integrated return probability @xmath440 , similarly to ( [ qual - sigma ] ) . for @xmath441
the propagator @xmath442 describes the ballistic motion in the absence of scattering , which is merely the undisturbed cyclotron rotation in the case of a strong magnetic field . since at @xmath443
the particle returns exactly to the original point , the integral @xmath440 diverges .
the encountered divergency signals that the neglect of smooth disorder is not justified , even though @xmath429 . indeed , with smooth disorder taken into account , the particle does not return exactly to the original point after a cyclotron revolution , see sec .
the return probability is then described by eqs .
( [ r(t ) ] ) , ( [ rsmooth ] ) with @xmath3 replaced by @xmath213 .
it is worth mentioning a similarity with the problem of memory effects in a system with strong scatterers , where even a weak smooth disorder turns out to be crucially important @xcite . to demonstrate the role of the smooth disorder on a more formal level
, we write down the angle - averaged propagator in the ballistic regime , @xmath444 , for the mixed - disorder model , ( , ) = .
[ d - mixed ] clearly , in both limits @xmath445 and @xmath446 this formula reduces to ( [ g0-wn-1 ] ) and ( [ < d > ] ) , respectively . in view of @xmath447
the last term in the denominator of ( [ d - mixed ] ) can be neglected , and we return to the expression for solely smooth disorder .
the presence of the term @xmath448 regularizes the @xmath249-integrals , thus solving the problem of ultraviolet - divergencies discussed above .
the characteristic momenta are thus determined by @xmath449 . therefore , despite the weakness of the smooth disorder , @xmath450 it is the first ( @xmath249-dependent ) rather than the third term which has to be retained in the denominator of ( [ d - mixed ] ) . in other words , in the ballistic regime and in a strong magnetic field the dynamics in the considered model
is governed by smooth disorder .
the above discussion demonstrates that at @xmath451 the kernel @xmath144 for the mixed - disorder model is given by ( [ bwq_general ] ) with propagators @xmath75 calculated in smooth random potential ( i.e. with white - noise disorder neglected ) . the time @xmath425 enters the result only through the matrices @xmath135 ( determined by the transport time @xmath452 ) and @xmath453 . using @xmath454
we find then that the resulting expression , b_xx & & 12_c^2 + & -&12_c^2_wn[-2 n_yn_y d ] + & -&12_c^2[-2 n_y d n_y ] + & + & 2_cn_ydn_xd- n_xdn_xd[bxxtotal ] is dominated by the first term corresponding to the first term in eq .
( [ brho - ball - wn ] ) .
this yields for @xmath455 b_xx ( , ) , [ bxx - mixed ] where @xmath456 . as in previous sections ,
we first calculate the conductivity correction for a point - like interaction . substituting ( [ bxx - mixed ] ) in ( [ sigma ] ) , we get _ xx =- e^2 2 ^ 2 v_0 ( _ sm)^1/24c_0(t)^1/2 , t1/_wn , [ sigma - mixed - ball ] with the constant @xmath457 as defined in eq .
( [ c0 ] ) . for an arbitrary ( not necessarily small ) value of the ratio @xmath458 the coefficient @xmath459 in ( [ sigma - mixed - ball ] )
is replaced by @xmath460 . for @xmath461 ( i.e. without white - noise disorder )
we then recover the ballistic asymptotics of eq .
( [ g0 ] ) . as in the case of purely smooth disorder
, the resistivity correction @xmath222 is related to @xmath191 via eq .
( [ parabola ] ) . comparing ( [ sigma - mixed - ball ] ) with ( [ g0 ] )
, we see that the correction @xmath222 is enhanced in the mixed - disorder model by a factor @xmath462 as compared to the purely smooth - disorder case . on the other hand ,
the scaling with temperature and magnetic field , @xmath463 , remains the same .
let us analyze now the crossover from the ballistic to the diffusive regime .
setting @xmath464 in ( [ sigma - mixed - ball ] ) , we find that the correction is parametrically large , @xmath465 . clearly , this does not match the diffusive contribution ( [ gmdiff ] ) , yielding @xmath466 at @xmath464 .
this indicates that returns without scattering on white - noise disorder continue to govern the correction in certain temperature window below @xmath467 , which normally belongs to the diffusive regime . to find the corresponding contribution , one should take into account the scattering - out term @xmath468 in the denominator of ( [ d - mixed ] ) , which yields g_1(x,)&=&2
( 2x ) ^1/2 _ 0^ + [ g1 ] & = & \ { ll ( 2)^1/2 , & x1 , + 4 c_0 ^1/2 x^-1/2 , & x 1 , .
where @xmath469 and @xmath470 .
to describe the temperature dependence of the interaction correction for all @xmath2 , we have to add here the diffusive contribution , which has the form ( [ gmdiff ] ) for @xmath4 and vanishes for @xmath194 .
this contribution corresponds to long times @xmath387 and describes the trajectories multiply scattered off white - noise disorder . since at @xmath464
the sum of the ballistic and diffusive contributions will be dominated by @xmath471 , the precise way of vanishing of the diffusive contribution at @xmath464 is inessential .
therefore , we can describe it by the function @xmath259 , eq .
( [ deltaint ] ) .
the resistivity correction for a system with mixed disorder and point - like interaction has thus the following form : = -(_c)^2k_fl v_0 g^mix_0(t,_sm/ ) , [ rho - mix ] where g^mix_0(x , ) & = & g_1(x , ) + g_0(x ) [ g - mix ] + & = & \ { ll -x + ( 2)^1/2 , & x1 , + 4 c_0 ^1/2 x^-1/2 , & x 1 , .
this result is illustrated in fig .
[ mix]a .
( a ) and @xmath472 ( b ) describing the temperature dependence of the resistivity correction due to point - like and coulomb ( exchange ) interaction , respectively , in the mixed - disorder model for different values of parameter @xmath473 ( from top to bottom ) .
dashed curves represent these functions for purely smooth disorder ( @xmath474 ) .
, title="fig:",width=302 ] ( a ) and @xmath472 ( b ) describing the temperature dependence of the resistivity correction due to point - like and coulomb ( exchange ) interaction , respectively , in the mixed - disorder model for different values of parameter @xmath473 ( from top to bottom ) .
dashed curves represent these functions for purely smooth disorder ( @xmath474 ) . ,
title="fig:",width=302 ] in the case of coulomb interaction , we have as usual a similar result for the exchange contribution = -(_c)^2k_fl g_f^mix(t,_sm/ ) , [ rho - mix - f ] with g_f^mix(x , ) & = & 1 2 g_1(x , ) + g_f(x ) [ g - mix - f ] + & = & \ { ll -x + ( /2)^1/2 , & x1 , + 2 c_0 ^1/2 x^-1/2 , & x 1 . .
this function is shown in in fig .
in fact , here the diffusive contribution can be described either by the function @xmath264 or by @xmath259 because in the diffusive limit they coincide up to a small constant .
since in the intermediate and ballistic regimes [ where @xmath264 and @xmath259 differ ] the contribution @xmath475 is dominant , the behavior of the diffusive contribution is of no importance , as in the case of point - like interaction . note that the ballistic contribution corresponds to the point - like interaction with @xmath476 , yielding a factor @xmath477 in front of @xmath478 as compared to ( [ g - mix ] ) .
this is because the dynamical part of screening is suppressed for all relevant @xmath479 in the whole range of temperatures , even for @xmath480 , where this contribution is important .
this also applies to the hartree contribution to the resistivity . within the `` @xmath348-approximation '' we have again an effectively point - like interaction with @xmath481 in the ballistic term
the result thus reads = 3(_c)^2k_fl g_h^mix(t,_sm/ ) , [ rho - mix - h ] where & & g_h^mix(x , ) = 12 f_0^(1+f_0^ ) g_1(x , ) + g^t_h(x ) + & & = \ { ll x + f_0 ^ 1+f_0^(2)^1/2 , & + & x1 , + - 2 c_0f_0^(1+f_0^)^1/2x^-1/2 , & x 1 . .
+ [ g - mix - h ] before closing this subsection , we briefly discuss the hall resistivity in the mixed disorder model . repeating the steps described above , we find that the ballistic contribution to @xmath379 also contains an extra factor @xmath482 , similarly to @xmath10 . for an arbitrary ( not necessarily small ) value of the ratio @xmath458 the coefficient @xmath483 in eqs .
( [ g0rhoxy ] ) , ( [ rhoxy - exchange ] ) is replaced by @xmath484(\tau_{\rm sm}/\tau)^{1/2}$ ] . in the case of a purely smooth disorder ( sec .
[ iii ] ) the resistivity correction in the ballistic regime is exponentially suppressed for @xmath38 because the particle can not return to the origin . when the short - range potential is present , the situation changes and the return probability is determined for @xmath194 by the single - backscattering processes .
the interaction - induced mr arises then due to the influence of the magnetic field on the probability of such return , as discussed in the end of sec .
[ iv ] . in this case
, there is no need to take the smooth potential into account and the mr is determined solely by the white - noise disorder .
let us calculate the corresponding correction using the ballistic form ( [ deltabrhoxx - ball ] ) of the kernel @xmath485 .
for a point - like interaction , substituting ( [ deltabrhoxx - ball ] ) in ( [ rho ] ) , we find the following ballistic ( @xmath194 ) asymptotics of the longitudinal mr , = - .
[ drhoxx - weak - wn - point ] in the case of coulomb interaction , @xmath485 is multiplied by the ballistic asymptotics of the interaction , eq .
( [ u - ball ] ) . substituting this product in eq .
( [ rho ] ) , we get the fock contribution to the mr in the form = - , t1 .
[ drhoxx - weak - wn - coul ] the corresponding hartree term also scales as @xmath486 .
it is worth noting that there is another contribution to the mr in this regime , which comes from the suppression of the triplet channel due to zeeman splitting @xmath329 rather than from the orbital effects .
this contribution is identical to that found in ref . for the ballistic magnetoresistance in a parallel magnetic field .
it also scales as @xmath486 in a weak magnetic field ; however , it contains an extra factor @xmath487 , as compared to ( [ drhoxx - weak - wn - coul ] ) .
this factor is small in typical experiments on semiconductor heterostructures where the effective mass of the carriers is much smaller that the bare electron mass .
we now turn to the hall resistivity . using ( [ bxy - zna - ball ] ) and
( [ deltabrhoxy ] ) , we find for @xmath423 and for arbitrary relation between @xmath86 and @xmath47 = [ drhoxy - weak - wn - point ] for the point - like interaction , and = [ drhoxy - weak - wn - coul ] for the coulomb interaction .
the result ( [ drhoxy - weak - wn - coul ] ) reduces in the limit @xmath167 to that obtained in ref . from the quantum kinetic equation .
we see that in view of a relatively small value of the numerical coefficient @xmath488 , the first ( @xmath6-independent ) term in square brackets in ( [ drhoxy - weak - wn - coul ] ) dominates for @xmath489 , so that the results of ref . are applicable in sufficiently broad range of magnetic fields .
for the corresponding hartree - correction to @xmath192 calculated within the `` @xmath348-approximation '' , we refer the reader to ref . .
in the preceding consideration , we assumed that the 2d system is isotropic .
while this is true for the majority of magnetotransport experiments we have in mind , there exists a number of important situations when the transport is anisotropic , @xmath490 .
first , such an anisotropy can be induced by the orientation of the 2d electron gas plane with respect to the crystal axes , see e.g. ref . for a measurement of the quantum correction for the ( 110 ) surface of the si - mosfet .
second , the electron - electron interaction may lead to spontaneous formation of a charge - density wave .
finally , the anisotropy may be induced by a one - dimensional periodic modulation ( lateral superlattice ) .
the latter example is of special interest in view of emergence of commensurability oscillations ( known as weiss oscillations)@xcite , and we will discuss it in more detail in sec .
[ vic ] .
the interaction - induced correction to the conductivity tensor of an anisotropic system was calculated for the diffusive regime and @xmath193 by bhatt , wlfle , and ramakrishnan @xcite .
they showed , in particular , that the quantum correction preserves the anisotropy of the quasiclassical ( boltzmann ) conductivity .
below we will generalize their result onto the case of a classically strong magnetic field , and , furthermore , will extend the consideration to the ballistic regime .
we begin by presenting a simple argument allowing one to estimate the conductivity correction in an anisotropic system ; we will confirm it by a formal calculation in sec .
[ vib ] . according to eq .
( [ qual - sigma ] ) , the relative correction to a diagonal component @xmath491 of the conductivity tensor is determined by the return probability ( and is , thus , the same for @xmath492 and @xmath493 ) .
this implies , in the diffusive regime [ sigma - mumu ] ~-re . ( dq )
|^=1/_=t , yielding _
xx~ e^2(_xx_yy)^1/2t[anis - diff - qual ] and analogously for @xmath494 . in the ballistic regime
the time - integrated return probability @xmath495 scales as @xmath297 [ see eqs .
( [ rsmooth ] ) and ( [ qual - ball ] ) ] , so that we have instead of ( [ anis - diff - qual ] ) , _ xx~ e^2(_xx_yy ) ( t)^-1/2 .
[ anis - ball - qual ] the explicit form of the function @xmath496 will be calculated below .
since the conductivity corrections ( [ anis - diff - qual ] ) and ( [ anis - ball - qual ] ) are only determined by the anisotropic diffusion , we expect that they do not depend on the particular source of anisotropy , in analogy with ref . .
an important feature of the results ( [ anis - diff - qual ] ) and ( [ anis - ball - qual ] ) is that they mix the components @xmath31 and @xmath497 of the conductivity tensor .
this will play a central role in our analysis of the interaction effect on the magnetoresistivity of modulated systems in sec .
[ vic ] .
it is worth mentioning that the validity of the formula ( [ anis - ball - qual ] ) for the ballistic regime is restricted on the high - temperature side by the condition @xmath498 , where @xmath499 is the time scale on which the anisotropic diffusion of the guiding center sets in .
the value of @xmath500 depends on the particular microscopic mechanism of anisotropy .
we will estimate @xmath500 and the behavior of the conductivity correction at @xmath501 for a modulated system in sec .
[ vic ] .
we proceed now with a formal calculation of the quantum correction to the conductivity of an anisotropic system in a strong magnetic field . as a model of anisotropy
, we will assume anisotropic impurity scattering , with a cross - section @xmath502 .
repeating the derivation performed in secs .
[ iia ] and [ iib ] , we find that the result ( [ sigma ] ) , ( [ bwq ] ) remains valid in the anisotropic case , with the matrix @xmath135 proportional to the corresponding ( anisotropic ) diffusion tensor @xmath503 , [ t_ab - anis ] t_= 2d_v_f^2 = 11+_c^2_x_y ( cc _ x & -_c _
x_y + _ c _ x_y & _ y ) , where @xmath504 and @xmath505 are the relaxation times for the corresponding components of the momentum .
we begin by considering the diffusive limit , when the leading contribution comes from three - diffuson diagrams , fig .
[ fig1]_d _ and _ e _ ( see sec . [ iic]1 ) , which are represented by the last term in eq .
( [ bwq ] ) . the singular contribution to the propagator @xmath75 , governed by the diffusion mode ,
has a form analogous to ( [ dmode ] ) , ^s(,;, ) , [ dmode - anis ] see appendix [ a5 ] for the derivation of ( [ dmode - anis ] ) and explicit expressions of @xmath506 . using ( [ dmode - anis ] ) and ( [ psirl - anis ] ) , we get ^s= 1 d _ q_q_-i [ < d>-anis - diff ] and b_xx ( , ) & & - ^sn_^s n_^s + & = & 4 v_f^2 .
[ bxx - anis - diff ] the result ( [ bxx - anis - diff ] ) can also be obtained with making use of the identity ( [ dndnd ] ) ; then it is sufficient to keep only the leading term ( unity ) in the expressions for functions @xmath506 entering ( [ dmode - anis ] ) . substituting ( [ bxx - anis - diff ] ) , ( [ < d>-anis - diff ] ) , ( [ screen ] ) in ( [ sigma ] ) , we obtain the final result for the conductivity correction in the diffusive regime , _
xx&= & e^22 ^ 2(d_xxd_yy)^1/2 t , [ anis - diff - xx ] + _ yy&= & e^22 ^ 2(d_yyd_xx)^1/2 t , [ anis - diff - yy ] in full agreement with a qualitative consideration of sec . [ via ] [ eq .
( [ anis - diff - qual ] ) ] .
the correction to the hall conductivity is zero in the leading ( @xmath25 ) order , as in the isotropic case .
for the point - like interaction , the result remains the same , up to a factor @xmath188 .
we now extend the consideration beyond the diffusive limit ( thus allowing for @xmath507 ) , assuming first the smooth disorder and concentrating on longitudinal components of the conductivity and resistivity tensors . in analogy with ( [ diffuson ] )
, the singular contribution @xmath171 to the propagator acquires then the form ( see appendix [ a5 ] ) ( , ; , ) & = & \{-iqr_c[(-_q)-(-_q ) ] } + & & [ anis - ball - diffuson ] where ( ) = 1-iq v_f _ c^2 ( 1 _ x_q+ 1 _ y_q ) .
[ chi - anis ] this yields = j_0 ^ 2(qr_c ) d_xxq^2_x+d_yyq^2_y - i [ < d>-anis - ball ] and b_xx ( , ) = 4 v_f^2 , [ bxx - anis - ball ] which differs from ( [ < d>-anis - diff ] ) , ( [ bxx - anis - diff ] ) by the factor @xmath508 only . in the ballistic limit @xmath509 the relevant values of @xmath510 are large , @xmath511 , so that the screening is effectively static and the interaction is effectively point - like with @xmath512 . substituting then ( [ bxx - anis - ball ] ) in ( [ sigma ] ) and rescaling the integration variables @xmath513 , we find _ yy&= & -e^24 ^ 2c_0(t_y)^-1/2 + & & 2*k * ( ) , [ anis - ball - yy ] + _ xx&= & d_yyd_xx _ yy , [ anis - ball - x ] where @xmath514 is the elliptic integral , _ 0^/2= * k * ( ) , 0<q<1 , [ k - elliptic ] and we assumed that @xmath394 is the easy - diffusion axis , @xmath515 .
let us analyze the obtained results in the limits of weak and strong anisotropy .
it is convenient to set @xmath516 , @xmath517 , @xmath518 , and to introduce a dimensionless anisotropy parameter @xmath519 . using the asymptotics of the elliptic integral , ( s)\ { ll ( 1+s^2 4 ) , & s1 , + , & 1-s 1 , .
[ asympt - elliptic - k ] we find _ xx & & -e^24 ^ 2c_0(t_y)^-1/2 + & & \ { ll 1 - 4 , & 1 , + , & 1 , .[limits - anis - ball ] and @xmath520 equations ( [ anis - ball - yy]),([anis - ball - x ] ) and ( [ limits - anis - ball ] ) confirm the qualitative arguments in sec . [ via ] ( based on the behavior of the return probability ) which led to eq .
( [ anis - ball - qual ] ) . in this subsection
, we apply the results of sec .
[ vib ] to a particularly important class of anisotropic 2d systems , namely , 2d electron gas subject to a periodic potential varying in one direction . such systems ( lateral superlattices ) have been intensively investigated experimentally during the last fifteen years . in a pioneering work@xcite , weiss ,
von klitzing , ploog and weimann discovered that even a weak one - dimensional periodic modulation with a wave vector @xmath521 may induce strong oscillations of the magnetoresistivity @xmath9 [ while showing almost no effect on @xmath522 and @xmath375 ] , with the minima satisfying the condition @xmath524 . here
@xmath390 and @xmath525 is the modulation period .
the quasiclassical nature of these commensurability oscillations was demonstrated by beenakker@xcite , who showed that the interplay of the cyclotron motion and the superlattice potential induces a drift of the guiding center along @xmath394 axis , with an amplitude squared oscillating as @xmath526 [ this is also reproduced by a quantum - mechanical calculation , see refs . ] . while ref
. assumed isotropic impurity scattering ( white - noise disorder ) , it was shown later that the character of impurity scattering affects crucially the dependence of the oscillation amplitude on the magnetic field .
the theory of commensurability oscillations in the situation of smooth disorder characteristic for high - mobility 2d electron gas was worked out in ref .
( see also numerical solution of the boltzmann equation in ref . ) and provided a quantitative description of the experimentally observed oscillatory magnetoresistivity @xmath527 .
the result has the form@xcite [ weiss - smooth ] _
xx_0= j_i(kr_c)j_-i(kr_c ) , where @xmath528 is the dimensionless amplitude of the modulation potential [ @xmath529 , and = 1 _ c_s \{1-^-1/2}. [ mu - weiss ] in the range of sufficiently strong magnetic fields eq . (
[ weiss - smooth ] ) describes the commensurability oscillations with an amplitude proportional to @xmath530 , [ weiss - cos - smooth ] _
xx_0 ^ 2 ( _ c ) ^2 k r_c^2(kr_c-/4 ) , for precise conditions of validity of ( [ weiss - cos - smooth ] ) , as well as for an analysis of the result ( [ weiss - smooth ] ) in the whole range of magnetic fields , the reader is referred to ref . . as to the modulation - induced corrections @xmath531 to the other components of the resistivity tensor , they are exactly zero within the quasiclassical ( boltzmann equation ) approach , independently of the form of the impurity collision integral@xcite .
such corrections appear in the quantum - mechanical treatment of the problem @xcite and are related to the de haas - van alphen oscillations of the density of states induced by the landau quantization of spectrum . as a consequence ,
these oscillations are exponentially damped by disorder , with the damping factor @xmath532 $ ] .
the phase of such quantum oscillations @xmath533 is opposite to that of quasiclassical commensurability oscillations in @xmath534 , eqs .
( [ weiss - smooth ] ) , ( [ weiss - cos - smooth ] ) .
indeed , oscillations in @xmath535 that are much weaker than those in @xmath534 , have the opposite phase , and vanish much faster with decreasing @xmath6 , were observed in ref . .
we will neglect these oscillations , which are exponentially weak in the range of magnetic fields considered in the present paper , @xmath536 .
we are going to show that the interaction - induced correction to resistivity also generates oscillations in @xmath537 , which are , however , unrelated to the dos oscillations of a non - interacting system and become dominant with lowering temperature . to demonstrate this , we apply the result of sec .
[ vib ] for the interaction - induced correction in an anisotropic system .
the anisotropy parameter is governed by the quasiclassical correction to @xmath10 due to modulation , = _
yy_xx-1 -1=_xx_0 [ beta - weiss ] and is given by eq .
( [ weiss - smooth ] ) .
for simplicity we will assume that the effect of modulation is relatively weak , @xmath538 .
( generalization to the large-@xmath539 case with making use of the corresponding formulas of sec .
[ vib ] is completely straightforward . ) using ( [ anis - diff - xx ] ) and ( [ anis - ball - x ] ) , we find the oscillatory correction to @xmath537 as a combined effect of the modulation and the coulomb interaction , & = & + & & \ { ll -t , & t1 , + ( t)^-1/2 , & t1 .
.[interact - weiss ] in the presence of strong scatterers ( mixed disorder model ) , the result for the ballistic regime is enhanced by the factor @xmath540 , as discussed in sec .
let us remind the reader that the result ( [ interact - weiss ] ) is valid for temperatures @xmath541 , where @xmath542 is the characteristic time on which the motion of the guiding center takes the form of anisotropic diffusion ( see sec . [ via ] ) . for the case of a modulated system with a smooth random potential
we find @xmath543 .
this is because on a scale shorter than @xmath499 the modulation leads to a drift of the guiding center along @xmath394 axis with the velocity depending on the coordinate @xmath544 of the guiding center @xcite , v_d(x)&=&-v_f2kr_c j_0(kr_c ) ( kx ) [ vdrift - mod ] + & & -(kr_c-/4)(kx ) . in a time
@xmath545 the position @xmath544 of the guiding center is shifted by a distance of the order of the modulation period @xmath546 due to the small - angle impurity scattering .
therefore , the drift velocity @xmath547 typically changes sign on this time scale , so that the drift is transformed to an additional diffusion process , with @xmath548 , in agreement with ( [ weiss - cos - smooth ] ) . to estimate the resistivity correction @xmath549 in the ultra - ballistic regime @xmath501
, we use the relation between the conductivity correction and the return probability ( sec .
the return probability @xmath550 after @xmath399 revolutions ( introduced in sec .
[ iv ] ) is modified by the modulation - induced drift in the following way : r_n^mod = r_n ( 1-n_c ) .
[ rn - mod ] according to ( [ qual - sigma ] ) , this yields an oscillatory correction to resistivity suppressed by a factor @xmath551 as compared to the second line ( ballistic regime ) of eq .
( [ interact - weiss ] ) . .
( a ) quasiclassical weiss - oscillations ; dashed line shows the resistivity in the absence of modulation .
( b ) interaction - induced quantum oscillations in @xmath537 for three temperatures .
the curves correspond to the values of the parameter @xmath552 ( from top to bottom ) , assuming mixed disorder .
dashed line represents the resistivity of the non - interacting system .
typical experimental parameters are used : effective mass @xmath553 g , electron density @xmath554 , modulation strength @xmath555 , modulation period @xmath556 nm , momentum and single - particle relaxation times @xmath557pc and @xmath558pc , respectively .
, title="fig:",width=302 ] .
( a ) quasiclassical weiss - oscillations ; dashed line shows the resistivity in the absence of modulation .
( b ) interaction - induced quantum oscillations in @xmath537 for three temperatures .
the curves correspond to the values of the parameter @xmath552 ( from top to bottom ) , assuming mixed disorder .
dashed line represents the resistivity of the non - interacting system .
typical experimental parameters are used : effective mass @xmath553 g , electron density @xmath554 , modulation strength @xmath555 , modulation period @xmath556 nm , momentum and single - particle relaxation times @xmath557pc and @xmath558pc , respectively . , title="fig:",width=302 ] let us summarize the results obtained in this subsection .
we have shown that in a periodically modulated system the interaction induces , in addition to the quadratic mr studied in secs .
[ iii ] and [ iv ] , an oscillatory contribution to the component @xmath537 of the resistivity tensor , which is not affected by modulation ( and thus shows no oscillations ) within the boltzmann theory . when the parabolic mr is negative ( meaning that the exchange contribution dominates ) , which is the case under typical experimental conditions and for not too high temperatures , these quantum interaction - induced oscillations in @xmath537 are _ in phase _ with classical oscillations in @xmath10 , as follows immediately from eq .
( [ interact - weiss ] ) [ see fig . [ weiss ] ] .
in other words , their phase is opposite to that of the above - mentioned contribution @xmath533 induced by the dos oscillations .
we come therefore to the following conclusion concerning the phase of the total oscillatory contribution to @xmath537 . while at sufficiently high temperatures the @xmath537-oscillations have , due to the contribution @xmath533 [ and possible due to the hartree counterpart of eq .
( [ interact - weiss ] ) ] , the phase opposite to @xmath534 , with lowering temperature the exchange contribution eq .
( [ interact - weiss ] ) starts to dominate , implying that @xmath537 oscillates in phase with @xmath10 .
furthermore , the both contributions are damped differently by disorder : the high - temperature out - of - phase oscillations @xmath533 vanish with lowering @xmath6 much faster that the low - temperature in - phase interaction - induced oscillations @xmath549 .
our results are in qualitative agreement with a recent experiment@xcite .
it was observed there that at sufficiently high temperature , @xmath559 , the oscillations in @xmath537 have the opposite phase with respect to @xmath10 , in accord with earlier experimental findings@xcite .
however , when the temperature was lowered , the phase has changed and @xmath537 started to oscillate in phase with @xmath10 , with an amplitude increasing with decreasing @xmath2 .
in addition to these novel oscillations , a smooth negative mr was seen to develop in the same temperature range .
the authors of ref . emphasized a puzzling character of the temperature dependence of the observed oscillations , which can not be explained by earlier theories@xcite discarding the interaction effects .
our theory leading to eq .
( [ interact - weiss ] ) provides a plausible explanation of these experimental findings .
quantitative comparison of the theory and experiment requires , however , a more systematic experimental study of the temperature dependence of the amplitude of @xmath537-oscillations in a broader temperature range .
let us summarize the key results of the present paper .
we have derived a general formula ( [ sigma ] ) , ( [ bwq_general ] ) for the interaction - induced quantum correction @xmath0 to the conductivity tensor of 2d electrons valid for arbitrary temperature , magnetic field and disorder range .
it expresses @xmath0 in terms of classical propagators in random potential ( `` ballistic diffusons '' ) . in the appropriate limiting cases , it reproduces all previously known results on the interaction correction ( see sec . [ iic ] ) . applying this formalism
, we have calculated the interaction contribution to the mr in strong @xmath6 in systems with various types of disorder and for arbitrary @xmath1 . in the diffusive limit , @xmath4
, the result does not depend on the type of disorder , as expected .
specifically , the mr scales with magnetic field and temperature as follows , @xmath560 and @xmath561 , in agreement with refs . and . in the ballistic limit , @xmath194 , the result
is strongly affected by the character of disorder . in sec .
[ iii ] we have performed a detailed study of the case of smooth disorder characteristic for high - mobility gaas heterostructures .
we have found that the temperature - dependent mr scales at @xmath382 as @xmath562 and @xmath563 .
in addition , there is a temperature - independent ( but larger ) contribution @xmath564 to the hall resistivity . in the opposite limit @xmath38
the mr is suppressed .
we have further considered a mixed disorder model , with strong scatterers ( modeled by white - noise disorder ) superimposed on a smooth random potential ( sec . [ v ] ) .
a qualitatively new situation arises when the momentum relaxation rate @xmath565 due to smooth disorder is much less than the total momentum relaxation rate @xmath47 .
such a model is believed to be relevant to si - based structures , as well as to gaas structures with very large spacer .
we have shown that in the ballistic limit and at @xmath382 the corrections to both longitudinal and hall resistivities are enhanced ( as compared to the case of smooth disorder ) by a factor @xmath566 . in the range of weaker magnetic fields , @xmath38 , the interaction - induced mr scales in the ballistic regime as @xmath567 and
@xmath568 $ ] . for a weak interaction ( @xmath270 )
the correction is dominated by the exchange contribution , implying that @xmath534 is negative and that the slope of @xmath379 decreases with increasing temperature .
this is true up to a temperature @xmath569 ( defined in sec .
[ iiid ] ) where the sign changes . in the case of a strong interaction
the magnitude of the hartree contribution ( and thus the sign of the total correction ) depends on angular harmonics @xmath570 of the fermi - liquid interaction ( sec .
[ iiie ] ) .
it is worth emphasizing that in contrast to the diffusive limit where only @xmath348 is relevant , in the ballistic regime all the fermi - liquid parameters are , strictly speaking , important , see eq .
( [ fermi - hartree - ball ] ) .
therefore , predictions of the `` @xmath348-approximation '' ( with only one fermi - liquid parameter retained ) should be treated with caution .
we have further applied our formalism to anisotropic systems ( sec .
[ vi ] ) and demonstrated that the correction mixes the components @xmath10 and @xmath537 of the resistivity tensor .
this result is of special interest in the case of systems subject to a one - dimensional periodic modulation ( lateral superlattice ; wave vector @xmath571 ) .
specifically , we have shown that the interaction induces novel oscillations in @xmath537 , which are in phase with quasiclassical commensurability ( weiss ) oscillations in @xmath10 .
our results for @xmath10 in the case of smooth disorder ( published in a brief form in the letter @xcite ) have been confirmed by a recent experiment on @xmath399-gaas system @xcite , which was performed in the broad temperature range , from the diffusive to the ballistic regime .
specifically , li _ et al . _
@xcite found that the mr scales as @xmath572 in strong magnetic fields .
the obtained temperature - dependence of the proportionality coefficient @xmath573 was in good agreement with our predictions .
very recently , olshanetsky _ et al . _
@xcite studied the mr in the ballistic regime in a si / sige structure of @xmath399-type , where both short- and long - range potential are expected to be present .
they found the interaction - induced correction to @xmath10 larger by a factor @xmath574 as compared to our prediction @xcite for the case of smooth disorder .
this conforms with the results of the present paper for the mixed - disorder model , where we find an enhancement of @xmath534 by a factor @xmath540 . as has been mentioned in introduction , the interaction - induced mr in the ballistic regime
was measured for the first time as early as in 1983 , by paalanen , tsui , and hwang@xcite , who studied gaas structures .
again , a parabolic temperature - dependent mr @xmath534 was obtained , in agreement with our findings . however , its magnitude was considerably ( roughly an order of magnitude ) larger compared with our theoretical result for the case of smooth disorder , as well as with the recent experiment @xcite .
we speculate that samples used in ref
. probably contained an appreciable concentration of background impurities , which has led to an enhancement of the interaction - induced contribution to resistivity , similarly to the recent work @xcite .
( indeed , the results for the mixed - disorder model shown in fig . [ mix ] may create an impression that the @xmath575 behavior extends up to @xmath576 , as was concluded in ref . .
) remarkably , the interaction - induced quantum correction to conductivity may serve as an indicator of the dominant type of disorder . to the best of our knowledge ,
no experimental study of the interaction effects on hall resistivity @xmath379 has been published .
this part of our predictions therefore awaits its experimental verification .
finally , our results for systems with one - dimensional periodic modulation are in qualitative agreement with the recent work by mitzkus _
@xcite , as we discussed in detail in sec .
quantitative comparison of the theory and experiment requires an experimental study of the temperature - dependence of novel oscillations ( found experimentally in ref . and theoretically in the present paper ) in a broader temperature range . before closing the paper
, we list a few further applications of our formalism and its possible generalizations .
first , our results can be generalized to frequency - dependent ( rather than temperature - dependent ) mr .
second , the interaction effects in systems of other dimensionality , as well as in macroscopically inhomogeneous systems , can be investigated by our general method .
third , the formalism can be used to calculate the phonon - induced contribution to resistivity , which becomes larger than that due to coulomb interaction at sufficiently high temperatures .
further , thermoelectric phenomena in the full range of magnetic fields and temperatures can be studied in a similar way .
finally , our approach can be generalized to the regime of still stronger magnetic fields , where the landau quantization can not be neglected anymore ; the work in this direction is in progress @xcite .
we thank p.t .
coleridge , a.v .
germanenko , z.d .
kvon , g.m .
minkov , c. mitzkus , a.k .
savchenko , and d. weiss for informing us on experimental results prior to publication and for stimulating discussions .
this work was supported by the schwerpunktprogramm `` quanten - hall - systeme '' and the sfb195 der deutschen forschungsgemeinschaft , and by the rfbr .
as discussed in sec . [ iia ] , diagrams _ f _ and _ g _ give rise , in addition to the contribution ( [ sigmaf+g ] ) , to a term of the type ( [ sigma - inel ] ) , characteristic for inelastic effects .
this yields at @xmath193 a disorder - independent correction to resistivity @xmath577 , see below .
note that such a contribution to resistivity would be obtained if one substitutes the inelastic relaxation rate of a clean 2d electron gas , @xmath578 in the drude formula ( [ drude ] ) .
however , such a procedure clearly makes no sense . indeed , in a translationally invariant system electron - electron collisions conserve the total momentum and
thus give no contribution to resistivity .
therefore , the correction ( [ sigma - inel ] ) ought to be canceled by some other contribution .
below we show explicitly that this is indeed the case , and that this second contribution is of the coulomb - drag type , described by the diagrams in fig .
[ drag ] . .
, width=302 ] for simplicity , we restrict our consideration here to the case of zero @xmath6 and white - noise disorder , which allows us to use the results of ref . for the coulomb drag .
note that while ref . considered the drag between two layers , we refer to the `` self - drag '' within a single layer . as we will see below ,
the characteristic momenta @xmath147 determining the contribution ( [ sigma - inel ] ) are large , @xmath579 . for this reason ,
there is no need to take into account impurity - line ladders while evaluating this term , similarly to the calculation of drag in ref . for a small inter - layer distance .
we thus have b_xx^f(,)&= & 12v_f^2 + & & re , [ deltabf ] where @xmath580 are the disorder - averaged retarded and advanced green s functions . using the identity @xmath581,\ ] ]
we reduce ( [ deltabf ] ) to the form b_xx^f(,)&= & -^2 v_f^2 p_x q_x + & & re[g_r(,p)g_a(-,p - q ) ] + & = & -^2 q_x^2 2 v_f^2 i m ( , ) , [ bf - impi ] where @xmath582 is the polarization operator ( [ polarization ] ) , ( , ) & = & g_r(,p)g_a(-,p - q ) + & & 2(qv_f- ) , [ impi - ball ] where @xmath583 is the step function . furthermore , the imaginary part of the interaction propagator within the rpa is proportional to @xmath584 u(,)=-|u(,)|^2 i m ( , ) .
[ imu ] substituting ( [ bf - impi ] ) and ( [ imu ] ) in ( [ sigma - inel ] ) , we finally obtain ^inel_xx&= & -e^2 ^2m^2 _ -^ + & & q_x^2|u(,)|^2[im(,)]^2 , [ sigma - in - impi ] this expression is identical , up to a sign , to the result of ref . for coulomb drag .
this demonstrates that two contributions indeed cancel each other , ^inel+^drag=0 .
[ nol ] using the explicit form of @xmath584 , eq .
( [ impi - ball ] ) , and of @xmath245 , eq .
( [ screen ] ) , in the ballistic regime , it is easy to estimate @xmath585 ( we assume here @xmath586 for simplicity ) , ^inel ~-e^2 ^2 t _ 0^t d^k_f q dqk_f^2 ~-e^2(t)^2 .
[ sigma - inel - estim ] as has been stated above , the @xmath147-integral is determined by the ultraviolet cutoff .
finally , we note that in double - layer system the inter - layer interaction does give rise to a correction @xmath585 to the driving - layer conductivity , which is equal in magnitude and opposite in sign to the transconductivity .
this effect is , however , reduced by a factor @xmath587 ( where @xmath588 is the interlayer distance ) , as compared to ( [ sigma - inel - estim ] ) , see ref . .
in this appendix we will derive the general expressions ( valid for arbitrary magnetic field ) for the kernels @xmath433 and @xmath589 in terms of the quasiclassical propagator for a white - noise random potential .
this will allow us to reproduce the results of refs . , where the interaction - induced corrections to @xmath31 and @xmath379 were studied for a white - noise disorder in the limit @xmath167 .
we will further apply the formalism to calculate the longitudinal mr and the hall resistivity in a _
finite _ magnetic field with @xmath38 .
the resistivity tensor in yet stronger magnetic field , @xmath382 , is studied , in the more general framework of a mixed disorder model in sec .
[ v ] . using eqs .
( [ bwq_general ] ) and ( [ brho ] ) , we get b_xx^()&=&1 2 ^ 2- 1 n_xn_x d + & + & 12-n_x d n_x + & -&2n_x d n_x d+2_cn_x d n_y d + & -&1-_c^2 ^ 2 ^ 2 n_x d n_x d + & + & 2_c n_x d n_y d[brhoxx - wn ] for the kernel describing the longitudinal resistivity , and b_xy^()&=&_c2- 1 n_xn_y d - n_x d n_y + & -&2n_x d n_y d - 2_cn_x d n_x d + & -&1-_c^2 ^ 2 ^2 n_x d n_y d + & -&2_c n_x d n_x d[brhoxy - wn ] for the hall resistivity .
the propagator @xmath590 in the case of white - noise disorder can be expressed through the propagator @xmath591 , obeying the liouville - boltzmann equation with only scattering - out term present in the collision integral , -i+ i q v_f ( -_q ) & + & _ c + 1 ] d_0(, ) + & = & 2(- ) .
[ lb - out - wn ] as in a zero magnetic field , the total propagator is given by the sum of the ladder - diagrams ( thus including the scattering - in processes ) , yielding ( , )=d_0(,)+ , [ prop - wn ] which we write symbolically as follows = d_0 + .
[ prop - wn - symb ] here we introduced a short - hand notation g_0(,)_0 = _
0(,;, ) [ g0-def ] for the angle - averaged scattering - out propagator .
it turns out that for a white - noise disorder the kernels @xmath433 and @xmath589 can be expressed in terms of @xmath592 ( and its derivatives with respect to @xmath147 and @xmath148 ) .
the solution of ( [ lb - out - wn ] ) is given by _
0(,;, ) & = & \{i q r_c[(-_q)-(-_q ) ] } + & & _ n=-^. [ d0-wn - solution ] it is worth mentioning that in the mixed - disorder model introduced in sec .
[ v ] with both , white - noise and smooth disorder present , the solution of the liouville - boltzmann equation also has the form ( [ prop - wn ] ) . in that case
, the propagator @xmath593 satisfies the liouville - boltzmann equation for a purely smooth disorder ( considered in appendix [ a4 ] ) with the replacement @xmath594 , where @xmath425 is relaxation time due to white - noise potential . using ( [ d0-wn - solution ] ) and a series representation for the bessel functions , we find [ see , e.g. ref . ]
g_0(,)&=&i_c_n + & = & i_c , [ g0wq - wn ] where @xmath595 is the bessel function and = _ c + i_c .
[ mu - def ] in the absence of magnetic field ( @xmath596 ) the propagators @xmath597 and @xmath437 acquire a simple form & & d_0(,;,)= , + [ d0b=0 ] + & & g_0(,)=1 .
+ [ g0b=0 ] to proceed further , we first reduce [ using ( [ prop - wn - symb ] ) ] the `` matrix elements '' appearing in ( [ brhoxx - wn ] ) and ( [ brhoxy - wn ] ) to the form containing only the propagators @xmath593 , & = & _ 0- g_0 , [ < d>-wn ] + & = & , [ < dd>-wn ] + n_xn _ d & = & , [ dnnd - wn ] + n_x d n _ & = & n_x d_0 n
_ + & + & n_x d_0 _ 0 n _ -g_0 , [ ndn - wn ] + n_x d n_&= & n_x d_0 n__0 + & + & , [ ndnd - wn ] + n_x d n _ d&= & + & + & , [ dndnd - wn ] where @xmath598 .
next , using ( [ d0-wn - solution ] ) and ( [ lb - out - wn ] ) and performing the averaging over @xmath599 , we can express the matrix elements involving @xmath593 via the propagator @xmath592 . introducing the notation @xmath600 , we get the following @xmath599-averaged matrix elements , _ 0d_0&=&-ig_0 , [ < d0d0 > ] + _ 0 n_xn_x d_0&= & -12 q^2v_f^2 ^ 2 + & + &
_ c^28 v_f^2(g_0q)^2 , [ < d0nx><nxd0 > ] + n_x d_0 n_x&= & _
c^24 v_f^2(^2 g_0q^2 + 1qg_0q ) + g_02 , + [ < nxd0nx > ] + n_x d_0 n_x d_0 & = & -w2 q v_f^2g_0q , [ < nxd0nxd0 > ] + n_x d_0_0 n_x d_0&= & [ 1-w g_0]2 q v_f^2g_0q , [ < nxd0><d0nxd0 > ] + _ 0 n_x d _ 0 n_x d_0&=&-1 4v_f^2(^2 g_0q^2 + 1qg_0q ) , + [ < d0nxd0nxd0 > ] + _ 0 n_x d_0 ^ 2 & = & -12 v_f^2(g_0q ) ^2 , [ < d0nxd0>2 ] for the `` longitudinal correlators '' , and _ 0 n_xn_y d_0&= & _ c 2 q v_f^2[1-w g_0]g_0q , [ < d0nx><nyd0 > ] + n_x d_0 n_y&= & _ c w2 q v_f^2g_0q , [ < nxd0ny > ] + n_x d_0 n_y d_0 & = & _ c 4 v_f^2(^2 g_0q^2 + 1qg_0q ) , [ < nxd0nyd0 > ] + n_x d_0_0 n_y d_0&= & _ c 4 v_f^2(g_0q ) ^2 , [ < nxd0><d0nyd0 > ] + _ 0 n_x d _ 0 n_y d_0&= & i2_c(g_0 + iwq v_f^2g_0q ) , + [ < d0nxd0nyd0 > ] for the `` hall correlators '' .
substituting eqs .
( [ < d>-wn])-([<d0nxd0nyd0 > ] ) in ( [ brhoxx - wn ] ) and ( [ brhoxy - wn ] ) , we obtain the kernels @xmath433 and @xmath589 averaged over @xmath599 , b_xx^()(,q)&=&(-g_0)^2 + & & \ { 2-g_02 ^ 2 . + & + & ig_0 -1qv_f^2 ^ 2g_0q + 1 4v_f^2 ^ 3(g_0q ) ^2 + & & + & + & .14v_f^2 ^ 2(^2 g_0q^2 + 1qg_0q ) } + [ brhoxx - wn - g0 ] + b_xy^()(,q)&=&(-g_0)^2 + & & \ { -i4_c^2(g_0 + iwq v_f^2g_0q ) .
c2q v_f^2 + & + & _ c g_04 v_f^2 ^ 2(^2 g_0q^2 + 1qg_0q ) + & + & ._c 4 v_f^2 ^ 2 + g_0 -g_0 ( g_0q ) ^2 } .
[ brhoxy - wn - g0 ] in zero magnetic field , we put @xmath601 and substitute @xmath602 in ( [ brhoxx - wn - g0 ] ) . after some algebra , we reduce the obtained expression for the kernel @xmath230 to the form b_xx^()(,q)&=&(qv_f)^2 2 ^ 3 s^3(s-1/)^3 + 3(qv_f)^2 4 ^ 2 s^3(s-1/)^2 + & + & + , + [ bxx - zna ] which agrees with eq .
( 16b ) of ref . up to an overall factor
@xmath603 related to different normalization . in the ballistic limit , @xmath194 , expanding ( [ bxx - zna ] ) in @xmath47
, one finds the leading contribution [ @xmath604 given by the last two terms in ( [ bxx - zna ] ) , b_xx^()(,q ) & & + , + & = & , + [ bxx - zna - ball ] where @xmath605^{1/2}$ ] .
substituting ( [ bxx - zna - ball ] ) in ( [ rho ] ) and using ( [ u - ball ] ) for exchange interaction , we reproduce the linear - in-@xmath2 correction to the resistivity in the ballistic regime , = -te_f .
[ rhof - ball - b0 ] within the approximation of isotropic interaction used in ref .
, the hartree term is determined by the triplet channel and is given by = -3 f_0 ^ 1+f_0^te_f .
[ rhoh - ball - b0 ] it is worth noting that one should exercise a certain caution when comparing the experimental data with the results ( [ rhof - ball - b0 ] ) and ( [ rhoh - ball - b0 ] ) , even in systems with short - range impurities .
first , the higher angular harmonics @xmath606 of the interaction @xcite ( neglected in the above approximation ) may change the numerical coefficient in front of the hartree term ( see discussion in sec .
[ iiie ] and in ref . ) .
second , anisotropy of the impurity scattering introduces an extra factor @xmath607 ( where @xmath408 is the effective impurity - backscattering probability ) in both exchange and hartree terms ( see sec . [ iic]3 and appendix [ a3 ] ) .
the anisotropy may arise due to some amount of smooth disorder present in any realistic system , due to a finite range of scatterers , or due to the screening of originally point - like impurities ( see sec .
therefore , the interaction parameter @xmath348 extracted from the measured linear - in-@xmath2 resistivity with the use of ( [ rhof - ball - b0 ] ) , ( [ rhoh - ball - b0 ] ) may differ considerably from that found from a measurement of other quantities ( e.g. the resistivity correction in the diffusive limit or the spin susceptibility ) . to find the leading contribution to @xmath589 in the limit of vanishing magnetic field
, we have to expand the propagator @xmath592 up to the second order in @xmath86 in the first term in curly brackets in ( [ brhoxy - wn - g0 ] ) .
this can be easily done by treating the term @xmath608 in ( [ lb - out - wn ] ) as a perturbation , which yields g_0(b0)&=&g_0(,q;b=0)+_c^2 h(,q ) + & = & 1s-_c^2 .
[ g_0(b ) ] after a simple algebra , we find @xmath589 in the following form , & = & ( qv_f)^2 ^2 s^3(s-1/)^3 + ( qv_f)^2 [ 2s -5 w]4 ^2 s^5(s-1/)^2 + & + & , which agrees with eq .
( 16a ) of ref . .
in the ballistic limit , @xmath194 , the leading contribution [ @xmath609 to @xmath589 has the form b_xy^()(,q)[6s_0 ^ 2 - 3is_0 + 5 ^2 ] .
[ bxy - zna - ball ] in arbitrary magnetic field , eq .
( [ brhoxx - wn - g0 ] ) can be also significantly simplified when the condition of the ballistic regime , @xmath194 , is assumed .
then the leading contribution to the longitudinal mr , @xmath610 is determined by the kernel b_xx^()(,q ) & & g_0 ^ 2 + + & + & ig_0 -_c^2 4 v_f^2(g_0q ) ^2 .
[ brhoxx - wn - ball ] the remaining terms in ( [ brhoxx - wn - g0 ] ) yield the contributions to the mr which are smaller at least by an additional factor @xmath611 . using ( [ g_0(b ) ] ) [ which tells us that for @xmath38 the magnetic - field - induced corrections to the propagator @xmath592 are small by a factor @xmath424 ] , we find that the mr for not very strong magnetic fields , @xmath38 , is determined by a quadratic in @xmath86 correction to the kernel @xmath433 , b_xx^()(,q ) & = & _ c^2 + & = & -_c^2 + & & -_c^2 ( 4s_0 ^ 4 + 13 is_0 ^ 3 .
25 ^2 s_0 ^ 2 - 35 i ^3 s_0 + 35 ^ 4 ) , [ deltabrhoxx - ball ] independently of the relation between @xmath86 and @xmath47 .
similarly , using ( [ g_0(b ) ] ) , one can find the correction to eq .
( [ bxy - zna - ball ] ) in a finite magnetic field @xmath38 , b_xy^()(,q ) = -_c^3[3s_0 ^ 2 + 7 ^ 2 ] .
[ deltabrhoxy ] again , this correction is independent of the relation between @xmath86 and @xmath47 .
the results ( [ deltabrhoxx - ball ] ) and ( [ deltabrhoxy ] ) are used for calculation of the interaction - induced corrections to @xmath10 and @xmath379 for the white - noise disorder and @xmath38 in sec . [ vb ] .
in this appendix we calculate the leading ballistic correction to the conductivity at @xmath193 for a generic scattering cross - section @xmath613 in the case of the coulomb interaction . as explained in sec .
[ iic]3 , this term ( proportional to @xmath1 ) is obtained by substituting the ballistic asymptotics ( [ bwq_ball ] ) of @xmath230 in the general formula ( [ sigma ] ) .
likewise , the interaction propagator @xmath275 entering ( [ sigma ] ) has to be replaced by u(,)=12 [ u - ball ] with the free propagator @xmath614 given by eq .
( [ d - free ] ) . performing the angular integration @xmath615 ,
we get & & _ xx - e^22 ^ 2 t i m _ 0^d ( ) + & & _ 0^q dq [ q^2-(+i0)^2]^1/2 [ q^2-(+i0)^2]^1/2 + i(+i0 ) + & & , [ sigma - ttau ] where we introduced the dimensionless variables @xmath616 and @xmath617 it is convenient to split the interaction propagator as follows 2u(,q)== ( 1- ) , [ u - split ] where @xmath618^{1/2}.$ ] the first term corresponds to a statically screened interaction and is equivalent to a point - like interaction with @xmath619 the second term results from the dynamical weakening of screening . as discussed in sec .
[ iic]3 , the contribution @xmath620 of the first ( constant ) term is proportional to the backscattering probability @xmath621 see eq .
( [ sigma_ball ] ) .
let us show that this follows also from eq .
( [ sigma - ttau ] ) . performing the variable change @xmath622
xx^(1 ) & & -e^22 ^ 2 t i m _ 0^d ( ) ( ) + ( ) & = & _ * c * ds_0 ( -i)(s_0 ^ 2+^2)s_0 ^ 2 + & & .
[ sigma1 ] the contour @xmath623 of integration over @xmath624 in eq .
( [ sigma1 ] ) is shown in fig .
[ contour ] . interchanging the order of integration over @xmath95 and @xmath624
, we see that for any @xmath625 ( i.e. @xmath626 ) the @xmath624-integral converges .
furthermore , transforming the integration contour @xmath627 as shown in fig .
[ contour ] , it is straightforward to reduce @xmath628 to an explicitly real form . therefore , only the singular point @xmath413 ( where the result of @xmath624-integration diverges as @xmath629 , implying that the imaginary part of @xmath628 is determined by a delta - function in @xmath95-integral ) contributes to ( [ sigma1 ] ) , so that @xmath630 . to find the corresponding coefficient
, one can consider the isotropic scattering @xmath631 and to integrate over @xmath95 first , yielding _
xx^(1)=e^2w ( ) t , [ sigma1pi ] in agreement with ( [ sigma_ball ] ) .
note that the integral over @xmath248 is formally divergent at the upper limit .
it should be cut off at @xmath632 yielding a temperature independent contribution @xmath633 which renormalizes the value of the drude conductivity . and
@xmath634 of integration over @xmath624 in eq .
( [ sigma1 ] ) , width=264 ] we now turn to the contribution @xmath635 of the second ( dynamical ) term in the interaction propagator ( [ u - split ] ) , which differs from eq .
( [ sigma1 ] ) by an extra factor @xmath636.$ ] rotating at @xmath625 the integration contour as before , we reduce the @xmath624-integral to the form ( @xmath637 ) ^2_^ , which is again real and thus yields no contribution to @xmath635 .
though the point @xmath413 is singular in this case as well , the singularity is only logarithmic ( @xmath638 ) , so that no contribution proportional to @xmath408 arises .
this can be easily checked by assuming @xmath631 and performing the @xmath95-integration first .
therefore @xmath639 and the linear - in-@xmath2 term is given by eq .
( [ sigma1pi ] ) . in the above consideration
we have expanded the ballistic propagator @xmath75 up to terms with one scattering event . in the case of small - angle scattering
this is justified provided @xmath640 while in the intermediate temperature range @xmath641 processes with many scattering events dominate ( though the particle motion is typically close to the straight line ) .
the term @xmath620 which is governed by anomalous processes of returns in a time @xmath642 is exponentially small in this case , see sec
. [ iic]3 . as to the @xmath635 contribution to the linear - in-@xmath2 term
, it remains zero in this case as well . to demonstrate this , we use eq .
( [ bwq ] ) . in the first and the third terms we can replace @xmath75 by the free propagator ( [ d - free ] )
, the fourth term gives no @xmath1 contribution , while in the second term we should take into account the angular diffusion ( [ csmooth ] ) around the straight trajectory , -n_xdn_x .
combining the contributions to @xmath230 of all the three terms , we get b_xx(,q)= ( 12 + 12 - 1 ) = 0 , [ bxx - ball - smooth ] so that the coefficient of the @xmath1-term indeed vanishes .
in this appendix , we will solve the classical equation for a propagator of a particle moving in a smooth random potential in a magnetic field , ( , ) + & = & 2(- ) .
[ lbsmooth ] here the polar angle of the velocity is counted from the angle of @xmath225 , @xmath643 .
we first consider the diffusive limit , @xmath644 , and solve this equation perturbatively in @xmath147 for arbitrary magnetic field .
putting @xmath244 , we obtain the solution in the form ( , )= _
n e^i n(-)-i+ in_c + n^2/ , [ q=0 ] which is just a standard expansion in eigenfunctions of the liouville - boltzmann operator .
we now treat the term @xmath645 as a perturbation .
the first - order correction to the eigenvalues @xmath646 vanishes , while the second order correction is _
n^(2)=q^2v_f^2211-(i_c+2n)^2 [ second - corr - eigen ] and can be neglected along with @xmath647 in all terms except for @xmath648 in the diffusive limit .
the first order correction to the right eigenfunction for @xmath649 reads _ 0,r^(1)()=-i q v_f 1+_c^2 ^ 2 [ + _ c ] , [ eigen ] while the left eigenfunction differs from ( [ eigen ] ) by a replacement @xmath650 .
thus , in the diffusive limit the propagator has the form @xmath651 \nonumber \\ & \times & \left[1-{i\ : q v_f\ : \tau\ ( \cos\phi'-\omega_c\tau\sin\phi ' ) \over 1+\omega_c^2\tau^2 } \right]\nonumber \\ & + & \sum_{n\neq 0 } { e^{i n(\phi-\phi')}\over in\omega_c + n^2/\tau}. \label{diff - prop}\end{aligned}\ ] ] in a strong magnetic field ( @xmath36 ) one can go beyond the diffusion approximation . in this case
one can represent the propagator in the form ( , ; , )=d(,;, ) , [ sin - trans ] and solve the equation for @xmath652 d(, ) + & = & 2(- ) [ rot - prop ] perturbatively in @xmath147 .
at @xmath244 we have the same solution ( [ q=0 ] ) as in the diffusive limit .
the first order correction to the eigenvalues is now produced by the @xmath653-term in ( [ rot - prop ] ) , @xmath654 with @xmath655 the diffusion constant in a strong magnetic field .
the second order corrections @xmath656 turn out to be small compared to @xmath657 for @xmath658 . as in the diffusive limit , for calculation of @xmath230 the corrections to the eigenfunctions
@xmath659 with @xmath660 can be neglected .
the first - order correction to @xmath661 is found to be ( we drop the term @xmath662 , since it does not contribute to @xmath230 in the leading order ) _ 0^(1)()-iqv_f(_c)^2 , [ psi-0-pert ] leading to eq .
( [ diffuson ] ) . to calculate the kernel @xmath235
, we need a more accurate form of the propagator .
therefore , we should analyze the corrections to the eigenvalues and eigenfunctions of the liouville - boltzmann operator to the next order in @xmath663 .
to do this , it is convenient to perform the transformation ( , ; , )&=&d(,;, ) + & & \{-iqr_c1+^2 [ ^2(- ) .
+ & + & ( - ) ] } , [ sin - cos - trans ] and introduce the dimensionless variables @xmath664 .
the equation for @xmath665 takes then the form ( , ; , ) + & = & 2(- ) , [ tilde - g ] where we performed a rotation @xmath666 . treating for @xmath667 $ ] ( i.e. @xmath658 in a strong magnetic field , @xmath668 ) the term = q^222 + 2iq [ deltal ] as a perturbation to the operator _ 0=-i2
+ q^22 + -^2 ^ 2 , [ l - unperturb ] we find the unperturbed solution _ 0(,)= 2_n , [ tilde - g0 ] and the first - order correction to the eigenvalues @xmath669 . calculating the @xmath649 eigenfunctions and eigenvalue up to the second order in the perturbation ( [ deltal ] ) we finally obtain the singular part of the propagator for @xmath668 with required accuracy , & & d^s(,;,)= 2 + & & [ dsing-2order ] , where @xmath670 and the functions @xmath671 are given by _
r , l(,q)&=&1 - 1 + & + & 1 ^ 2 .
[ chirl-2 ] as to the regular part of the propagator , for @xmath660 it is sufficient to calculate the eigenfunctions to the first order in the perturbation , which yields & & d^reg(,;,)= 2 + & & _
n0e^i n(- ) , [ dreg-2 ] where _ r , l(,q;n)&=&1-i q + & & q^242 .
+ [ psirl - n-2 ] the results ( [ dsing-2order])-([psirl - n-2 ] ) allow us to calculate the kernel @xmath672 in the first non - vanishing order in @xmath673 , see sec .
in this appendix , we assume that the collision integral @xmath73 induces a transport anisotropy , i.e. that the scattering cross - section @xmath674 is _ not _ a function of @xmath675 . the propagator @xmath279 satisfies the equation & & ( , ; , ) + & & = 2(- ) , [ lb - anis ] where ( ) = [ ( ) -()]w(, ) .
[ coll - int - anis ] we first consider the diffusive limit and concentrate on the leading contribution @xmath171 governed by the diffusion mode .
this requires finding a lowest eigenvalue @xmath676 of the operator in the l.h.s . of ( [ lb - anis ] ) and the corresponding left and right eigenfunctions . treating the term @xmath677 perturbatively as in appendix [ a5 ] ,
l()= 1-i q v_f ( _ c + c)^-1 ( -_q ) [ psirl - anis ] and _ 0=d_q_q_-i , [ eigen - anis ] with the diffusion tensor d_=v_f^2 n _ ( _ c + c)^-1 n_. [ diff - tens - anis ] we thus get the result ( [ dmode - anis ] ) for the singular contribution @xmath171 , with @xmath506 given by ( [ psirl - anis ] ) . in a strong magnetic field ( @xmath53 ) , we can go beyond the diffusive limit . proceeding as for an isotropic system , we perform the transformation ( [ sin - trans ] ) . treating the @xmath147-dependent terms in the obtained equation for @xmath678 as a perturbation and keeping the singular contribution
only , we come to the result ( [ anis - ball - diffuson ] ) , where @xmath679 can be represented symbolically as ( ) = 1 + ( ) ^-1c(-_q ) . [ chi - anis - c ] according to ( [ dd])-([dndnd ] ) , we only need to calculate averages of the type @xmath228 and @xmath680 , so that it is sufficient to keep the zero and first harmonics in @xmath95 in eq .
( [ anis - ball - diffuson ] ) . using [ t_xy ]
n_cn_= ( a. sergeev , m.yu .
reizer , v. mitin , cond - mat/0302278 .
this work considers the effects of electron - phonon interaction , as well as of finite thickness of a 2d layer , in zero magnetic field and for white - noise disorder .
the condition of the ballistic regime arises naturally also in the context of fluctuation phenomena in superconductors , see the recent review a.i .
larkin and a.a .
varlamov in _ handbook on superconductivity : conventional and unconventional superconductors _ edited by k. bennemann and j.b .
ketterson , ( springer , berlin , 2002 ) [ cond - mat/0109177 ] and references therein .
since characteristic times corresponding to the temperature @xmath2 and the cyclotron frequency @xmath86 are @xmath681 ( see the end of sec .
[ iiib ] ) and @xmath682 , respectively , a more accurate form of this condition is @xmath683 .
determination of the exact numerical value of @xmath373 requires a more accurate treatment of the ballistic propagator at large momenta and frequency , @xmath684 , than used in course of derivation of ( [ bxy ] ) .
|
we study the interaction - induced quantum correction @xmath0 to the conductivity tensor of electrons in two dimensions for arbitrary @xmath1 ( where @xmath2 is the temperature and @xmath3 the transport scattering time ) , magnetic field , and type of disorder .
a general theory is developed , allowing us to express @xmath0 in terms of classical propagators ( `` ballistic diffusons '' ) .
the formalism is used to calculate the interaction contribution to the longitudinal and the hall resistivities in a transverse magnetic field in the whole range of temperature from the diffusive ( @xmath4 ) to the ballistic ( @xmath5 ) regime , both in smooth disorder and in the presence of short - range scatterers .
further , we apply the formalism to anisotropic systems and demonstrate that the interaction induces novel quantum oscillations in the resistivity of lateral superlattices .
|
models of interacting electrons in one spatial dimension are very valuable for the understanding of strongly correlated systems .
this is because there exist theoretical methods enabling us to determine their physical properties reliably .
indeed , by combining perturbative renormalization group@xcite , bosonization , and bethe ansatz techniques , a wealth of interesting phases in one dimension has been discovered . while some properties of these phases are unique in one dimension , others have their higher dimensional analogs . for example the independent gapless spin- and charge- excitations and the vanishing quasiparticle weight of the _ luttinger liquid _ @xcite are unique in 1d .
however , the fact that it has a finite charge compressibility and drude weight is analogous to a normal metal in higher dimensions . as another example , like systems in higher dimensions , a mott insulating state is realized at half filling for repulsive interactions .
however , the fact that antiferromagnetic long range order is absent and that spin 1/2 excitation exists in the half - filled mott state are special features of 1d .
furthermore , in one dimension there exists a phase , the _ luther - emery liquid _
@xcite , which exhibits a spin gap and no charge gap .
in addition , as in the _ luttinger liquid _ , the dc electric conductivity is infinite .
the above characteristics suggest that the _ luther - emery liquid _ is a 1d analog of a superconductor .
however , up until very recently an important question remained unanswered : `` do electrons pair in the _ luther - emery liquid _ ? ''
the best way to answer that question is to determine whether the magnetic flux period is @xmath3 or @xmath1.@xcite however , since the spin and charge degrees of freedom are manifestly separate in the effective theory describing the _ luther - emery liquid _ , and the vector potential enters only in the charge action , it is difficult to see why the flux period for a _ luttinger liquid _ and a _ luther - emery liquid _ should be different .
+ in a recent paper we addressed these issues in the one - dimensional @xmath4-@xmath5-@xmath6-model in the limit of vanishing exchange couplings .
@xcite fortunately , both a spin gapless phase as well as a spin gapful phase appear in this limit.@xcite in ref .
we have demonstrated that while the flux period is @xmath7 in the former@xcite , it indeed becomes @xmath1 in the latter .
in particular we have shown that as a function of the aharonov - bohm flux , the ground state energy of a spin gapped ring is periodic with period @xmath1 .
due to one - dimensionality the energy barrier between adjacent minima is proportional to the inverse circumference @xmath2 of the ring . for definiteness
, we therefore define the function [ eq1 ] ( ) = _ ll ( e_0()-e_0(0 ) ) where @xmath8 is the ground state energy of the system as function of flux . despite the above progress ,
the question `` do all _ luther - emery liquids _ exhibit an @xmath1 flux period , and hence electron pairing ? '' remains to be answered . in this paper , we show that the answer to the above question is indeed affirmative .
technically we start from the luttinger hamiltonian with the @xmath9 channel scattering.@xcite we bosonize this model using the constructive formalism @xcite which provides rigorous operator identities on the hilbert space of the finite size system .
we show that due to a set of constraints on the total charge / spin number / current operators@xcite , the state of the spin sector impacts the charge sector through a twisting of the boundary condition . as a result , when the spin sector is gapped by the large momentum transfer two body scattering ,
the charge channel flux period becomes @xmath1 .
+ in the literature , the fact that there exist constraints on the total charge / current operators in bosonization has been employed by loss@xcite for spinless fermion systems to study particle number parity effects . regarding spinful fermions , ref .
used a method similar to ours to determine the flux period for the hubbard model .
however , the author concluded that the flux period is always @xmath1 regardless of whether a spin gap exists , which we believe to be in error .
furthermore , a common reasoning encountered in the literature is to attribute the @xmath1 flux period to the dominance of singlet superconducting ( ss ) correlations at long distance and low energy , rather than to the appearance of a spin gap .
it has , however , been noted that states with dominant charge density wave ( cdw ) correlations may also feature this anomalous flux period ( see , e. g. , refs . ) .
here we argue that this is just the case when there is a spin gap . in this case
, it is natural to interpret the state as being formed by cooper pairs .
the degree of coherence of these pairs will determine if the state is more appropriately thought of as cdw - like or ss - like on not too large length scales .
in this picture , one naturally expects the flux period to be one half of a flux quantum . in the following
, we will show that regardless of the correlation functions in the charge sector , the existence of a spin gap _ alone
_ indeed causes the @xmath1 flux period in systems with even particle number . + the structure of this paper is as follows : in section [ geology ] we present the luttinger model with large momentum scattering and state the selection rules between charge and current quantum numbers that characterize its hilbert space . in section [ bosonize ]
we briefly review the formalism of constructive bosonization and introduce some notation . in section [ proof ]
we complete the proof that the flux period will be @xmath1 in the presence of a spin gap , and contrast this behavior with that expected in the spin gapless case for even and odd particle number .
we will also comment on finite temperature effects here .
our conclusions are summarized in section [ conclusion ] .
appendix [ appa ] discusses the finite size refermionization of the spin part of the hamiltonian , supplementing our line of arguments given in the bulk of this paper .
appendix [ appb ] is devoted to the use of conjugate phase variables in the construction of klein factors .
the tomonaga - luttinger hamiltonian describes a gas consisting of right and left moving chiral fermions , each suffering small - momentum transfer scattering in a one - dimensional system of size @xmath2 : & & h_tl = h_0+h_2+h_4 [ htl ] + & & h_0=_r , k , s(r v_fk-):c^_rksc_rks:[h0 ] + & & h_2=_q , s , s(g_2||_s , s+g_2_s , -s)_+,s(q)_-,s(-q)[h2 ] + & & h_4=_r , q , s , s(g_4||_s , s+g_4_s , -s):_r , s(q)_r , s(-q ) : + & & [ h4 ] here @xmath10 denotes the allowed momenta under periodic boundary condition , the fermion operator @xmath11 annihilates a right ( @xmath12 ) or left ( @xmath13 ) moving fermion with momentum @xmath14 and spin @xmath15 ( see fig .
[ tl ] ) , @xmath16 , and :o : - _ 0.[fnormal ] right- and left - moving branches of the luttinger model .
the crosses denote the allowed momenta @xmath10 for periodic boundary conditions .
the dark shaded region represents the occupied momentum states in the `` vacuum '' @xmath17 .
the vacuum chemical potential lies between the last occupied and first unoccupied states as indicated by the horizontal line .
the light shaded region corresponds to a different filling .
the fermi momentum corresponding to the latter is given by @xmath18.,width=188,height=188 ] in the above , @xmath19 denotes the expectation value taken in the vacuum state defined as the ground state of eq .
( [ h0 ] ) .
the density operators appearing in eqs .
( [ htl])-([h4 ] ) are defined as [ rhors ] _
r , s(q)_k : c^_r , k+q , sc_r , k , s : the @xmath20-component of these operators , n_r , s_r , s(0 ) measures the extra number of @xmath21 type fermions added on top of the vacuum .
all four integers @xmath22 are conserved by @xmath23 .
these quantum numbers play an important role in the rest of the paper .
their importance in the bosonization procedure has been stressed by heidenreich et al.@xcite and haldane@xcite . out of the four operators @xmath22 we can form the following linearly independent number and current operators : [ qnumbers ] & n_=_r , sn_r , s , j_=_r , srn_r , s + & n_=_r , ssn_r , s , j_=_r , srsn_r , s , where the indices @xmath24 and @xmath25 stand for charge and spin respectively
. it will be important in the following to note that in any one - band model with single particle states symmetrically occupied between @xmath18 and @xmath26 , the total particle number is actually given by @xmath27 the reason for this is that the states at @xmath28 , which consist of 4 degenerate states in the luttinger model rather than 2 , have not been included in the definition of the @xmath22 ( see fig .
[ tl ] ) .
there are important relations between the integer quantum numbers defined in eq .
( [ qnumbers ] ) .
for example @xmath29 are either all odd or all even .
in addition , the average of @xmath30 and @xmath31 has the same even - odd parity as the average of @xmath32 and @xmath33 , while they both have opposite even - odd parity as the average of @xmath34 and @xmath31 .
these constraints are summarized by the following `` selection rules''@xcite : [ sel ] @xmath35 which follow from the definitions eq .
( [ qnumbers ] ) and the fact that the @xmath22 are integer . for most of the paper ,
we shall primarily concentrate on the case where @xmath34 is even . while selection rule eq .
( [ sel1 ] ) then requires the same of all the other quantum numbers , it is the selection rule eq .
( [ sel2 ] ) that imposes a coupling between the spin and charge quantum numbers which ultimately determines the value of the flux period . + the tomonaga - luttinger hamiltonian eq .
( [ htl ] ) is exactly solvable .
@xcite the solution describes a system with gapless spin and charge excitations .
a spin gap may be opened by the addition of the following large - momentum transfer scattering term@xcite : h&=&h_tl+h_1[g - ol ] + h_1&=&h_1,||+h_1 , + & = & -_k , k,q , s , s(g_1||_ss+g_1_-s
s ) + & & : c^_+,k+q , sc_+,k , s::c^_-,k - q , sc_-,k,s:[h1 ] when the number of particles is incommensurate with the number of lattice sites , eq .
( [ g - ol ] ) is the generic hamiltonian including the most relevant two - body scattering terms .
the inclusion of @xmath36 destroys the exact solubility of the model , and at the same time it destroys the conservation of @xmath37 .
however since @xmath36 changes @xmath37 in multiples of @xmath38 , the parity @xmath39 remains conserved . as a result
the selection rules eq .
( [ sel1 ] ) , eq .
( [ sel2 ] ) remain valid even in the presence of @xmath36 .
under suitable choices of parameters , eq . ( [ g - ol ] ) can describe a translationally invariant system of spin-@xmath40 fermions with a spin gap but no charge gap , i.e. , a _ luther - emery liquid_. in the rest of the paper we study the dependence of the ground state energy of such a model as a function of an applied aharonov - bohm flux .
technically we employ the constructive bosonization method@xcite extensively reviewed in refs . and .
in the following we shall just summarize the main bosonization rules . due to the following commutation relation between the density operators = kl2_rr_ss_k ,
k.we define boson creation operators for each momentum @xmath41 and each spin @xmath15 : & & b^_s(q)=_r(rq)_r , s(q ) + & & [ b_s(q),b^_s(q)]=_s , s_q , q , [ b_s(q),b_s(q)]=0,where @xmath42 is the heaviside step function .
the bosonization of the local fermion operators _ r , s(x)=_k e^ikxc_r , s(k ) then proceeds by means of the introduction of a non - hermitian bosonic field [ varphi]_r , s(x)=-n_r , s+i_q0(rq)e^iqx - rq/2b_s(q ) , + in terms of which the fermion creation operators can be written@xcite as:^_r , s(x)=a_r , se^i^_r , s(x)e^i_r , s(x ) e^i_r , s,[bosid]where the factor a_r , s = e^i(r_sn_-r , s+s_sn_r , s),[ars]is introduced to ensure the proper anticommutation relations between the fermion operators ( [ bosid ] ) carrying different @xmath43 and @xmath15 .
it commutes with all the spatially dependent fields in eq .
( [ bosid ] ) .
a positive infinitesimal @xmath44 was introduced in eq .
( [ varphi ] ) to ensure the convergence of commutators between operators .
the operator @xmath45 is conjugate to @xmath22 , = i[comm0 ] .
it is customary to further define local hermitian fields each associated with the spin ( @xmath46 ) or charge ( @xmath24 ) degrees of freedom , [ phi]&&_,(x ) = _ rr(_r,(x)_r,(x)+ ) + _ , + & & _ , = _
rr(_r,_r , ) , as well as their `` dual '' fields , [ theta ] & & _ , ( x)=_r ( _ r,(x)_r,(x)+ ) + _ , + & & _ , = _ r(_r,_r , ) written in terms of the spin and charge boson operators [ mode ] b_,(q)=(b_(q)b_(q ) ) the above local fields read ( @xmath47 ) [ modeex ] _ ( x)&=&_-- _ q0(q ) + & & e^-iqx-|q|/2(b^_(q)+b_(-q))[phiexp ] + _ ( x)&=&_--_q0 + & & e^-iqx-|q|/2(b^_(q)-b_(-q)).[thetaexp ] from eqs . and it is evident that @xmath48 and @xmath49 are the winding numbers of @xmath50 and @xmath51 respectively , and @xmath52 and @xmath53 are the spatial averages of @xmath50 and @xmath51 .
it is simple to check that @xmath52 and @xmath53 are the conjugate operators of @xmath49 and @xmath48 respectively , i.e. , [ comm1 ] = i,=i .
the inclusion of the zero modes @xmath54 and @xmath55 in eqs . and ensures the proper commutation relations of these fields when the system size @xmath2 is finite [ comm2 ] = i_,(x - x ) .
( [ comm2 ] ) suggests that the conjugate operator of @xmath56 is proportional to @xmath57 , i.e , [ comm3 ] & & _ ( x)=-_x_(x ) + & & = i_,(x - x ) .
similarly the conjugate operator of @xmath58 is proportional to @xmath59 , i.e. , [ comm4 ] & & ( x)=-_x_(x)[dens ] + & & = i_,(x - x ) . the physical spin or charge density is given by eq .
( [ dens ] ) , and in the absence of an applied flux , the physical ( spin or charge ) current density is given by [ current ] j_(x)=k_v__(x ) = -k_v__x_(x ) , which follows from the bosonized hamiltonian given below . in the above expressions & & v_= + & & k_= + & & g_2,= + & & g_4,=
.since @xmath60 is just the density of right moving fermions minus the density of left moving fermions , it is appropriate to interpret the coefficient @xmath61 in eq .
( [ current ] ) as the renormalized fermi velocity of the system . in terms of @xmath56 and
@xmath58 the bosonization identity eq .
( [ bosid ] ) reads @xmath62 here , @xmath63 denotes boson normal ordering : all powers of the fields @xmath64 are to be moved to the left of powers of the fields @xmath65 , whereas positive powers of the operator @xmath66 are to appear on the very right , and negative powers of the same operator are to appear on the very left of the expression . by means of eq .
( [ bosid2 ] ) the selection rules eq .
( [ sel ] ) become equivalent to the requirement _
r , s(x)=_r , s(x+l ) r , s .
this clearly illustrates the topological origin of these rules .
+ we may now write the hamiltonian eq .
( [ g - ol ] ) entirely in terms of the bosonic fields introduced above .
the tomonaga - luttinger part of the hamiltonian , including the large momentum scattering term with parallel spin , takes the following quadratic form : @xmath67 where the operators @xmath68 are related to those in eq .
( [ mode ] ) by a bogoliubov transformation .
the large momentum - transfer scattering with antiparallel spin term becomes @xcite : @xmath69 note that the coefficient of @xmath38 in the argument of the cosine assures that the operator does not go out of the physical subspace , as explained above .
the weak coupling renormalization group flow of the system eq .
( [ g - ol ] ) is well known@xcite : for @xmath70 , the operator @xmath71 is relevant and a spin gap will be opened .
this is the case we will focus on in the following . for spin @xmath72 invariant systems @xmath73 . in that case
@xmath70 requires @xmath74 , as discussed by luther and emery.@xcite
by virtue of eqs . and , the model eq .
( [ g - ol ] ) takes the form [ g - ol2 ] h = h_tl+h_1h_+h_where @xmath75 and @xmath76 act exclusively on charge- and spin- degrees of freedom respectively .
the eigenstates are thus direct products of charge states and spin states [ factor ] |c>|s > , and the ground state energy is the sum of spin and charge energies [ energy ] e_0=e_0^c+e_0^s . when @xmath36 causes a spin gap to open up , the spin sector of the model eq .
( [ g - ol ] ) is described by a gapped sine - gordon field theory . in the following
we shall focus on the @xmath77 sector , which is where the gapped spin ground state lies . in this sector
@xmath71 is relevant , and one may interpret the cosine term in eq .
( [ h12 ] ) as a steep potential experienced by @xmath78 . in the limit of infinite system size where true symmetry breaking is possible
, one may think of @xmath78 as being locked to one of the minima of the cosine potential .
when this happens @xmath79 , the spatial average of @xmath56 , will take a c - number value equal to the respective minimum value of @xmath80 . at first , let us neglect the selection rules eq .
( [ sel ] ) .
that is , we start by looking at the problem in the space @xmath81 introduced in section [ bosonize ] , where in particular @xmath82 is an independent integer valued quantum number
. then we may regard the conjugate variable @xmath79 as an angular variable with period @xmath83 .
this notion becomes precise if we identify @xmath79 with its `` lattice version '' discussed in appendix [ appb ] , which we shall do for the present purpose.@xcite within @xmath84 there are four inequivalent minima of the cosine term in eq .
( [ h12 ] ) , and the corresponding ground states in the spin sector can be labeled as [ philock ] |0 > , |/2 > , | > , |3/2>,[fgd]where _
| > = |>.as discussed earlier , the operator ( -1)^j_/2(-ij_/2)commutes with @xmath85 , hence its eigenvalues can be used to classify the spin ground states . unfortunately the states given in eq .
( [ fgd ] ) are not eigenstates of @xmath86 . following appendix [ appb ] ,
it is easy to show that _ |>=(+)| > , where the eigenvalue on the right hand side is to be understood modulo @xmath83 .
we may hence choose the global phases in eq .
( [ philock ] ) such that |z/2>=^z|0>. it is thus easy to form linear combinations |>=_z=0 ^ 3 ^-z ^z|0 > such that |>=|>.we are now in a position to enforce the selection rules . given @xmath87 , the selection rule eq .
( [ sel1 ] ) requires @xmath37 to be even . as a result
only @xmath88 are allowed .
we label these two states by [ pm]|+ > , |->.thus actually , the ground state is only two - fold degenerate .
this degeneracy becomes further lifted in the case of a finite system size @xmath2 , to be discussed next . for finite @xmath2 ,
the notion that the field @xmath80 is locked to a classical value is no longer valid .
in fact for finite @xmath2 , even @xmath79 is subjected to quantum fluctuations .
this is explicit in eq .
( [ htldiag ] ) , where the variable conjugate to @xmath79 , namely @xmath37 , enters the hamiltonian when @xmath2 is finite .
thus the spin ground state can no longer be thought of as one of the `` locked '' spin states given by eq .
( [ philock ] ) . on the other hand , since @xmath89 remains a good quantum number , the states in eq .
( [ pm ] ) are still well defined as the respective ground states in the @xmath90 sectors of the spin hilbert space .
we note that the spin states in eq .
( [ pm ] ) thus defined are not strictly degenerate for finite @xmath2 .
it is important to observe , however , that the difference in energy between these two states vanishes exponentially in the system size @xmath2 .
one way to see this is the well known fact that the gapped sine - gordon field theory is the low energy effective theory of a dimerized spin-@xmath40 chain.@xcite here , the @xmath91 are respectively the symmetric and antisymmetric combination of the two dimer patterns . since the two dimer patterns differ by a macroscopic number of degrees of freedom , the tunnel splitting between these two states should vanish exponentially with the system size . a slightly more direct way to see the above
is offered by the well known mapping between the gapped sine - gordon theory and the massive thirring model.@xcite we will elaborate on this point in appendix [ appa ] .
the advantage of this method is that at the special luther - emery point , it allows us to study the effect of a finite temperature . for the purpose of this paper we may ignore the above exponentially small energy difference between the states eq .
( [ pm ] ) .
this is because such a tiny difference will drop out of in the limit taken in eq .
( [ eq1 ] ) . in this sense
we may still speak of a degeneracy in the spin sector of the model , and regard the spin contribution @xmath92 in eq .
( [ energy ] ) as essentially independent of @xmath89 in the spin gapped case .
naively the spin degeneracy discussed above seems to suggest that the ground state of the full hamiltonian eq . ( [ g - ol2 ] )
is degenerate .
however this is not so , and the reason for this is the selection rule eq .
( [ sel2 ] ) . to demonstrate that let us assume the total particle number to be @xmath93 , whereas @xmath77 . according to the selection rule eq .
( [ sel2 ] ) the spin states @xmath91 may not be combined with the same charge state .
the spin state @xmath94 may only be combined with a charge state whose current quantum number @xmath31 is an odd multiple of @xmath95 and hence non - zero .
the presence of a non - zero current will cost an energy of order @xmath96 as is evident from eq .
( [ htldiag ] ) .
the state @xmath97 , on the other hand , may be combined with a charge state of zero current , which minimizes the charge energy . as a result
there is an energy splitting @xmath98 between the lowest energy state in the @xmath99 and @xmath100 sectors .
we note that an analogous result was discussed by haldane@xcite for the case of a vanishing spin gap and a finite charge gap at commensurate band fillings . in contrast , here we are interested in the effect of an applied aharonov - bohm ( ab ) flux , which is of interest only when the charge sector is gapless .
the coupling to a vector potential @xmath101 is determined by gauge invariance and can be worked out from the minimal coupling requirement .
we only consider the constant vector potential @xmath102 corresponding to an ab flux .
the correct coupling to @xmath0 then follows from the formal replacement@xcite [ abflux ] _
r , s^(x ) e^-ix_r , s^(x ) in the hamiltonian , where a charge @xmath103 is assumed . here
, the boundary conditions of the field @xmath104 remain the same , while the right hand side of eq .
( [ abflux ] ) will in general satisfy different boundary conditions . by eq .
( [ bosid2 ] ) , this is equivalent to the following replacement in the hamiltonian eq .
( [ htl2 ] ) and the current eq .
( [ current ] ) : [ mc ] & & _ ( x ) _ ( x)-x or , by eq .
( [ thetaexp ] ) , simply [ replace ] j _ j_+4 . from eqs . ,
the energy versus flux function @xmath105 in eq .
( [ eq1 ] ) is given by [ result ] e()= _ j_=-2 , 0 , 2 v_k _
( j_+4)^2 here , all multiples of @xmath95 are allowed values for @xmath31 by selection rule eq .
( [ sel1 ] ) .
this leads to the various branches shown in fig .
[ branches ] .
the alternating labels of @xmath90 reflect the fact that the spin state has to be adjusted according to selection rule eq .
( [ sel2 ] ) whenever @xmath31 is changed by 2 . in the presence of a spin gap , however , this does _ not _ affect the energy in the limit of eq .
( [ eq1 ] ) , as discussed above . as a consequence , eq .
( [ result ] ) has an exact period equal to half a flux quantum , shown by the lower envelope in fig .
[ branches]a ) .
we note that these findings are in complete agreement with those obtained in ref .
for the @xmath4-@xmath5-@xmath6 model . the amplitude of the ground state energy modulations is apparently given by [ deltae ] e = k_v_= v_f^ * . where @xmath106 is the renormalized fermi velocity introduced above
the corresponding modulations of the charge current for a given quantum number @xmath31 are given by [ current2 ] |i_&= & -e_0^l dx j_(x ) + & = & v_f^ * _ 0^l dx ( _ x_(x)- ) + & = & -(j_+4 ) + & = & -c e()/l .
the current is thus diamagnetic for @xmath107 and is given by a sawtooth curve in general which one obtains by taking the derivative of the envelope in fig .
[ branches]a ) .
the amplitude of the current is given by @xmath108 , which is the same as that of spinless particles@xcite , although the flux period is halved .
note that this observation is consistent with the notion that the charge of the carriers is effectively doubled . to establish the fact that the @xmath1 flux period is due to the presence of a spin gap
, it is prudent to demonstrate the change of flux period when the spin gap collapses .
first , let us assume @xmath93 as before .
our discussion from the preceding sections generalizes most easily to the case of a vanishing spin gap , if we also assume isotropy in the spin sector : in this case , su(2 ) invariance requires the parameter @xmath109 to be unity at the luttinger liquid fixed point .
we will comment on the general non - isotropic case below .
for gapless spins , the operator @xmath71 in eq .
( [ g - ol2 ] ) is irrelevant , and we may expect to get qualitatively correct results by omitting it . with this simplification , the spin sector becomes analogous to the charge sector , and in particular @xmath37 can be regarded as a good quantum number .
the @xmath99 spin ground state
@xmath97 then has @xmath110 , whereas the state @xmath94 lives in a degenerate doublet space with @xmath111 .
this then raises the corresponding spin energy of the @xmath94 state by a term of order @xmath112 , as shown explicitly in eq .
( [ htldiag ] ) . as a consequence ,
the @xmath100 branches are shifted upward with respect to the @xmath99 branches ( fig .
[ branches]b ) ) which destroys the @xmath1 periodicity of @xmath105 .
the spin current carrying @xmath100 states may ( but need not ) be shifted up in energy so much that @xmath113 ceases to be a metastable minimum of @xmath105 .
this is just the case for a non - interacting system . in the case where the uplifting of the @xmath100 state is not as large
, @xmath113 persists to be a metastable minimum in the energy versus flux curve .
the difference between the ground state energy at @xmath114 and @xmath113 is thus given by : @xmath115 interestingly , eq .
( [ shift ] ) provides information about the luttinger parameters of the _ spin sector_. it has long been known that for a luttinger liquid , the luttinger parameters can be determined from of the ground state properties.@xcite this technique is often applied to infer the charge luttinger parameters , e. g. by calculating the ground state energy as a function of particle density and magnetic flux.@xcite eq .
( [ shift ] ) shows that the same technique may be used to infer spin luttinger parameters , provided that @xmath116 holds . in @xmath72 invariant systems
, the spinon - velocity @xmath117 may thus be obtained .
note that in this case , the @xmath111 , @xmath77 states corresponding to the @xmath118 branches in fig .
[ branches]b ) are degenerate with states having @xmath110 , @xmath119 , which carry no spin current but have a net azimuthal spin projection @xmath120 .
this degeneracy follows from eq .
( [ htldiag ] ) with @xmath121 .
the latter states , however , will generally be lower in energy for spin gapless systems _ without _
su(2 ) invariance .
this follows because one has @xmath122 in this case , since @xmath70 would always lead to a spin gap .
we thus predict that the branches corresponding to the metastable minima in fig .
[ branches]b ) , if present , will have a net spin , rather than a net spin current , in models without su(2 ) invariance . in this case , @xmath117 in eq .
( [ shift ] ) is to be replaced by @xmath123.@xcite when @xmath124 , all the patterns in fig .
[ branches ] are shifted horizontally by @xmath1 . in this case
the global minima in fig .
[ branches]b ) are located at odd multiples of @xmath1 .
hence the function @xmath105 can distinguish the cases @xmath124 and @xmath93 in the case of gapless spins , but not in the case of gapped spins .
the same result had also been observed for the @xmath4-@xmath5-@xmath6 model.@xcite we now turn to the case of odd particle number @xmath125 . in this case
the selection rules eq .
( [ sel1 ] ) requires both @xmath126 and @xmath37 to be odd , reflecting the fact that there must be a dangling spin .
( of course , with a dangling spin the system can not have a spin gap . )
now the quantum number @xmath127 may take the values @xmath128 .
the two corresponding subsets of the spin state space are related by the transformation @xmath129 , which leaves the hamiltonian invariant . as a consequence ,
the spin ground states @xmath130 are exactly degenerate , and an exact @xmath1 periodicity is obtained for the hamiltonian at any system size , regardless of whether @xmath36 is relevant or not .
also , since @xmath31 is now odd as well , the pattern shown in figure [ branches]a ) will be shifted horizontally by @xmath131 .
hence the global minima of @xmath132 will be located at odd multiples of @xmath131 in this case .
the results presented in the preceding section are exact for the hamiltonian eq .
( [ g - ol ] ) , which is believed to be the low energy effective theory for all one - dimensional systems with gapless , linearly dispersing charge degrees of freedom.@xcite these results may thus be expected to be representative for this entire universality class . to rigorously justify this point , the effects of higher order , less relevant operators
should be included into the model studied above .
we will not carry out such a detailed analysis here .
rather , we will point out some expected modifications due to less relevant operators , and argue for the robustness of the basic results derived above by comparing them to special examples of microscopic models , where the features of @xmath132 are known analytically or numerically .
the flux period of the repulsive hubbard model was studied in ref . .
these results agree well with our findings for the spin gapless case .
in particular , for odd particle number @xmath34 the global minima of @xmath132 are at odd multiples of @xmath131 , and the flux period is @xmath1 . while it may seem surprising that the flux period does not distinguish the spin gapless , odd particle number case from the spin gapped case ( except for the position of the minima ) , the microscopic origin of the @xmath1 period is of a rather different nature in the two cases . a more subtle effect may demonstrate this : if one calculates the @xmath132 of _ free _ electrons for odd @xmath34 , one indeed finds that @xmath1 is the flux period .
however one also finds that there exist corrections to the @xmath1 period at order @xmath134 in @xmath105 .
these corrections are due to the band curvature neglected in the hamiltonian eq .
( [ g - ol ] ) .
similar corrections to @xmath135 also exist at odd @xmath34 for the @xmath4-@xmath5-@xmath6 model .
they can be calculated using the method discussed in ref .
.@xcite such corrections in powers of @xmath134 , however , were not found in the @xmath4-@xmath5-@xmath6 model for the spin gapped case , where only exponentially small corrections were observed .
we thus argue that corrections to the @xmath1 flux period generally scale as @xmath134 in the odd @xmath34 case , while they are exponentially small in the spin gapped case .
the behavior in the latter case can be attributed to the fact that the spin gap generally causes an exponentially small sensitivity to boundary conditions in the spin sector .
this point will be further clarified in appendix [ appa ] .
the analysis in the preceding section predicts the ground state to be unique on the branches of @xmath132 which contain the global minima .
we expect this to be obeyed by general hamiltonians .
however , the four - fold degeneracy which we found between the @xmath77 , @xmath136 and @xmath119 , @xmath137 states on the metastable branches in the isotropic , even @xmath34 , gapless spin case is an artifact of our restriction to the luttinger hamiltonian eq .
( [ htl ] ) .
rather , the true eigenstates are given by a triplet and a singlet to be formed from these four states , giving rise to a small splitting .
however , except for this effect , the implied degeneracies at the crossings between the branches remain valid : although the conservation of @xmath31 is approximate once higher order operators are allowed , a change of @xmath31 by 2 implies a change of momentum by @xmath138 . hence the states at a branch crossing will not be mixed , and the cusps in @xmath132 will remain sharp for general models.@xcite finite size studies of the hubbard model@xcite
show that the patterns displayed in fig .
[ branches ] indeed emerge very clearly in numerical simulations carried out at moderate system size , both for the spin gapped ( attractive ) and gapless ( repulsive ) case .
we thus conclude that all systems which can be characterized as luther - emery liquids have the @xmath1 flux period . in particular , deviations from the patterns in fig .
[ branches ] such as the appearance of additional minima at higher fractions of a flux quantum must be attributed to finite size effects .
such additional minima at @xmath139 are known to occur in the large @xmath140-limit of the hubbard model @xcite , or the small @xmath5-limit of @xmath4-@xmath5-type models @xcite , for _ fixed _ system size .
the criterion for such finite size effects to disappear is that the amplitude of the oscillations @xmath141 from eq .
( [ deltae ] ) is small compared to any other energy scale of the system . in the above cases , the relevant competing scale is @xmath142 .
the associated crossover is clearly observed in ref . ,
where the @xmath4-@xmath143 model is studied : for @xmath144 , the model displays the spin gapped behavior shown in fig .
[ branches]a ) .
this is a consequence of the ising spin gap of this model .
the similar crossover for the repulsive hubbard model is shown in ref . , where the pattern of fig .
[ branches]b ) emerges ( with the necessary shift for @xmath124 ) .
we note that the appearance of local minima separated by half a flux quantum from the global minima in the repulsive hubbard model is sometimes interpreted as a sign of pairing.@xcite we stress , however , that this case _ does not _ meet the criterion of a @xmath145 flux period as we define it , since a small splitting of order @xmath146 remains between the two types of minima of @xmath105 , which does not vanish as the system size is taken to infinity . finally , we briefly comment on the expected generalization of our findings to finite temperature .
the behavior stated below can be verified straightforwardly at the special solvable luther - emery point of the model eq .
( [ g - ol ] ) ( see appendix [ appa ] ) . for @xmath147 , we consider the modulations of the free energy @xmath148 as a function of flux .
the particle number is held fixed , i.e. , the averages are taken in the canonical ensemble .
( if the particle number were allowed to fluctuate , the even / odd effects discussed above would considerably weaken the sensitivity to flux . ) under these conditions , the observations made above for the ground state energy will carry over to the free energy as long as @xmath149 .
however , the limit of eq .
( [ eq1 ] ) is not to be taken here , because the amplitude of the free energy modulations is proportional to @xmath150 rather than @xmath151 when @xmath152 . in the spin
gapped , even particle number case it remains true that terms violating the @xmath1 periodicity are suppressed by a factor @xmath153 , where @xmath154 is the spin gap . comparing the two exponential factors
, these @xmath1 violating terms will be negligible until @xmath155 is of the order of the spin gap . at this temperature ,
the amplitude of @xmath132 is already exponentially suppressed , provided that the system is large enough such that @xmath156 is satisfied .
this again shows that the @xmath1 period will be obeyed as long as @xmath151 is the smallest energy scale ( other than temperature ) of the system .
we note that the cusps between the branches in fig . [ branches ] will be smoothened by thermal fluctuations , giving rise to a finite negative curvature and paramagnetic@xcite effects .
in this paper , we demonstrated that the ground state energy of luther - emery liquids will generally exhibit an @xmath1 flux period .
while this statement holds in a strict sense in the limit of large system size , finite size deviations are expected to be exponentially suppressed in the system size .
this result had been anticipated in an earlier work on a particular microscopic realization of the luther - emery liquid , the @xmath4-@xmath5-@xmath6-model.@xcite here , we generalized the result of ref . by showing that the @xmath1 flux period is implied by the widely accepted low energy effective theory describing such a phase . as a result
, we clarify why the state of the spin sector impacts upon the flux period when it is commonly believed that in one dimension spin and charge decouple at low energies .
an important aspect of our findings is that in systems with even particle number @xmath34 , the @xmath1 period is triggered by the spin gap ( i.e. pairing ) alone and is independent of whether the superconducting pair - pair correlations are the dominant long - distance / time correlation function .
this may be of particular value for the correct interpretation of numerical work .
in addition , we have also discussed the expected finite temperature generalization of our findings . in ref .
we stressed the su(2 ) invariance of the model discussed there .
this requirement has been relaxed in the present discussion , where we did not enforce su(2 ) invariance . instead , only the weaker requirement of a conserved @xmath157-component of the spin @xmath158 was found necessary . in the anisotropic case
, we must also require that the spin gapped ground state has @xmath159 ( see footnote ) , which should be automatic in the isotropic case .
our findings underline the intuitive notion that every spin gapped system with linearly dispersing charge modes should share some features of a superconductor .
this work has been supported by doe grant de - ac03 - 76sf00098 .
it is well known that the sine - gordon model [ hsigma ] h_&&= dx + & & - dx , can be mapped onto the massive thirring model.@xcite this mapping permits a rather direct demonstration of the exponentially small energy difference between the @xmath99 and @xmath100 states discussed in section [ gapped ] .
in addition , when @xmath160 , i.e. at the luther - emery point , the massive thirring model reduces to a massive free fermion hamiltonian which allows the exact calculation of various physical quantities . in particular finite temperature results can be obtained at the luther - emery point easily . in the notation established in section [ bosonize ] ,
the mapping onto the massive thirring model can be performed by the introduction of the following spinless fermion operators : @xmath161 here , the symbol @xmath162 denotes a normal ordering convention analogous to that defined in section [ bosonize ] , but where @xmath65 is replaced by the field @xmath163 it is interesting to note that the operators @xmath164 , in terms of which the spin hamiltonian eq .
( [ hsigma ] ) is best analyzed , lead out of the physical hilbert space @xmath165 : apart from affecting the spin current quantum number @xmath37 , they also change the total number of net excited spins , @xmath126 , by @xmath166 . in the physical hilbert space
, such a change must always go along with a change of charge quantum numbers , which are not affected by @xmath164 .
it is quite natural that the action of a single `` fractionalized '' operator such as @xmath164 will lead out of the space of physical states . at this point
the larger space @xmath81 discussed in section [ bosonize ] becomes indispensable , as it allows us to define operators such as @xmath164 in the first place .
it is clear , however , that these operators enter the hamiltonian only in appropriate pairs , which leave the physical subspace invariant . using standard methods reviewed in ref . , it is straightforward to show that the field defined in eq .
( [ referm ] ) satisfies the required anticommutation relations .
the additional factor @xmath167 in eq .
( [ referm ] ) will be commented on below .
for now we note that it gives rise to the following boundary conditions for the spinless fermion fields : [ boundary ] _
r(x+l)=e^i(n_+j_/2 - 1)_r(x ) as is relevant to section [ proof ] , in the following we will concentrate on the case @xmath77 and @xmath37 even .
( [ boundary ] ) then tells us that the sector @xmath100 is represented by fermions obeying periodic boundary conditions , and the @xmath99 sector by fermions obeying antiperiodic boundary conditions . note that without the additional twist in eq .
( [ referm ] ) , it would have been vice versa . using the methods discussed in ref .
, one may now show the equivalence of eq .
( [ hsigma ] ) and h_mtm&= & _ r_0^l dx ( -irv ) : _
r^(x)_x_r(x ) : + & + & _ r_0^l dx m(i)^r : _
r^(x)_-r(x ) : + & + & g_r_0^l dx + & & h_+[hmtm]where@xcite [ params ] v&&=(+4k _ ) + g&&=(-4k _ ) + m&&=here , the symbol @xmath162 is as defined in eq .
( [ fnormal ] ) , but the vacuum state is now the @xmath168 state that is annihilated by the field eq .
( [ phitilde ] ) .
ground state of the fermion hamiltonian eq .
( [ hmtm ] ) for @xmath169 .
the linear branches indicate the massless case where both @xmath126 and @xmath37 are good quantum numbers .
crosses indicate allowed momenta in the @xmath99 sector , boxes those in the @xmath100 sector .
the encircled areas indicate the occupied states on each branch for the ground states in the @xmath77 , @xmath110 sector ( crosses ) , as well as the @xmath77 , @xmath170 sector ( boxes ) .
the latter state is degenerate with the @xmath77 , @xmath171 ground state , where the @xmath28 mode is occupied on the @xmath13 branch instead . for non - zero mass and @xmath77
, a gap opens up and the upper band remains empty while the lower band is completely filled , where the allowed momenta depend on @xmath89 as shown.,width=192,height=192 ] when @xmath70 the sine - gordon model is massive , and the physics of eq .
( [ hmtm ] ) is given by massive spinless fermions .
in particular , the fermionic interaction @xmath172 vanishes for @xmath173 , and the hamiltonian eq .
( [ hmtm ] ) becomes that of a massive free fermion model .
this is the special point identified by luther and emery.@xcite at the luther - emery point the fermion dispersion relation is given by @xmath174 . in a range of @xmath109 values around @xmath40 , the @xmath172-term only gives rise to quantitative corrections .
we now return to the factor @xmath167 in eq .
( [ referm ] ) ( or the boundary conditions eq .
( [ boundary ] ) ) and show that with this factor ( or the boundary condition specified in eq .
( [ boundary ] ) ) the boson and fermion theories are consistent . to illustrate that we compare the ground state degeneracies in the non - interacting massless case for the bosonic and fermionic theories , i. e. , we let @xmath173 and @xmath175 , which results in @xmath176 . in the absence of @xmath177 both @xmath32 and @xmath33
are good quantum numbers .
let us denote the ground state in the @xmath178 sector by @xmath179 . from eq .
( [ htldiag ] ) we recall that @xmath180 is the non - degenerate global ground state of the spin sector , whereas the states @xmath181 , @xmath182 form a degenerate doublet . that this also holds in the fermionic representations of the model is just achieved by the boundary conditions eq .
( [ boundary ] ) ( see fig .
[ ref ] and the caption ) .
note that since the fermions eq . ( [ referm ] ) are _ derived _ in terms of bosons , the fermion occupancies claimed below fig .
[ ref ] follow by derivation , not by definition .
in fact one may easily evaluate , where & & c_r^(k)_0^ldx e^ikx_r^(x ) + & & k = n+(n_+j_/2 - 1 ) by plugging in eq .
( [ referm ] ) , and verify that the occupancies identified for the various states in fig . [ ref ] are correct.@xcite the spin sector of the model is now represented in terms of fermions satisfying a conventional boundary condition given by eq .
( [ boundary ] ) .
the difference between the ground state energy for @xmath99 and @xmath100 thus becomes the change in the fermion ground state energy induced by a change of the boundary condition , or equivalently the modulation of the fermion ground state energy caused by an ab flux .
when the sine - gordon model is massive , the fermions form an insulating state .
then , the sensitivity of their ground state energy to the boundary condition will vanish exponentially with the system size , as is well known from the general arguments given by kohn@xcite and thouless@xcite .
the greatest advantage of the refermionization occurs at the luther - emery point .
for in that case , a non - trivial interacting bosonic theory is mapped onto a free fermion theory .
in particular , at the luther - emery point @xmath169 one obtains from a direct calculation that the energy difference between the ground states for @xmath99 and @xmath100 vanishes as @xmath183 , as we claimed earlier .
the hilbert space of the luttinger hamiltonian eq .
( [ htl ] ) , denoted as the `` physical '' hilbert space @xmath165 in the bulk of the paper , can be decomposed as @xmath184 here , the spaces @xmath185 contain all the degrees of freedom associated with the bosonic excitation spectrum , whereas @xmath186 contains the degrees of freedom of the operator @xmath22 .
the subscript `` @xmath187 '' stands for `` lattice '' and reminds us of the discrete nature of @xmath22 in the physical hilbert space : the hilbert space basis of @xmath186 is given by a set of non - degenerate eigenstates of @xmath22 , whose spectrum consists of all integer numbers . in the process of bosonization , however , we introduce new linear combinations @xmath48 , @xmath49 of the @xmath22 ( eq . ( [ qnumbers ] ) ) .
the spectrum of these new operators is likewise integer , yet not all possible combinations of integer eigenvalues are physically allowed .
this `` residual coupling '' is not evident from commutation relations , since the operators @xmath48 , @xmath49 all commute as the @xmath22 do .
hence , once we bosonize an enlarged hilbert space becomes more natural , where the eigenvalues of the operators @xmath48 , @xmath49 are independent .
this is the hilbert space @xmath81 of `` fractional '' excitations .
the physical subspace @xmath165 is then characterized by the fact that the selection rules eq .
( [ sel ] ) are satisfied .
it is clear that in @xmath81 , the spectrum of the @xmath22 must also contain certain fractional values .
formally , we find it convenient to introduce an even larger hilbert space @xmath188 , where the spectrum of the @xmath22 is _
continuous_. the benefit of this is that the conjugate phase @xmath189 of these operators then becomes meaningful .
this , in turn , allows us to construct unitary `` ladder operators '' which change the eigenvalue of the @xmath22 by _ arbitrary _ amounts .
this formalism is of particular advantage in appendix [ appa ] , where fractionalized spin fermion operators are constructed .
below we present some fine details of this embedding of @xmath165 into the larger space @xmath188 . for this purpose let us consider a single operator @xmath190 and its conjugate variable @xmath191 such that @xmath192=i\ ] ] holds .
an analogy to the quantum mechanics of a point particle moving in one dimension is obtained if we identify @xmath193 and @xmath194 , where @xmath195 and @xmath196 are the coordinate and the momentum of the particle . in this context , it is familiar how to construct a hilbert space @xmath197 such that @xmath190 and @xmath191 are well defined on a dense set , eq .
( [ commb ] ) is satisfied and the spectrum of @xmath190 is unbounded : it is the hilbert space of square integrable functions of the variable @xmath34 . from the commutation relation eq .
( [ commb ] ) , it is clear that the spectrum of both @xmath190 and @xmath191 has to be continuous and unbounded . in particular
, we can construct shift operators @xmath198 satisfying @xmath199=a \ , e^{i\hat\varphi a},\ ] ] which shift the value of @xmath190 by an arbitrary amount @xmath200 .
we note that , as is familiar from the point particle analogy , the `` position '' and `` momentum '' eigenkets @xmath201 and @xmath202 are not strictly contained in the hilbert space @xmath197 of `` proper '' vectors , but are `` generalized '' states in the usual sense . here , we will not attempt to introduce a different notation for proper and generalized states , nor for the proper hilbert space and its extension containing generalized states .
we refer the reader to ref . and references therein for details , and simply note that the kets @xmath201 and @xmath203 satisfy @xmath204 suppose now that a physical problem is defined on a subspace @xmath205 , which is given by the discrete `` lattice '' represented by the eigenkets @xmath201 for integer @xmath34 . since these kets from a countable hilbert space basis in @xmath205 , it is natural and convenient to introduce a new scalar product on @xmath205 via : @xmath206 this differs from the scalar product in @xmath197 only by an infinite multiplicative factor . in the above
, @xmath207 denotes the same vector as @xmath201 , but endowed with a different scalar product .
( [ scalphy ] ) means that the @xmath201 for integer @xmath34 become a complete orthogonal set of _ proper _ vectors within @xmath205 . within @xmath205 ,
one may now define a `` crystal momentum '' operator @xmath208 , whose eigenkets are defined to be @xmath209 one observes that these eigenkets are periodic in @xmath210 with period @xmath83 , hence for definiteness the eigenvalues must be restricted to lie within the `` brillouin zone '' @xmath211 $ ] , where @xmath212\ ] ] holds . if we now denote the restriction of @xmath190 to @xmath205 by @xmath213 , we find that the commutator @xmath214 $ ] is _ not quite _ analogous to eq .
( [ commb ] ) .
this has been examined in detail by schnhammer.@xcite however , for the applications we have in mind here , this difference never matters .
this is so since all physical observables , including the hamiltonian , depend on @xmath208 only via integer powers of @xmath215 , and since the equations hold .
hence @xmath215 and @xmath217 _ act identically _ on @xmath205 , and we may express all observables in terms of _ either _ of these operators . by means of eq .
( [ xtal ] ) , the kets @xmath218 are identified with the following kets in @xmath197 : @xmath219 note that the norm of the right hand side with the scalar product in @xmath197 formally computes to @xmath220 , which by comparison with eq .
( [ scalphi ] ) is larger by an infinite multiplicative constant than @xmath221 .
recall that the same relation also holds for @xmath222 and @xmath223 by definition of the scalar product in @xmath205 .
( [ phiphy ] ) makes it clear that the kets @xmath218 are _ not _ to be identified with the kets @xmath224 .
the latter are not periodic in @xmath210 , and can not be constructed solely from kets @xmath201 with integer @xmath34 .
however , @xmath225 holds .
thus whenever it is clear from the context that we are working in @xmath205 , we may drop all labels `` @xmath187 '' , keeping the periodicity of the states @xmath226 in mind .
we can now define the hilbert space @xmath188 introduced in the main part of the paper as @xmath227 the advantage of embedding @xmath165 into the space @xmath188 is that now the space @xmath228 , satisfying @xmath229 , can be generated easily through the action of the operators @xmath230 , @xmath231 on @xmath165 .
the bookkeeping is greatly simplified by the simple commutation relation of @xmath189 and @xmath22 , eq .
( [ comm0 ] ) , valid on @xmath188 .
we finally note that @xmath81 can be written as a product analogous to eq .
( [ hphys ] ) , involving a space containing bosonic degrees of freedom and four discrete `` lattice '' spaces containing the degrees of freedom of the quantum numbers @xmath48 , @xmath49 .
all the above therefore holds in an analogous way for @xmath81 , and in particular lattice versions of the phase operators @xmath232 , @xmath233 can be constructed if desired . in ref .
, it has been noted that the use of the canonical commutation relations eqs .
, in constructing the klein factors does indeed yield correct results in bosonization , even though these relations can not be rigorously justified for operators that are restricted to a discrete space .
we believe that the embedding procedure discussed here provides a proper explanation for this observation .
|
we study the dependence of the ground state energy on an applied aharonov - bohm flux @xmath0 for the luttinger model with large momentum scattering . employing the method of finite size bosonization
, we show that for systems with a spin gap but with gapless charge degrees of freedom , the ground state energy has an exact period of @xmath1 , i. e. _ half _ a flux quantum , in the limit of large system size @xmath2 .
finite size corrections are found to vanish exponentially in @xmath2 .
this behavior is contrasted to that of the spin gapless case , for both even and odd particle number .
generalizations to finite temperature are also discussed .
|
the bose - einstein condensate ( bec ) of chromium ( @xmath0cr ) , where long range dipolar interaction between atoms plays the dominant role , is a novel kind of nonlinear system becoming available to experiments @xcite .
properties of dipole - dipole ( dd ) interactions , namely their long range character and anisotropy , allow dipolar condensates to exhibit many unusual properties not found in becs with just contact interactions @xcite . in particular the existence of stable isotropic and anisotropic two dimensional ( 2d ) solitons has been predicted for such cold quantum gases @xcite .
recently the bright solitons in quasi-1d dipolar bec with competing local and nonlocal interactions have been studied in works @xcite .
the long - range dipolar interactions become dominant when the local part is detuned to zero by the feshbach resonance ( fr ) techniques , as in the experiment on observation of anderson localization in non - interacting cold quantum gases @xcite . in this particular case
the pure dipolar bright soliton can be observed .
the propagation of such solitons under joint action of the trap and a barrier potential , including processes of crossing the barrier by the soliton during its oscillations in the trap are of the greatest interest to investigate .
such processes have a fundamental importance for the problems of entanglement of quantum solitons in cold dipolar gases .
an interesting limit is the case of strong nonlocality when the dynamics of wavepackets become almost linear .
this paper is devoted to the investigation of the properties of cold quantum gases in the presence of long range dipolar interactions at the mean field level .
first we study oscillations of the dipolar solitons in trap potential using the variational approach ( va ) and the scattering theory .
second we consider the soliton transmission , reflection and splitting through a barrier placed in the center of the trap .
particular attention will be devoted to a process of coalescence of colliding wavepackets at the barrier which could be important for the design of beamsplitters and matter wave interferometers using matter wave solitons@xcite .
finally , by means of numerical simulations of the original dynamical equation we verify the predictions of the va and analyze the results beyond the analytical predictions .
we consider the quasi-1d dipolar bec loaded in a parabolic trap with a barrier at the center of the trap .
the governing equation is the 1d gross - pitaevskii equation ( gpe ) with a nonlocal interaction term @xcite : @xmath1 here @xmath2 corresponds to the transverse trap frequency , @xmath3 , and @xmath4 is the magnetic dipole moment oriented along the @xmath5-direction .
the parameter @xmath6 is connected to the angle @xmath7 which the dipoles form with the longitudinal axis @xmath5 .
if dipoles rotate rapidly in the plane perpendicular to the axis @xmath5 , the parameter @xmath6 can vary from 1 for dipoles oriented along the @xmath5-axis ( @xmath8 ) , to @xmath9 in the case of perpendicular orientation to the @xmath5-axis ( @xmath10 ) @xcite .
the wave function is normalized to the number of atoms comprising the bec , @xmath11 .
the hamiltonian for our model has the following form @xmath12\psi(x).\end{aligned}\ ] ] now we define dimensionless parameters : @xmath13 where @xmath14 is the characteristic scale of the long - range dipolar interactions , and @xmath15 is the background value of the atomic scattering length . the dimensionless s - wave scattering length @xmath16 is expressed in units of @xmath15 due to the choice of the normalization of the wave function .
( [ e1 ] ) now can be written in the dimensionless form as follows : @xmath17where @xmath18 is the mean - field wave function of the condensate , @xmath16 is the local contact interaction term , @xmath19 is the nonlinear coefficient responsible for long range dipolar interactions , assumed to be time - dependent @xcite .
the external trapping potential @xmath20 is @xmath21 the frequency @xmath22 is the longitudinal frequency of the trap , @xmath23 and @xmath24 are the amplitude and the width of the barrier .
potential @xmath25 for the case of a broad soliton , @xmath26 , can be approximated by a delta - barrier @xmath27 with @xmath28 .
the wave function @xmath29 has normalization @xmath30 .
the following two forms for the kernels in the nonlocality term are possible @xmath31 the former kernel corresponds to the dipolar bec in a quasi-1d trap @xcite , while the latter one , which contains a cutoff parameter @xmath32 also proposed for dipolar bec in @xcite ( see also @xcite ) is more convenient for analytical treatment . using the matching conditions for @xmath33 @xmath34
it can be found that @xmath35 .
the comparison of profiles of the kernels ( [ r1 ] ) and ( [ r2 ] ) at this choice of @xmath32 shows very good fit .
the cutoff parameter @xmath32 has the meaning of an effective size of the dipole and the value is fixed by interpolating the function @xmath36 by @xmath37 .
actually , it takes the value of the order of the transverse confinement length , which makes the model one - dimensional , and is the unit length in eq .
( [ gpe ] ) . in the limit @xmath38 , where dd interaction effects dominate over the contact interaction effects , both response functions behave as @xmath39 .
thus , in the following we will use the kernel function @xmath37 as it is more simpler for analytical treatments in the description of dipolar effects in bec , using for example va , where the integrals with @xmath37 can be calculated in explicit form .
to describe the soliton propagation we employ the va @xcite .
the lagrangian for eq.([gpe ] ) has the following form : @xmath40 to derive equations for soliton parameters we use the gaussian ansatz @xmath41 ^ 2}{2a^2(t ) } + ib(t)[x-\zeta(t)]^2 \right.\nonumber\\ & + & \left .
ik(t)[x-\zeta(t)]+i\phi(t)\right\},\end{aligned}\ ] ] where @xmath42 are the soliton amplitude , width , chirp , coordinate of the center of mass , wave vector and linear phase , respectively . calculating the averaged lagrangian @xmath43 with this ansatz
we obtain @xmath44 - v_0 g(a,\zeta , l),\ ] ] where @xmath45const and @xmath46 in computing the integral the shift @xmath47 , @xmath48 and a transformation to the new variables @xmath49 , @xmath50 and @xmath51 have been used . considering the euler - lagrange equations @xmath52 we can derive the system of evolution equations for parameters @xmath53 of the solution ( [ ansatz ] ) .
variations on @xmath54 , @xmath55 , @xmath56 and @xmath57 , respectively , give the following equations @xmath58
@xmath59 @xmath60 @xmath61 the equation for the phase @xmath62 is decoupled from the system , and we did not write it here .
finally , from ( [ avl ] ) we obtain the system of equations for the soliton width and the center of mass @xmath63 where @xmath64 thus the evolution is described by the dynamics of two coupled nonlinear oscillators . from the first equation we can find the fixed point for the soliton width @xmath65 and from the second we can calculate the effective potential @xmath66 for the center of mass of the soliton .
the description of the dipolar soliton dynamics by the system of two coupled nonlinear oscillators was applied successively , for example , in the work @xcite where soliton - soliton scattering in the dipolar bec placed in unconnected layers was considered.this method was also used to observe oscillations of the solitons profile in the quasi-1d dipolar bec in ref.@xcite .
we also want to note that the gaussian trial function gives good results for describing the stationary states and dynamics in quasi-1d and -2d geometries ( see for example @xcite ) . in some cases , for 2d geometries , when the purely dipolar bec profile ( for parameters unstable to collapse ) has blood - cell forms , this ansatz failed , and it is necessary to choose the trial function as the sum of gaussian profiles @xcite . in our case
the failure of the ansatz can occur when the pulse is splitted into two or more parts by the defect .
we start with the case of competing nonlinearities ( @xmath67 ) , namely we consider an attractive local nonlinearity , @xmath68 , and a repulsive dd interaction , @xmath69 , or vice versa . by using ansatz @xmath70 and considering @xmath71 we get the stationary equation @xmath72
let us first consider stationary solitons without external potential ( @xmath73 ) and without defect ( @xmath74 ) . in fig.[q_1_g ] and
[ g_1_q ] by solving numerically eq.([gpe_st ] ) ( by newton iteration algorithm ) we present the existence curves for families of the solutions with different combination of local and nonlocal terms . in fig.[q_1_g](a ) we start with the simple local case ( @xmath75 and @xmath76 ) for which the norm of the solution @xmath77 goes to 0 as the chemical potential @xmath78 goes to 0 . by increasing gradually the coefficient @xmath79 in the nonlocal term we pass through the critical value of @xmath80
for which , at some @xmath81 , the existence curve starts to have a local minimum becoming _ bounded _ by critical value @xmath82 which means that below @xmath83 the solutions do not exist . for two branches in fig.[q_1_g](a ) corresponding to the purely local case ( @xmath75 , @xmath76 ) , and with competing local and nonlocal terms , ( @xmath75 , @xmath84 ) we calculated numerically the width of the stationary solutions @xmath85 and
compared it with the results of the va taken from the condition @xmath86 ( see eq.([ua ] ) ) .
the results are shown in fig.[q_1_g](b ) . for the purely local case ( @xmath76 ) one
observes good agreement between the width calculated numerically and from the va .
_ however by increasing the repulsive nonlocal term the discrepancy between @xmath87 and @xmath65 starts to grow and for @xmath88 results from the va for small @xmath89 does not agree with the width of the solution calculated numerically . _ in fig.[dyn ]
we have checked dynamically the stability of the solutions at the black and red points indicated in fig.[q_1_g](a ) where the derivative @xmath90 has different signs . according to vakhitov - kolokolov stability criterion
these two solutions should have different stability . in fig.[dyn](a )
the profiles of these two solutions are shown and in the panels ( b ) , ( c ) their evolutions are presented . as it is expected
the solution with @xmath91 is stable while the solution with @xmath92 is unstable , which is confirmed by direct numerical simulation in fig.[dyn](b ) , ( c ) .
similar analysis for the existence of the solutions with attractive nonlocal interaction ( @xmath93 ) and with the presence of the local repulsive interaction ( @xmath94 ) is presented in fig.[g_1_q ] .
now , contrary to the previous case , the norm of the solutions is always unbounded which means that @xmath95 as @xmath96 .
also in this case the width of the solutions calculated numerically from eq.([width ] ) and analytically from the va are in very good agreement ( see fig.[g_1_q](b ) ) .
now we consider the dynamics of the soliton with local and nonlocal interactions in the presence of the parabolic potential @xmath97 and the localized defect @xmath98 . since we are interested in competing local and nonlocal nonlinearities ( condition @xmath67 ) in the following we will consider two possible combinations separately : i ) @xmath99 , @xmath100 ; and ii ) @xmath94 , @xmath101 .
let us first concentrate on the case of attractive local ( @xmath75 ) and repulsive nonlocal ( @xmath102 ) interactions when the existence curve is bounded by the critical norm @xmath83 , which occurs above @xmath103 . in fig.[fig1 ] the initial profile of the unperturbed stable soliton taken from the existence curve in fig.[q_1_g ] at @xmath104 and shifted to the position @xmath105 , as well as the corresponding external parabolic potential with the defect are shown .
the shift from the center is used to get oscillations of the soliton in the defect - free parabolic potential . as soon as the soliton is shifted from the minimum of the parabolic potential it starts to accelerate to the center and
the soliton velocity at the center , @xmath106 , will depend on the strength of the parabolic potential @xmath22 and magnitude of the shift @xmath107 according to @xmath108 . in this way by changing the strength of the trap with fixed @xmath107 we can control the soliton velocity in the process of the interaction of the soliton with the defect . in the defect - free parabolic trap the frequency of oscillations of the soliton coincides with the trap frequency ( @xmath109 ) . by fixing the amplitude @xmath23 and the width @xmath110 of the defect and letting the soliton collide with the defect we found three characteristic regimes of the soliton - defect interaction depending on the incoming velocity . in fig.[fig2 ] we present two regimes .
the first one corresponds to a relatively small soliton velocity ( weak trap ) when the soliton is reflected from the defect and becomes `` closed '' in the semi - space ( @xmath111 ) of the parabolic potential ( see fig.[fig2](a ) ) . by increasing the soliton velocity by changing the strength of the parabolic potential we found another limiting case when the soliton has enough kinetic energy to pass through the defect ( see fig.[fig2](b ) ) . to visualize the corresponding evolutions we calculated the norms in the right part and in the left part from the defect as @xmath112 and @xmath113 .
the results are shown in fig.[fig2](c),(d ) which confirm the above mentioned soliton behavior .
we also compared direct numerical calculation of the soliton dynamics with the results obtained from the va . in fig.[fig2](a),(b ) by dashes lines we present the trajectories of the oscillations of the center of mass of the soliton calculated from eqs.([ua ] ) , ( [ uz ] ) .
a more interesting effect of the coherent splitting of the soliton by the defect is observed for intermediate values of @xmath22 .
this is the case when the va fails and only numerical calculation will be presented .
we found ( see fig.[fig3 ] ) that taking the strength of the parabolic potential around @xmath114 one can observe splitting of the soliton into two quasi - symmetrical parts .
this can be seen from the evolution of the norms @xmath115 and @xmath116 in fig.[fig3](b ) , ( c ) .
it should be stressed here that after splitting the soliton lost its solitonic identity as soon as in this case the existence curve has bounded norm @xmath83 and the norm of the each parts is below this critical value ( @xmath117 and @xmath118 ) . to confirm this behaviour we switch off the parabolic trap and take a soliton with initial velocity @xmath119 far from the defect . as it is shown in fig.[fig3_v ] , after interaction with the defect the soliton splits into two packets which transform continuously into linear waves . in the presence of the trap
these linear packets do not escape but they are reflected by the parabolic potential and return to the center where again produce the initial soliton ( `` fusion '' ) in the left or right part of the defect and continue this process through several periods . in fig.[fig3](a )
one observes that after splitting and returning to the center the soliton continues to move towards the negative @xmath120 passing completely through the barrier while in [ fig3](b ) soliton after fusion is reflected from barrier .
it should be stressed that very tiny changes in the initial conditions could affect the condition for the soliton splitting and the dynamics of the soliton after fusion ( transmission or reflection scenarios ) . in this case
the va can be used to find the condition for soliton splitting . as follows from eqs.([var_a ] ) , ( [ ua ] ) , when the kinetic energy of the effective particle @xmath121 is less then the barrier effective potential height @xmath122 , a full reflection of the soliton occurs . in the opposite case
we have full transmission ( compare figs.[fig2 ] ) .
therefore we can assume that when @xmath123 , the partial reflection / transmission should take place .
considering this condition as the soliton splitting condition , we obtain : @xmath124 by comparing the numerical results with the analytical ones taken from ( [ x0 ] ) one observes excellent agreement ( see fig.[x0_v0 ] ) . for smaller strength of the repulsive nonlocal nonlinearity , when the norm of the solution is unbounded , one observes a rather different picture of splitting of the soliton at the defect ( see figs.[fig3_g0 ] ) comparing fig.[fig3 ] and fig.[fig3_g0 ] one can conclude that in the case @xmath102 the soliton transforms into two quasi - identical wave packets in fig.[fig3](a),(b ) ( one also observes this effect looking at @xmath125 when along a half period @xmath126 ) while in the case @xmath76 there is no splitting of the soliton . instead of that it transforms into the defect mode localized at the defect .
this can be verified in fig.[q_1_g_v ] by comparing the existence curves for the case without defect with curves calculated in the presence of the defect .
as one can see in fig.[q_1_g_v ] by switching the defect on the existence curves go upper and what is essential is that the critical value of @xmath79 at which the existence curves become bounded also decreases with presence of the defect . as an example
the existence curve for @xmath84 in the defect - free case , @xmath74 , is unbounded while in the presence of the defect with @xmath127 , the existence curve becomes bounded . considering the case
@xmath76 the soliton at @xmath104 and @xmath74 has the same number of particles as the defect mode at @xmath128 and @xmath127 ( see the lower horizontal dotted line in the left panel of fig.[q_1_g_v ] ) . to check this we calculated the defect mode for @xmath128 in the presence of the defect and compared it with the defect mode obtained by direct dynamical calculation where the initial soliton for @xmath104 and @xmath74 has the same number of particles as the defect mode . in the right panel of fig.[q_1_g_v ] we compare these two defect modes ( time @xmath129 corresponds to the instant when @xmath130 during interaction of the soliton with the defect ) .
in the case @xmath84 the situation is different .
the number of particles in the soliton at @xmath104 and @xmath74 is below the critical number of particles , @xmath83 , needed to generate the defect mode in the presence of the defect ( there is no intersection of the horizontal dotted with the upper dashed line ) . in this case
one can observe splitting of the soliton into two linear wave packets after interaction with the defect .
[ [ localization - delocalization - transition - governed - by - the - nonlocal - term . ] ] localization - delocalization transition governed by the nonlocal term .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us consider a linear variation in time of the strength of the nonlinear nonlocal coefficient of the form @xmath131 where @xmath132 and @xmath133 correspond to the initial strength of the nonlocal coefficient and its value at the turning point , @xmath134 , with the return to the initial value at @xmath135 .
tuning the dipolar interactions in quantum gases can be achieved using for example time dependent control of the anisotropy of dipolar interactions suggested in @xcite . as it was shown in section 4.1 , in the presence of the attractive local and the repulsive nonlocal interaction
( see fig.[q_1_g ] ) one can have a transition between unbounded and bounded cases of the existence curves . using a time dependent nonlocal coefficient @xmath19 in fig .
[ deloc ] we present evolutions of the profile of the soliton density with @xmath136 and @xmath137 ( here @xmath138 corresponds to the case when the norm of the initial soliton coincides with the norm at the global minimum of the existence curve @xmath83 ) .
as one can see in the former case there is no delocalization transition and the solution remains localized , while in the latter case at some instant the norm of the solution becomes smaller then the critical norm @xmath83 and the norm of the solution any more pretends to the existence curve and therefore the solution decays into linear waves .
returning to the initial value @xmath139 in the first case the solution reconstructs its initial form while in the second case it remains delocalized .
now let us consider the opposite situation when the solitonic structure is supported by an attractive nonlocal interaction @xmath93 and the local interaction is repulsive , @xmath140 . in fig.[fig11 ] the corresponding initial profile of the nonlocal soliton in the parabolic trap with defect
as in the previous case the shift from the center of the trap is needed to observe oscillations of the solitons and eventual interaction with defect placed in the center of the trap .
two simple scenarios of the interaction of the soliton with the defect in the case of a strong parabolic trap are shown in fig.[q-05](a ) , ( c ) . in the first case
the incoming soliton has enough kinetic energy to go through the defect . in this interaction
the soliton lost part of the energy and therefore became locked by the defect in the semi - space @xmath141 .
for a stronger trap the soliton has enough kinetic energy to go through the barrier several times ( see fig.[q-05](b ) , ( d ) ) .
a similar picture of the scattering of the pure dipolar soliton can be observed by decreasing the strength of the local interaction @xmath16 to zero ( see fig.[q0 ] ) .
as one can see the pure dipolar soliton does not present very robust dynamics , what can be explained by the weak force of the nonlocal interaction .
in conclusion , we have investigated the dynamics of bright solitons in a dipolar condensate loaded into a parabolic trap with a barrier potential at the center . using a variational approach and scattering theory ,
we have studied the reflection , transmission and splitting of bright solitons on the defect in the presence of the trap potential .
we explore the different sets of parameters changing the relative strength of the local and nonlocal ( dipolar ) interactions .
the case when the local nonlinearity dominates is close to the settings investigated early @xcite and well understood now .
the intermediate cases when @xmath142 and the pure dipolar nonlinearity @xmath143 are considered here .
we show that coherent splitting of the dipolar bright soliton exists .
the splitting parts are reflected by the trap and recombine into a single soliton , performing a few oscillations under the trap potential .
the condition for soliton splitting is derived , which is in excellent agreement with numerical simulations of the full nonlocal gross - pitaevskii equation .
localization - delocalization phenomena governed by the variation of dipolar interactions in time has been studied for the cases @xmath144 and @xmath145 .
the reconstruction of the soliton to its initial form is observed for the former case , while in the latter case the soliton transforms into the linear waves .
the obtained results can be interesting for the study of soliton dynamics in the case of very small atomic scattering length when the dipolar interactions effects start to play a dominant role .
several experiments were performed recently in @xmath146li with a very small atomic scattering length tuned via the broad @xmath147 feshbach resonance with the number of atoms in the soliton of the order of @xmath148 @xcite .
the barrier potential was generated by a near - resonant cylindrically focused laser beam .
it can be useful for the design of matter wave beamsplitters and matter wave interferometers using the fusion of the solitons . the related problem of quantum scattering of solitons by the potential _
well _ for bec with local interactions has been studied in @xcite . for the case of the dipolar solitons
this problem will be considered separately .
fka was supported by the 7th european community framework programme under the grant piif - ga-2009 - 236099 ( nomatos ) .
99 pedri p and santos l 2005 phys . rev . lett . * 95 * , 200404 ; lashkin vm 2007 phys .
rev . a * 75 * , 043607 ; tikhonenkov i , malomed ba and vardi a 2008 phys .
a * 78 * , 043614
; tikhonenkov i , malomed ba and vardi a 2008 phys . rev . lett . * 100 * , 090501 .
|
the propagation of solitons in dipolar bec in a trap potential with a barrier potential is investigated .
the regimes of soliton transmission , reflection and splitting as a function of the ratio between the local and dipolar nonlocal interactions are analyzed analytically and numerically .
coherent splitting and fusion of the soliton by the defect is observed .
the conditions for fusion of splitted solitons are found .
in addition the delocalization transition governed by the strength of the nonlocal dipolar interaction is presented .
predicted phenomena can be useful for the design of a matter wave splitter and interferometers using matter wave solitons .
|
the next generation blast experiment @xcite ( blast - tng ) is a suborbital balloon payload that seeks to map polarized dust emission in the 250 @xmath0 m , 350 @xmath0 m and 500 @xmath0 m wavebands .
the instrument utilizes a stepped half - wave plate to reduce systematics .
the general requirement of the detectors is that they are photon - noise - limited and dual - polarization sensitive . to achieve this goal
, we are developing three monolithic arrays of cryogenic sensors , one for each waveband .
each array is feedhorn - coupled and each spatial pixel consists of two orthogonally spaced polarization - sensitive microwave kinetic inductance detectors @xcite ( mkids ) fabricated from a ti / tin multilayer film . in previous work
, we demonstrated photon - noise - limited sensitivity in 250 @xmath0 m waveband single polarization devices @xcite . in this work ,
we present the first results of dual - polarization sensitive mkids at 250 @xmath0 m .
our detection scheme utilizes feedhorn / waveguide , front - side optical coupling and places the inductive section of a lumped - element kinetic inductance detector @xcite one - quarter wavelength away from a reflective backshort .
this coupling approach is described in detail in hubmayr et al @xcite .
the devices are fabricated from tin / ti proximitized films , which have a tunable and spatially uniform tc @xcite .
the tin(ti ) thickness is 4(10 ) nm , which sets tc=1.35 k. by stacking a number of bilayers , we may tune the sheet resistance of the film without altering tc .
we chose to stack four bilayers and add a protective tin cap layer , which produces @xmath2 = 20 @xmath3 and @xmath4 = 22.5 ph/@xmath5 .
the tin / ti multilayer effectively reduces the sheet impedance by a factor of four as compared to the tin / ti / tin trilayer films used in our previous single - pol devices @xcite .
this allows us to reduce the absorber width by a factor of four in the new dual - polarization devices , which is critical to minimize their cross - polar coupling .
[ fig : noncrossoverlayout ] shows the photolithography mask design used to produce + dual - polarization - sensitive mkids from these films .
there are two mkids per spatial pixel , one per linear polarization .
each mkid contains a 5 @xmath0 m finger / gap interdigitated capacitor of total area 0.68 mm@xmath6 and a 3.2 @xmath0 m wide inductor that spans the length of the 180 @xmath0 m wide waveguide diameter for a total volume of 154 @xmath0m@xmath7 and 230 @xmath0m@xmath7 for the x and y inductors , respectively .
this combination of @xmath8 and @xmath9 produces resonance frequencies near @xmath10 @xmath11 1 ghz .
each mkid couples to a 340 @xmath0 m wide microstrip transmission line ( the silicon wafer is the dielectric and the device box is the ground plane ) via an interdigitated coupling finger of designed @xmath12 @xmath11 40000 - 50000 .
the two inductors within a pixel are orthogonally aligned in order to obtain dual - polarization - sensitivity . by making the y - polarization inductor discontinuous , both mkids are defined in one device layer without requiring electrical cross - overs . in electro - magnetic simulations of a simplified model of just the antennas ,
the inherent asymmetry of the design produces a band averaged ( 1 - 1.4 thz ) co - polar coupling of 79 ( 75)% in the continuous ( discontinuous ) absorbing inductor .
these simulations also suggest the expected cross - polar coupling to be @xmath12% , but if the vacuum gap between the wafer and the feedhorn / waveguide becomes too large , other structures in the detector design could begin to produce an additional cross - pol contribution .
five - pixel prototype arrays have been fabricated on a @xmath13 ( 2.5 mm ) detector pitch .
an array couples to a matching array of aluminum , direct - machined feedhorns .
the horns are a three step modified potter horn that has been designed for minimized beam asymmetries while achieving a 30% fractional bandwidth @xcite@xcite .
the 1.0 thz low edge of the band is defined by waveguide . a quasi - optical low - pass filter mounts in front of the feedhorns , which defines the 1.4 thz high edge of the passband @xcite .
we mount this detector package to the cold stage of an adiabatic demagnetization refrigerator and operate the array at 100 mk in the measurements described below .
m diameter waveguide which illuminates the inductors is depicted by the shadowed circular region .
_ right : _ a magnification of the inductive meanders which act as the polarization - sensitive absorbers .
the two 3.2 @xmath0 m thick inductors are non - intersecting with the y polarization detector ending 2 @xmath0 m before intersecting the x polarization detector .
( color figure online . ) ]
polarization characterization was performed on the prototype array described in section [ sec : det ] in a cryostat that is optically coupled to the room with appropriate quasi - optical filtering .
in addition , a 1.8 mm thick piece of eccosorb mf-110 microwave absorber was installed to decrease the optical loading on the detectors , ensuring their operability when viewing a 300 k thermal load .
the microwave absorber has an anti - reflective coating and has a calculated band - averaged transmission of 0.93% .
the detectors are coupled to a 1050 c to 20 c chopped thermal source that underfills the beam of the feedhorns .
a rotatable wire grid polarizer which has an induced cross - pol of less than 0.5% is placed between the chopped source and the cryostat window .
we determine the polarization properties by measuring the amplitude of the response of the detectors as a function of the angular position of the polarizer .
the result produces a sinusoidal signal , shown in fig .
[ fig : poleff ] .
we determine the cross - polar coupling , or the minimum of the amplitude response , by fitting the data to a sine wave .
the results of the fit suggests the detector cross - polar coupling is 2.6% and 2.8% for the x and y polarizations , respectively .
however , these values are most likely an upper limit as the fit lies above the minimum data points ( 1.7% and 1.9% for x and y , respectively ) .
regardless , this result is consistent within the uncertainty of the hfss simulations for this pixel geometry , and is an improvement over the 11% cross - polar coupling measured in the previous experiment@xcite . and @xmath14 for the x and y polarization detectors , respectively .
( color figure online . ) ] we determine the detector passbands by use of a fourier transform spectrometer ( fts ) , which was purged with nitrogen gas to minimize atmospheric attenuation .
the fts is coupled to the same 1050 c thermal source used in the polarization measurements . in this configuration ,
the source fills the beam and is no longer chopped , and the input polarizer is removed .
the bandpasses for each x and y polarization detector are averaged over multiple fts scans to reduce spectral noise , and the result is shown in fig .
[ fig : bandpass ] . while the cut - off frequency ( 1.4 thz ) , which is defined by the low - pass filter , is uniform , there is a clear difference in cut - on frequencies between the two detector polarizations .
the x polarization cut - on is 1054.1 ghz , while the y polarization cut - on is 1033.7 ghz .
this 20.4 ghz discrepancy in the low frequency edge is likely due to a slightly oval - shaped waveguide ( 3.55 microns larger in y - pol than x - pol ) , which is produced by using a standard twist drill bit . to address this problem ,
the waveguide will be undercut and reamed to produce a uniform circle at the desired waveguide diameter .
we determine the sensitivity to thermal radiation by coupling the detectors to a beam - filling temperature - controlled blackbody load .
the measurement approach is identical to that described in hubmayr et al @xcite .
while the detectors were held at a bath temperature of 100 mk , the frequency noise and detector response was measured as a function of blackbody temperature 3 k to 22 k. our previous trilayer films @xcite as well as other devices fabricated from tin @xcite show a near linear responsivity to photon load .
we observe the same phenomenon in these multilayer films as seen in fig .
[ fig : nep_dualpol ] .
the device responsivity ranges from -20 to -24 ppm / pw for a sample of mkids in the 5-pixel array .
the frequency noise of the detector taken at each temperature step is converted to a noise equivalent power ( nep ) by utilizing the local slope of the detector responsivities in fig .
[ fig : nep_dualpol ] .
the results are fitted using a best fit nep model which accounts for individual sources of noise .
more detail on this can be found in hubmayr et al @xcite .
this best fit model is used to determine a detector optical coupling efficiency of @xmath1175% .
these results also confirm that the multilayer films demonstrate photon noise limited sensitivity above 1 pw of loading .
the expected photon noise of the instrument for the 250@xmath0 m array is shown as the blue bar at 13.8 pw .
( color figure online . ) ]
we demonstrate dual - polarization sensitive thz pixels comprised of microwave kinetic inductance detectors . these detectors have good polarization isolation , demonstrating a cross - polar coupling of at most 2.6% and 2.8% in the x and y polarizations , respectively .
in addition , these detectors demonstrate background limited sensitivity above 1 pw of loading .
this work represents a viable path towards production of large arrays of thousands of polarization sensitive mkids for observations in the sub - mm which will be deployed on the blast - tng instrument .
99 b. dober , p.a.r .
ade , p. ashton , f.e .
angil , j.a .
beall , d. becker , k.j .
bradford , g. che , h .-
cho , m.j .
devlin , l.m .
fissel , y. fukui , n. galitzki , j. gao , c.e .
groppi , s. hillbrand , g.c .
hilton , j. hubmayr , k.d .
irwin , j. klein , j. van lanen , d. li , z .-
y li , n.p .
lourie , h. mani , p.g .
martin , p. mauskopf , f. nakamura , g. novak , d.p .
pappas , e. pascale , f.p .
santos , g. savini , d. scott , s. stanchfield , j.n .
ullom , m. underhill , m.r .
vissers , and d. ward - thompson , _ proc .
spie _ * 9153 * , 91530h , ( 2014 ) , doi:10.1117/12.2054419 .
n. galitzki , p.a.r .
ade , f.e .
angil , p. ashton , j.a .
beall , d. becker , k.j .
bradford , g. che , h .-
cho , m.j .
devlin , b.j .
dober , l.m .
fissel , y. fukui , j. gao , c.e .
groppi , s. hillbrand , g.c .
hilton , j. hubmayr , k.d .
irwin , j. klein , j. van lanen , d. li , z .- y .
lourie , h. mani , p.g .
martin , p. mauskopf , f. nakamura , g. novak , d.p .
pappas , e. pascale , g. pisano , f.p .
santos , g. savini , d. scott , s. stanchfield , c. tucker , j.n .
ullom , m. underhill , m.r .
vissers , and d. ward - thompson , _ j. astron
. instrum . _ * 03 * , 02 , 1440001 , ( 2014 ) , doi:10.1142/s2251171714400017 .
j. hubmayr , j. beall , d. becker , h .-
cho , m. devlin , b. dober , c. groppi , g. c. hilton , k.d .
irwin , d. li , p. mauskopf , d. p. pappas , j. van lanen , m.r .
vissers , y. wang , l.f .
wei , and j. gao , _ appl .
lett . _ * 106 * , 73505 , ( 2015 ) , doi:10.1063/1.4913418 .
e. pascale , p.a.r .
ade , f. angil , s. benton , m.j .
devlin , b. dober , l. fissel , y. fukui , n. gandilo , j. gundersen , p.c .
hargrave , j. klein , a.l .
korotkov , t.g .
matthews , l. moncelsi , t.k .
mroczkowski , c.b .
netterfield , g. novak , d. nutter , l. olmi , f. poidevin , g. savini , d. scott , j.a .
shariff , j.d .
soler , n.e .
thomas , m.d.p .
truch , c.e .
tucker , g.s .
tucker , and d. ward - thompson , _ proc .
spie _ * 8444 * , 844415 , ( 2012 ) , doi:10.1117/12.927211 .
|
polarization - sensitive , microwave kinetic inductance detectors ( mkids ) are under development for the next - generation blast instrument ( blast - tng ) .
blast - tng is a balloon - borne submillimeter polarimeter designed to study magnetic fields in diffuse dust regions and molecular clouds .
we present the design and performance of feedhorn - coupled , dual - polarization sensitive mkids fabricated from tin / ti multilayer films , which have been optimized for the 250 @xmath0 m band .
measurements show effective selection of linear polarization and good electrical isolation between the orthogonally crossed x and y detectors within a single spatial pixel .
the detector cross - polar coupling is @xmath13% .
passband measurements are presented , which demonstrate that the desired band - edges ( 1.0 - 1.4 thz ) have been achieved .
we find a near linear response to the optical load from a blackbody source , which has been observed in previous devices fabricated from tin .
blackbody - coupled noise measurements demonstrate that the sensitivity of the detectors is limited by photon - noise when the optical load is greater than 1 pw .
|
symmetric informationally - complete positive operator - valued measures ( sics ) @xcite represent a general form of measurement in quantum theory . as the more familiar projective measurements are associated to an orthonormal basis in hilbert space , a sic is associated to an over - complete set of @xmath0 unit vectors in @xmath1 such that the absolute value of the scalar product between any distinct two is always constant , i.e. @xmath2 @xmath3 sics have practical applications in quantum state tomography @xcite , quantum communication @xcite , quantum cryptography @xcite , classical high precision radar @xcite and classical speech recognition @xcite .
a significant amount of work is aimed towards proving their existence in all dimensions although no general proof is currently known .
numerical studies have successfully found sics when @xmath4 @xcite and analytical solutions have been published for 20 of these dimensions ( see references in @xcite plus recent solutions for @xmath5 in @xcite and @xmath6 in @xcite ) .
such results promote the belief that sics can always be found , but the current solutions are rather dimension - dependent and do not provide an overall coherent picture .
all known sics are group covariant , meaning they can be obtained from the action of a group on a single fiducial vector .
the vast majority of sics are covariant with respect to the weyl - heisenberg ( wh ) group and if the dimension is a prime , it has been shown this is the only possible group @xcite . in this paper
we only consider such sics .
there are a few curious properties of sics that are not yet understood , but clearly hint at an underlying mathematical structure
. this scenario could support suggestions that sics should play a deeper role in quantum mechanics @xcite .
the most well - known of these properties is the conjecture of zauner symmetry , which states that every sic vector is an eigenvector of a certain order three unitary matrix @xcite .
another property stems from current sic solutions ; in a preferred basis , the components of the sic vectors are expressed as nested radicals .
such complex numbers are quite exceptional and arise because the galois group of the polynomial equation one must solve is solvable @xcite .
however , it is a third observation that we shall focus on in this paper , namely the connection between sics and algebraic curves in dimension 3 .
the wh group appears naturally in the theory of elliptic curves @xcite , suggesting this might be an advantageous framework for the study of sics .
furthermore , there is a very close connection between elliptic curves and linear dependencies among sic vectors in dimension 3 , first noted by lane hughston @xcite .
this leads to the hesse configuration in the complex projective plane a particular structure well - known in algebraic geometry .
we aim to explore this special connection , looking for analogous linear dependencies in higher dimensions .
it was our hope that another configuration in complex projective space might reveal some new structure behind sics and broaden the areas from which they can be studied .
this did not turn out quite as planned for all the sics we looked at but we do see similarities for certain sics in dimension 8 , which could signal another connection to elliptic curves .
no sets of @xmath7 linearly dependent sic vectors exist for @xmath8 , 5 or 7 , but we do find them again when @xmath9 , 9 and 12however , in these latter cases the linear dependencies are not uniquely tied to sics .
additionally , we found an unexpected method of generating 2- and 3-dimensional sics from the linear dependencies among vectors in 6- and 9-dimensional sics , respectively
. we do nt know whether this can be generalised to all dimensions divisible by 3 .
we observe that the reverse of our problem to show that there exists an open set of fiducials giving rise to wh orbits without linear dependencies has been considered for signal processing purposes and has been solved in prime dimensions @xcite and more recently in all finite dimension @xcite .
in section 2 we review some basic facts about the weyl - heisenberg and clifford groups , and explain notations and terminology used in the rest of the paper .
section 3 introduces the hesse configuration and linear dependencies in sics in dimension 3 .
section 4 proves that linear dependencies can always be found in wh orbits when the fiducial vector is an eigenvector of an order 3 unitary matrix . in dimensions divisible by 3 and dimensions equal to 8 mod 9
, these wh orbits include sics .
in section 5 , we give numerical results from our computer search for dependencies in dimensions 4 to 9 .
we focus specifically on data relating to the sic case in dimensions 6 and 9 . in section 6 , we prove the sets of linearly dependent vectors in dimension 6 contain 2-dimensional sics and show that sets in dimension 9 contain 3-dimensional sics .
we do not have data for dimension 12 or higher , so do not know whether this pattern continues for all dimensions divisible by 3 .
in which case , this could open a new avenue for a sic existence proof .
section 7 turns away from sics altogether and proves that linear dependencies will always arise in wh orbits with a fiducial vector invariant under a unitary matrix whose order is a non - trivial divisor of @xmath7 in dimensions where @xmath7 is a product of distinct prime numbers .
we give an example of this using linear dependencies in even dimensions .
section 8 relates our results to a recent representation of the clifford group @xcite .
section 9 summarises our results and points at some possible clues for the sic existence problem .
in this section we review some basic facts about the weyl - heisenberg ( wh ) and clifford groups which will be needed in the sequel . for more details see ( for example ) @xcite .
let @xmath10 be the standard basis in dimension @xmath7 .
define the operators @xmath11 and @xmath12 by @xmath13 where addition of ket - labels is mod @xmath7 and @xmath14 .
the weyl - heisenberg displacement operators are then defined by @xmath15 where @xmath16 , @xmath17 . with this definition @xmath18 where the symplectic form @xmath19 is defined by @xmath20 it is convenient to introduce the notation @xmath21 we then have @xmath22 and @xmath23 (
note that if @xmath7 is even @xmath24 and @xmath25 implies @xmath26 ) .
we define the weyl - heisenberg group to consist of all operators of the form @xmath27 , with @xmath28 , @xmath29 ( where @xmath30 is the set @xmath31 equipped with arithmetic modulo @xmath32 ) . in many situations
the phase factor in the first of equations ( [ eq : dprd ] ) is not important .
it is accordingly convenient to introduce the notation @xmath33 to signify that @xmath34 and @xmath35 are equal up to a phase .
we then have @xmath36 for all @xmath37 , @xmath38 . neglecting phases in this way
the wh group reduces to the @xmath0 operators @xmath39 ( in more technical language these facts can be expressed by the statement that the collineation group is @xmath40 ) .
similarly , if we ignore phases then the orbit of a vector @xmath41 under the wh group reduces to the set @xmath42 ( in more technical language this set is the orbit of @xmath41 considered as an element of @xmath43 under the action of the collineation group ) . in the sequel , by an abuse of terminology
, we shall refer to this set as the wh orbit of @xmath41 .
the symplectic group @xmath44 consists of all matrices @xmath45 such that @xmath46 , @xmath47 , @xmath48 , @xmath49 and @xmath50 ( mod @xmath32 ) . to each such matrix
there corresponds a unitary @xmath51 , unique up to a phase , such that @xmath52 if @xmath47 is relatively prime to @xmath32 we have the explicit formula @xmath53 where @xmath54 is the multiplicative inverse of @xmath47 ( mod @xmath32 ) and @xmath55 is an arbitrary phase .
if @xmath47 is not relatively prime to @xmath32 we use the fact @xcite that @xmath56 has the decomposition @xmath57 where @xmath58 , @xmath59 both are relatively prime to @xmath32 so that @xmath60 , @xmath61 can be calculated using equation ( [ eq : ufdef ] ) .
@xmath51 is then given by @xmath62 up to an arbitrary phase .
we shall refer to unitaries of the form @xmath63 as symplectic unitaries .
the clifford group then consists of all products @xmath64 .
we have @xmath65 for all @xmath66 .
as mentioned in the introduction it turns out that every known wh sic fiducial vector is an eigenvector of a canonical order @xmath67 clifford unitary and , conversely , that in every dimension where an exhaustive search has been made every canonical order @xmath67 clifford unitary has a sic fiducial vector as one of its eigenvectors @xcite . in these statements
the term `` canonical order @xmath67 clifford unitary '' refers to a unitary of the form @xmath68 where @xmath37 is arbitrary and @xmath69 is any element of @xmath44 with the property @xmath70 if @xmath71 one needs to impose the additional requirement that @xmath72 .
it is then guaranteed that , with a suitable choice of the phase @xmath55 , @xmath73 is order @xmath67 .
the matrix @xmath69 is always conjugate @xcite , either to the zauner matrix @xmath74 or its square or , in the case of dimensions @xmath75 with @xmath76 , to the matrix @xmath77 in this paper we will confine ourselves to the case of fiducials which are eigenvectors of @xmath78 .
choosing the phase in equation ( [ eq : ufdef ] ) to be @xmath79 one finds @xcite that the eigenvalues of @xmath78 are @xmath80 , where @xmath81 .
we denote the three corresponding eigenspaces as @xmath82 , and @xmath83 , and give their dimensions in table 1 .
.multiplicities of the eigenvalues of @xmath78 for different dimensions . [ cols="^,^,^,^",options="header " , ] next we define three new operators @xmath84 and @xmath85 as follows .
@xmath86 note that @xmath87 and @xmath88 have the same linear span .
we will prove that @xmath89 implies @xmath90 and @xmath91 , and that @xmath92 implies @xmath93 and @xmath94 .
from that one can prove that the span is 2-dimensional .
the following properties of @xmath84 and @xmath85 can be derived straightforwardly from their definitions : @xmath95 @xmath96 moreover , @xmath97 and @xmath98 are rank-3 projection operators , and they are orthogonal to each other .
one can also verify the following commutation relations between @xmath99 and @xmath100 @xmath101 these equations tell us how @xmath102 and @xmath85 permute the eigenspaces of @xmath100 . for example , if @xmath103 is an eigenvector of @xmath100 with eigenvalue @xmath104 , then @xmath105 , so @xmath106 is an eigenvector of @xmath100 with eigenvalue 1 .
the full action of @xmath107 and @xmath85 on the eigenspaces of @xmath100 is described in figure 2 ( @xmath108 commutes with @xmath100 so it simply leaves the eigenspaces invariant ) .
( solid arrow ) and @xmath85 ( dotted arrow ) on the eigenspaces of @xmath109 .
@xmath110 ( 2d ) and @xmath111 ( 1d ) span @xmath112 ( 3d ) .
@xmath113 ( 1d ) and @xmath114 ( 1d ) span @xmath115 ( 2d ) .
@xmath116 ( 1d ) is @xmath83 . ]
let @xmath117 , @xmath118 , @xmath119 , @xmath120 and @xmath121 be non - zero eigenvectors of @xmath109 belonging to the eigenspaces @xmath122 , @xmath114 , @xmath116 , @xmath111 and @xmath113 respectively .
except for @xmath117 , the rest of them are unique up to a scalar , because the corresponding eigenspaces are 1-dimensional .
as seen from the diagram , we have @xmath123 .
we are going to prove that @xmath124 , so that @xmath90 for any @xmath89 .
suppose otherwise , that @xmath125 .
it has to be the case that @xmath126 ( for any choice of @xmath117 ) , because otherwise @xmath127 will be a non - zero vector in @xmath113 , which is 1-dimensional , which implies @xmath128 for some non - zero @xmath46 , and therefore @xmath129 , contradicting the assumption that @xmath125 .
since @xmath130 is a non - zero vector in @xmath116 , which is 1-dimensional , we must also have @xmath131 for some non - zero @xmath47 , and therefore @xmath132 so from the assumption that @xmath125 , we have obtained @xmath133 , which implies @xmath134 , which means that @xmath135 is orthogonal to the 4-dimensional subspace spanned by @xmath110 , @xmath114 and @xmath116 .
this contradicts @xmath98 being a rank-3 projection operator .
thus , we conclude that @xmath136 , and that @xmath90 for any @xmath89 .
the identity @xmath91 immediately follows from @xmath137 .
note that @xmath138 ( because @xmath135 is orthogonal to @xmath139 ) is a non - zero vector in @xmath140 .
@xmath141 is not proportional to @xmath103 because @xmath141 lies in @xmath112 and @xmath103 lies in @xmath115 .
therefore @xmath103 , @xmath142 , @xmath143 and @xmath141 indeed fully span a 2-dimensional subspace .
it s a similar argument to show that @xmath144 , which implies @xmath145 and @xmath146 for any @xmath92 .
we can thus explain the occurrence of the @xmath147-dimensional sics one finds when @xmath9 .
unfortunately , one ca nt explain the @xmath67-dimensional sics one finds when @xmath148 in the same way .
if one uses the construction described in the proof of theorem 4 when @xmath148 one obtains 9 vectors which do not belong to a @xmath67-dimensional subspace and for which the overlaps are not constant ( they actually take two different values ) .
the construction also fails to produce sics in dimensions @xmath149 and @xmath150 .
we summarise the situation in table 5 , which shows the dimensionality of the subspace spanned by the @xmath0 vectors obtained using the construction of theorem 4 as a function of the eigenspace from which @xmath41 is chosen .
nevertheless , the fact remains that if , in dimension @xmath151 , instead of arbitrarily choosing a vector in @xmath115 or @xmath83 as in theorem 4 , one starts from an arbitrarily chosen fiducial vector @xmath41 in the zauner subspace ( not necessarily a sic fiducial ) , then one always does find @xmath67-dimensional sics among the normal vectors to the linearly dependent sets generated by @xmath41 .
if this phenomenon repeated in higher dimensions it would open up the intriguing possibility that one might be able to get a proof of sic existence in this way .
c | c c c c & @xmath9 & @xmath148 & @xmath152 & @xmath153 + + @xmath140 & 4 & 8 & 12 & 15 + @xmath115 & 2 & 7 & 8 & 15 + @xmath83 & 2 & 6 & 8 & 10 +
we have been focussing so far on the zauner unitary because , along with its conjugates , it is the symplectic unitary which appears to be relevant to the sic existence problem .
however , there are many other symplectic unitaries which give rise to linear dependencies , as we now show . before proceeding
further we need to take account of a subtlety which is important in even dimensions .
suppose @xmath7 is even , so that @xmath154 .
given @xmath155 let @xmath156 be @xmath69 reduced mod @xmath7 .
we define 1 . the @xmath32-order of @xmath69 to be its order considered as an element of @xmath44 .
2 . the @xmath7-order of @xmath69 to be the order of @xmath157 considered as an element of @xmath158 .
this distinction was not important in the case of the zauner matrix since its @xmath32-order is the same as its @xmath7-order .
however , the two orders are , in general , different . turning to @xmath159
, we say that @xmath37 , @xmath160 are @xmath7-distinct if @xmath161 mod @xmath7 .
so as to be able to make statements which apply to both even and odd dimensions we will use the terms `` @xmath7-order '' and `` @xmath7-distinct '' in odd dimensions also , defining them in the obvious way .
let @xmath162 ( @xmath7 even or odd ) .
we say that @xmath69 generates linear dependencies in all eigenspaces if , for each eigenvector @xmath41 of @xmath163 , there exist @xmath164 which are @xmath7-distinct and such that the vectors @xmath165 are linearly dependent . continue to suppose @xmath166 , and let @xmath167 be the @xmath7-order of @xmath69 .
we say that @xmath37 is @xmath69-full if the points @xmath168 are @xmath7-distinct .
we then have [ thm : genlindep ] let @xmath166 and let @xmath167 be the @xmath7-order of @xmath69 .
suppose 1 .
2 . @xmath7 is a multiple of @xmath167 .
[ trcond ] @xmath170 .
[ fullcond ] there exist @xmath7 points in @xmath171 which are @xmath7-distinct and @xmath69-full then @xmath69 generates linear dependencies ( in the sense defined above ) in all eigenspaces .
we first observe that @xmath172 has at least one eigenspace with dimension @xmath173 .
in fact , let @xmath174 be the spectral decomposition of @xmath172 , where @xmath175 and @xmath55 is a phase . by assumption
@xmath176 it follows that the numbers @xmath177 can not all be the same .
consequently we must have @xmath178 for at least one value of @xmath179 .
it follows from condition [ fullcond ] that it is possible to choose @xmath69-full points @xmath180 such that the @xmath7 points @xmath181 are @xmath7-distinct .
let @xmath41 be any eigenstate of @xmath172 and define @xmath182 for all @xmath179 @xmath183 where @xmath184 are the eigenspaces @xmath185 with the appropriate ordering ( note that @xmath186 is nt necessarily the eigenspace corresponding to the projector @xmath187 defined above ) .
we showed that @xmath188 for some @xmath179 .
for that value of @xmath179 the vectors @xmath189 are linearly dependent , implying that the vectors @xmath190 are linearly dependent . as an illustration of this theorem
consider the parity matrix @xmath191 we have @xmath192 and @xmath193 ( respectively @xmath55 ) if @xmath7 is even ( respectively odd )
. also the number of @xmath69-full points in @xmath194 is @xmath195 ( respectively @xmath196 ) if @xmath7 is even ( respectively odd ) .
so the theorem shows that @xmath197 generates linear dependencies in every eigenspace for all even dimensions .
it will be seen that the conditions of the theorem are not very restrictive .
there are @xmath0 points in @xmath171 so one would expect that the requirement that @xmath7 of them be @xmath69-full should not be difficult to satisfy . similarly with the requirement that @xmath170 . for a large class of dimensions we can make this statement sharp : [ thm : lindepgen ]
let @xmath198 for @xmath199 distinct odd prime numbers @xmath200 .
then , with the sole exception of the identity matrix , every matrix @xmath201 whose order is a factor of @xmath7 generates linear dependencies ( in the sense defined above ) in every eigenspace .
the proof of this theorem will occupy us for the rest of this section .
it depends on the tensor product representation of the clifford group which is described in appendix b in @xcite , and which we now briefly review . for each @xmath202 and @xmath203 in the range @xmath204
let @xmath205 be the unique element of @xmath206 such that @xmath207 mod @xmath208 .
then the chinese remainder theorem @xcite tells us that the map @xmath209 is an isomorphism of the ring @xmath210 onto the ring @xmath211 with inverse @xmath212 where @xmath213 is the multiplicative inverse of @xmath214 considered as an element of @xmath206 . given @xmath215 then define @xmath216 the map @xmath217 ( respectively @xmath218 ) is an isomorphism of modules ( respectively groups ) .
we have @xmath219 for all @xmath69 , @xmath37 .
let @xmath220 be @xmath167-dimensional hilbert space and let @xmath221 be the unitary defined by @xmath222 where @xmath223 is the standard basis in dimension @xmath167 .
then @xmath224 modulo the phase ambiguity in the definitions of @xmath225 .
[ lm : ffull ] let @xmath198 for @xmath199 distinct odd prime numbers @xmath200 .
let @xmath226 be arbitrary .
then the number of @xmath69-full points in @xmath171 is @xmath227 .
let @xmath228 be the number of @xmath69-full points in @xmath229 and @xmath230 the number of @xmath231-full points in @xmath232 .
use the isomorphisms defined in equations ( [ eq : phi ] ) and ( [ eq : psi ] ) to identify @xmath69 with @xmath233 and @xmath37 with @xmath234 .
let @xmath167 be the order of @xmath69 and @xmath235 the order of @xmath231 .
then @xmath236 suppose @xmath237 is @xmath231-full in @xmath232 for each @xmath203 .
then @xmath238 implies @xmath239 is a multiple of @xmath235 for each @xmath203 , which in turn implies @xmath239 is a multiple of @xmath167 .
so @xmath37 is @xmath69-full in @xmath240 .
so @xmath241 .
the problem thus reduces to showing that @xmath242 for all @xmath203 . to see this we use the analysis in @xcite
let @xmath243 and let @xmath244 ( respectively @xmath245 ) be the set of quadratic residues ( respectively non - residues ) in @xmath206 . if @xmath246 then @xmath231 is conjugate to a diagonal matrix @xmath247 if @xmath237 is non - zero then @xmath248 if and only if @xmath239 is a multiple of @xmath235 .
so @xmath249 . if @xmath250 we can diagonalize @xmath231 over the extension field @xmath251 .
so we again have that @xmath231 is conjugate to a matrix of the form @xmath247 where @xmath46 now belongs to @xmath251 instead of @xmath206 .
we again deduce that @xmath252 .
finally , if @xmath253 we have that @xmath231 is conjugate to the matrix @xmath254 for some @xmath47 . if @xmath255 then @xmath252 ; otherwise @xmath256 .
[ lm : espce ] let @xmath198 for @xmath199 distinct odd prime numbers @xmath200 .
then @xmath257 for all @xmath258 . in view of equation ( [ eq : symutensor ] )
it is enough to show that @xmath259 for all @xmath203 . to see this
let @xmath260 we have @xcite @xmath261 where @xmath262 and @xmath55 is a phase .
so @xmath263 where @xmath264 and @xmath265 .
performing the gauss sums @xcite we find @xmath266 where @xmath267 where @xmath268 is the legendre symbol , and where we have used the multiplicative property of the legendre symbol together with the fact that @xmath269 for all non - zero @xmath270 .
we see that in every case @xmath271 immediate consequence of theorem [ thm : genlindep ] and lemmas [ lm : ffull ] , [ lm : espce ] . it would be interesting to see if one can generalise theorem [ thm : lindepgen ] to the case of even dimensions , and odd dimensions for which the prime factorization contains primes occurring with multiplicity @xmath272 .
in the motivating three - dimensional example there was an interesting interplay between linear dependencies in wh orbits on the one hand , and preferred orthonormal bases on the other .
it is not difficult to find fiducial vectors giving rise to a wh orbit with @xmath7 linearly dependent sets of @xmath7 vectors each , in such a way that their normal vectors form an orthonormal basis : if the generator @xmath12 is diagonal , any fiducial vector with a zero entry will give rise to such a linearly dependent set under repeated action with @xmath12 .
acting with @xmath11 on the fiducial will permute its entries , and one ends up with @xmath7 linearly dependent sets in the orbit .
the situation is more interesting if the dimension is a square number , @xmath273 . for this case
it has recently been shown that there exists a choice of basis in which the entire clifford group is represented by monomial matrices , i.e. by permutation matrices in which the unit entries have been replaced by phase factors @xcite .
this representation stems from the observation that in square dimensions @xmath274 = 0\ ] ] thus @xmath275 and @xmath276 can be simultaneously diagonalised , and the resulting basis turns out to have the above property .
denote the basis vectors by @xmath277 , where @xmath278 are integers modulo @xmath167 , and let us refer to the basis as the monomial basis .
the monomial representation of the clifford group is given by @xmath279 @xmath280 where @xmath281 .
@xmath282 where @xmath283 note that equation ( [ therep ] ) assumes @xmath47 is relatively prime to @xmath32 , otherwise one has to use the decomposition given in equation ( [ eq : primedecomp ] ) of section 2 . evidently the @xmath7 orthogonal rays that make up the basis is in itself an orbit of the full clifford group .
moreover , one sees by inspection that the vector @xmath284 is left invariant by the full symplectic group whenever @xmath7 is odd . in this sense
the monomial basis is a distant cousin of the ` special sic ' in three dimensions or perhaps of the mub in three dimensions .
now suppose that @xmath285 , and consider the zauner matrix @xmath286 if we scrutinise the explicit expression ( [ therep ] ) we find that there are three non - zero diagonal elements in the unitary operator @xmath287 .
if we adjust the phase @xmath55 according to zauner s recipe one of them equals @xmath288 , and the other two equal @xmath289 .
it follows that any vector in the @xmath290 eigenspace must have one component equal to 0 .
since the wh group contains @xmath7 diagonal elements ( namely the elements of the subgroup generated by @xmath275 and @xmath276 ) this means that the wh orbit generated from a fiducial vector in @xmath290 contains @xmath7 vectors confined to a subspace orthogonal to one of the vectors in the monomial basis .
since the entire wh group is given by monomial matrices the orbit can be divided into @xmath7 such linearly dependent sets , spanning @xmath7 subspaces each orthogonal to a vector in the monomial basis . in particular this
applies to the known sics in these dimensions .
a similar argument applies if @xmath291 .
we consider the symplectic matrix @xmath292 after adjusting the arbitrary phase so that @xmath293 and so that the largest eigenspace corresponds to the eigenvalue 1 , we find that @xmath294 has two diagonal non - zero elements , one equal to 1 and the other equal to @xmath295 .
following the same logic as above we see that every wh orbit that includes a fiducial invariant under @xmath294 must be divided into @xmath7 linearly dependent sets of @xmath7 vectors orthogonal to one of the vectors in the monomial basis
. variants of this argument can be used also if the dimension is @xmath296 , where @xmath179 is any integer . in this case
the hilbert space splits into a direct sum of @xmath297 copies of a @xmath179-dimensional hilbert space , and there exists a representation of the clifford group in which every group element acts within and permutes these subspaces @xcite .
this representation is @xmath179-nomial , in the sense that it is given by a permutation matrix of size @xmath297 in which the unit entries have been replaced by matrices of size @xmath179 .
for instance , one finds in dimension eight that any vector in the smallest eigenspace of the zauner unitary has two zero entries , and again there will be linear dependencies in a wh orbit arising from a fiducial in this eigenspace .
a sic fiducial does in fact exist in this eigenspace @xcite .
the hesse configuration singles out ` special sics ' in dimension 3 by the higher number of linear dependencies among their vectors .
we searched for similar patterns in higher dimensions through exhaustive numerical searches for sets of @xmath7 linearly dependent vectors in sics in dimensions @xmath8 to 8 and partial searches in dimensions @xmath148 and 12 . in dimension 8
, the sics with fiducials in the smallest subspace of the zauner unitary , @xmath83 , also stand out for exhibiting a higher number of linear dependencies than other wh orbits with a fiducial vector from the same subspace .
this could extend the dimension 3 connection between elliptic curves and sics ( which gives rise to the hesse configuration ) to dimension 8 and , as sic fiducials always lie in @xmath83 for dimensions equal to 8 mod 9 , possibly to an infinite family of dimensions . in the dimensions we looked at , sets of @xmath7 sic vectors
are usually linearly independent .
we proved that it is always possible to find dependencies in sics for dimensions divisible by 3 ( where the sic fiducials sit in @xmath112 ) and for dimensions equal to 8 mod 9 ( where the sic fiducials sit in @xmath83 ) .
however , this dependency structure in dimensions divisible by 3 also occurs for any wh orbit when the fiducial vector is taken from @xmath112 and so is not a property uniquely tied to sics .
we also proved that linear dependencies will arise for wh orbits whose fiducial vector is in @xmath112 for dimensions @xmath298 ; wh orbits whose fiducial vector is in @xmath115 for dimensions @xmath298 and @xmath299 ; and wh orbits whose fiducial vector is in @xmath83 for all dimensions .
these theorems do not account for all the linear dependencies ; our numerical results often revealed a higher number of linearly dependent sets of vectors than the theorems predict .
looking in detail at linear dependencies in dimension 6 revealed some surprising structure .
firstly , the normal vectors ( i.e. the vectors orthogonal to each 5-dimensional subspace spanned by 6 linearly dependent vectors ) collect into orthogonal sets of 4 when the hw orbit is a sic .
when the orbit is not a sic , these normal vectors lose their orthogonality pattern .
secondly , the normal vectors of dependencies from both sics and non - sics form 30 2-dimensional sics .
this repeats in dimension 9 , where some of the normal vectors form 3-dimensional sics .
we did not find 4-dimensional sics from normal vectors in dimension 12 , but our search was not exhaustive and we can not rule out their existence either .
we also proved a number of theorems for linear dependencies using unitaries other than the zauner unitary .
our results represent the beginning of a theory for when linear dependencies occur in wh orbits .
this is of interest in itself .
also , just possibly , they may contain an important clue for the sic existence problem .
ib and kb thank markus grassl for taking an interest in this problem when we visited singapore , and especially for drawing our attention to the linear dependencies in the `` odd '' sic in eight dimensions .
we also thank a referee for constructive comments .
hbd was supported by the natural sciences and engineering research council of canada and by the u.s .
office of naval research ( grant no . n00014 - 09 - 1 - 0247 ) .
ib was supported by the swedish research council under contract vr 621 - 2010 - 4060 .
dma was supported in part by the u.s .
office of naval research ( grant no .
n00014 - 09 - 1 - 0247 ) and by the john templeton foundation .
research at perimeter institute is supported by the government of canada through industry canada and by the province of ontario through the ministry of research & innovation .
99 g. zauner , _ quantendesigns .
grundzge einer nichtkommutativen designtheorie _ phd thesis , univ . wien ( 1999 ) ; english translation _ quantum designs : foundations of a non - commutative design theory _ int . j. quant
* 9 * 445 ( 2011 ) j. m. renes , r. blume - kohout , a. j. scott , and c. m. caves , _ symmetric informationally complete quantum measurements _ journal of mathematical physics * 45 * 2171 ( 2004 ) a. j. scott , _ tight informationally complete quantum measurements _ j. phys .
a * 39 * 13507 ( 2006 ) h. zhu and b .-
englert , _ quantum state tomography with fully symmetric measurements and product measurements _ phys . rev . a * 84 * 022327 ( 2011 ) c. a. fuchs and m. sasaki ,
_ squeezing quantum information through a classical channel : measuring the quantumness of a set of quantum states
_ quantum inf .
* 3 * 377 ( 2003 ) o. oreshkov , j. calsamiglia , r. muoz - tapia and e. bagan , _ optimal signal states for quantum detectors _ new j. phys .
* 13 * 073032 ( 2011 ) b .- g .
englert , d. kaszlikowski , h. k. ng , w. k. chua , j. ehek and j. anders , _ efficient and robust quantum key distribution with minimal state tomography _ arxiv:0412075 ( 2004 ) t. durt , c. kurtsiefer , a. lamas - linares and a. ling , _ wigner tomography of two - qubit states and quantum cryptography _ phys .
rev . a * 78 * 042338 ( 2008 ) s. d. howard , a. r. calderbank and w. moran , _ the finite heisenberg - weyl groups in radar and communications _ eurasip j. appl .
sig . process .
* 2006 * 85865 ( 2006 ) m. a. hermann and t. strohmer , _ high - resolution radar via compressed sensing _
ieee trans .
on sig . process . * 57 * 2275 ( 2009 ) r. balan , b. g. bodmann , p. g. casazza and d. edidin , _ painless reconstruction from magnitudes of frame coefficients _
j. fourier anal .
appl . * 15 * 488 ( 2009 ) a. j. scott and m. grassl , _ sics : a new computer study _ j. math
. phys . * 51 * 042203 ( 2010 ) d.m .
appleby , i. bengtsson , s. brierley , m. grassl , d. gross and j .- .
larsson , _ the monomial representations of the clifford group _ quantum inf .
* 12 * 0404 ( 2012 ) d. m. appleby , i. bengtsson , s. brierley , .
ericsson , m. grassl and j .- .
larsson , _ systems of imprimitivity for the clifford group
_ arxiv:1210.1055 ( 2012 ) h. zhu , _ sic - povms and clifford groups in prime dimensions _ j. phys .
a : math . theor . * 43 * 305305 ( 2010 ) c. a. fuchs and r. schack , _ quantum - bayesian coherence _
arxiv:0906.2187v1 ( 2009 ) d. m. appleby , _ sic - povms and the extended clifford group _ j. math .
phys . * 46 * 052107 ( 2005 ) d. m. appleby , h. yadsan - appleby and g. zauner , _ galois automorphisms of a symmetric measurement _
arxiv:1209.1813 ( 2012 ) d. mumford , _ tata lectures on theta _ , vols .
i , ii , and iii , birkhaser ( 1983 , 1984 , 1991 ) l. hughston , _
d=3 sic - povms and elliptic curves _
, perimeter institute , seminar talk , october 2007 ; available online at http://pirsa.org/07100040/ g. e. pfander , gabor frames in finite dimensions , in p. g. casazza and g. kutyniok ( editors ) _ finite frames : theory and applications _ , birkhser boston , 2013 j. lawrence , g. e. pfander and d. walnut , _ linear independence of gabor systems in finite dimensional vector spaces _
j. fourier anal . and appl .
* 11 * 715 ( 2005 ) r .- d .
malikiosis , _ a note on gabor frames in finite dimensions _ arxiv:1304.7709 ( 2013 ) s. flammia , _ on sic - povms in prime dimensions _ , j. phys .
a , * 39 * , 13483 ( 2006 ) o. hesse , _ ber die wendepuncte der curven dritter ordnung _ j. reine angew . math .
* 28 * 97 ( 1844 ) m. artebani and i. dolgachev , _ the hesse pencil of plane cubic curves _
lenseignement mathmatique * 55 * 235 ( 2009 ) w. k. wootters , _ a wigner - function formulation of finite - state quantum mechanics _ ann . phys .
ny * 176 * 1 ( 1987 ) i. bengtsson , _ from sics and mubs to eddington _
. ser . * 254 * 012007 ( 2010 ) w. o. alltop , _ complex sequences with low periodic correlations _ ieee transactions on information theory * 26 * 350 ( 1980 ) k. blanchfield , _ mutually unbiased bases , weyl - heisenberg orbits and the distance between them _ aip conf . proc .
* 1508 * 359 ( 2012 ) m. grassl , private communication h. e. rose , _ a course in number theory _
oxford university press , oxford , 1994 d. m. appleby , _ properties of the extended clifford group with applications to sic - povms and mubs _ arxiv:0909.5233 ( 2009 )
|
five years ago , lane hughston showed that some of the symmetric informationally complete positive operator valued measures ( sics ) in dimension 3 coincide with the hesse configuration ( a structure well known to algebraic geometers , which arises from the torsion points of a certain elliptic curve ) .
this connection with elliptic curves is signalled by the presence of linear dependencies among the sic vectors .
here we look for analogous connections between sics and algebraic geometry by performing computer searches for linear dependencies in higher dimensional sics .
we prove that linear dependencies will always emerge in weyl - heisenberg orbits when the fiducial vector lies in a certain subspace of an order 3 unitary matrix .
this includes sics when the dimension is divisible by 3 or equal to 8 mod 9 .
we examine the linear dependencies in dimension 6 in detail and show that smaller dimensional sics are contained within this structure , potentially impacting the sic existence problem .
we extend our results to look for linear dependencies in orbits when the fiducial vector lies in an eigenspace of other elements of the clifford group that are not order 3 .
finally , we align our work with recent studies on representations of the clifford group .
|
one - dimensional systems of interacting particles or spins show interesting collective effects when the system approaches an ordered state as the temperature approaches zero @xcite .
the static properties of such systems are dominated by a diverging correlation length , and generally well understood
. however , dynamic properties are more varied and intricate .
of particular interest is the way in which domains of ordered phases grow when the system is quenched from a disordered state to an ordered one .
the coarsening dynamics that ensues can often be modelled through the kinetics of domain walls ; a well - known example is the glauber - ising chain , in which domain walls diffuse and annihilate upon contact , corresponding to the kinetics of the reaction @xmath11 0 @xcite . in this paper
we study a system of arrowheads on a continuous 1d line , as depicted in fig .
[ fig : arrow_heads ] .
arrowheads resemble bent core or banana - shaped molecules , assemblies of which are known to form ordered phases in higher dimensions @xcite .
our objective in studying the simpler one - dimensional problem is to understand , qualitatively and quantitatively , the elements that go into the formation of large stacks of similarly oriented arrowheads .
these elements turn out to derive from a set of interlinked themes : entropy - driven ordering ; spin models with assymetric pairwise interaction ; domain wall kinetics with alternating diffusing and stationary walls ( @xmath12 ) . as summarized in the discussion below
, this sequence leads to novel effects in dynamics , both for a finite number of walls , as well as for a finite wall density in the thermodynamic limit .
each zero - area arrowhead points right ( @xmath13 ) or left ( @xmath14 ) , and stochastically attempts to change its location ( via diffusion ) or orientation ( by flipping ) , respecting a no - overlap constraint all the while , as discussed in section ii . since only hard - core interactions operate ( tantamount to the no - overlap constraint ) it is evident that purely entropic effects must be responsible for driving the order . in equilibrium
, this is brought out explicitly by tracing over locations of arrowheads , thereby generating an effective interaction between successive arrowheads , involving their orientations ( section iii ) .
this technique has successfully been used in the past to study 1d assemblies of interacting particles @xcite . in our case ,
interestingly , this interaction is not symmetric under the interchange of the orientations of a near neighbour pair .
a transfer matrix calculation then allows the equation of state and correlation functions to be calculated .
our model is closely related to the ` chiral ' ising models of @xcite , as discussed in more detail in section ii .
when the degree of alignment is large , there are long stretches of aligned arrowheads , with successive stretches being separated by domain walls .
a significant outcome of the non - symmetric interactions is that there are two species of domain walls . in the case of primary interest to us , one species ( @xmath1 ) diffuses , while the other ( @xmath2 ) is immobile , with reaction kinetics @xmath12 .
this has a crucial bearing on the dynamics .
this is already evident with a finite number of walkers ; as shown analytically in section iv , the full survival probability decays exponentially in time , in strong contrast to the power law decays that characterize @xmath4 @xcite , pertinent to relaxation in the glauber - ising model .
moreover , with a thermodynamically large number of walkers , numerical simulations reveal interesting effects ( section v ) .
further , during the approach to steady state ( section vi ) , the density of domain walls falls as @xmath15 as in the ising case .
persistence properties show greater variation .
the persistent fraction of arrowheads is found to decay as a power law @xmath16 , with @xmath8 close to @xmath9 , in contrast to the exact value @xmath17 for the ising case @xcite .
the global persistence decays with the same power @xmath9 , a result which can be justified through theoretical arguments , as in the ising case @xcite . finally ,
on allowing the @xmath2 walls to diffuse , albeit less slowly than the @xmath1 s ( section vii ) , a degree of nonuniversality is revealed : the exponent @xmath8 is found to depend continuously on the ratio of diffusion constants , even though the fraction of domain walls continues to decay as @xmath15 .
this is reminiscent of the results of @xcite , where a continuous variation of exponents was found as a function of coupling constants , though in different quantities .
our model consists of @xmath18 arrowheads , each of length @xmath19 on a 1d line of length @xmath20 , with a hard core constraint which implies no overlap ( fig .
[ fig : arrow_heads ] ) .
while @xmath20 is taken to be of fixed value with periodic boundary conditions while discussing dynamics , it proves expedient to allow @xmath20 to fluctuate in a constant pressure ensemble in order to derive static properties ; in that case , the number density @xmath21 and other thermodynamic quantities are well - defined in the thermodynamic limit . a microscopic configuration @xmath22 is specified by the set @xmath23 where @xmath24 is the coordinate of the vertex of the @xmath25 arrowhead while @xmath26 represents the orientation , with the positive sign corresponding to @xmath13 and the negative sign to @xmath14 .
the full dynamics involves attempts to make either a spin flip or displacement of a randomly chosen arrowhead @xmath27 , as follows + ( i ) it can attempt to flip about its vertex @xmath28 + ( ii ) it can attempt a displacement @xmath29 where @xmath29 is chosen with uniform probability in the interval @xmath30 $ ] where @xmath31 is a fixed length . + the attempted moves in ( i ) and ( ii ) are accepted only if they do not lead to an overlap with other arrowheads , and do not lead to crossing arrowheads in case ( ii ) . as discussed in section iii , these moves respect detailed balance and hence lead the system to an equilibrium state .
this state is characterized by a correlation length which diverges as @xmath32 , approaching an ordered state with like orientations ( @xmath33 or @xmath34 ) .
since both arrowhead flip and displacement attempts are involved in the dynamics , the time - dependent properties of the system would be expected to depend on parameters which govern these , both for autocorrelation functions in steady state , and for the dynamics of approach to steady state .
our preliminary numerical studies indeed indicate that at finite densities @xmath35 , the dynamic power laws do depend on the displacement range @xmath31 . in this paper , however , we are primarily interested in the approach to the fully ordered state which is reached only in the limit @xmath32 . accordingly , we will retain only the re - orientation move ( i ) ; the displacement move ( ii ) is ineffective in this limit .
hence the allowed dynamical moves for the central member in triplet of successive arrowheads are : @xmath36 note that the triplets @xmath37 , @xmath38 , @xmath39 , @xmath40 can not evolve in the limit considered , owing to the no - overlap constraint .
these rules have an important implication for interfaces which separate ordered segments of similarly - oriented arrowheads . evidently , the interfaces are of two types : ( @xmath1 ) ... @xmath41 .... and ( @xmath2 ) .... @xmath42 ... . while @xmath1 interfaces can evolve ( and move in the process ) , @xmath2 interfaces are static .
this distinction is ultimately responsible for the difference of behaviour in the coarsening dynamics vis a vis the glauber - ising model , where both types of interface evolve and diffuse at equal rates . in order to investigate the effects of allowing @xmath2 interfaces to diffuse , though more slowly than @xmath1 interfaces , in section vi we allow configurations @xmath39 and @xmath40 to evolve @xmath43 with @xmath44 .
evidently , by varying the ratio @xmath45 between 0 and 1 , we generate a family of models which interpolates between arrowhead model and the ising model .
the dynamical evolution rules are closely related to those considered by kim et al @xcite .
these authors studied nonequilibrium ising models with dynamics which they termed ` chiral ' , namely with different transition rates at ( @xmath46 ) and ( @xmath47 ) kinks .
besides nonconserving single spin flips ( analogous to arrowhead reorientations ) , they allowed asep - like moves which conserve spin ; these have an important effect on the dynamics , as discussed in subsequent sections .
in this section , we consider the equilibrium properties of the arrowhead model introduced in section ii and shown in fig .
[ fig : arrow_heads ] .
evidently the hard - core constraint between arrowheads plays a crucial role in determining the set of allowed microscopic configurations , i.e. the possible arrowhead vertex coordinates and `` spin '' orientations \{@xmath48 } which specify each allowed configuration @xmath22 .
we first observe that the condition of detailed balance @xmath49 is valid provided prob(@xmath22 ) is chosen to be equal for every allowed configuration @xmath22 .
this is because an allowed arrowhead flip @xmath50 @xmath51 @xmath52 occurs at the same rate as the reverse move , and this is true also of every displacement move @xmath24 @xmath51 @xmath24 + @xmath31 which leads to an allowed configuration , and its reverse . since every allowed configuration has equal energy , correlations between arrowheads develop purely from entropy .
it is well known that in systems with hard core interactions , entropy can lead to a tendency towards ordering @xcite .
indeed , as we will see below , this tendency is present in our system as well , and leads ultimately to a diverging correlation length with increasing arrowhead density . as for several 1d systems with hard objects on a line @xcite
, our system of arrowheads can be solved exactly by integrating over the coordinates \{@xmath24 } , thereby generating an effective interaction between nearest neighbour spins .
the important point in our case is that this interaction is not symmetric under the interchange of spins in a near neighbour pair of ising spins . as discussed in section ii
, our model is closely related to the chiral ising models discussed in @xcite which have a similar asymmetry ( though no handedness distinction ) .
the equilibrium properties of the resulting spin system can be obtained using a transfer matrix technique , resulting in closed form expressions for the equation of state and correlation functions .
we also study the extent of spatial persistence in this system .
it is convenient to embody the hard - core constraint between arrowheads by introducing an orientation - dependent potential energy of interaction @xmath53 we work in a constant pressure ensemble , where the pressure @xmath54 and temperature @xmath55 are specified .
the coordinate @xmath56 of the first particle is held fixed ( though @xmath57 can flip ) , while all other @xmath24 can fluctuate , implying that the total length of the system @xmath58 can fluctuate as well .
the corresponding partition function is then @xmath59 \end{split}\ ] ] for a specified set of spin orientations @xmath50 it is straightforward to perform the integrals over @xmath24 sequentially over @xmath27 with the result @xmath60 where @xmath61 ; @xmath62 . in terms of the transfer operator @xmath63 with matrix elements @xmath64 ,
we may write @xmath65 which may readily be evaluated by diagonalizing @xmath66 .
let us define @xmath67 .
then the eigenvalues @xmath68 are given by ( @xmath69 ) , with corresponding right eigenvectors @xmath70 and left eigenvectors @xmath71 .
the ( unnormalized ) entries of @xmath72 are ( 1,@xmath73 ) while those of @xmath71 are ( @xmath74 ) .
we find @xmath75\ ] ] in the limit of large @xmath18 , we obtain @xmath76 recalling that the average system length @xmath77 = @xmath78 we may find the number density @xmath79 as a function of @xmath80 and @xmath54 , yielding the equation of state @xmath81 the contribution to the equation of state coming from the arrowhead configurations is embodied in the second term on the right hand side .
this term is a correction to the ideal gas contribution @xmath82 ; it arises from the hard core interaction between arrowheads .
interestingly , it is small at both high and low values of @xmath83 .
when @xmath83 is small , arrowheads are well separated and their orientation is unimportant , so that they approximate a free ideal gas . on the other hand ,
when @xmath83 is large , the large entropic cost of the pair sequence ( @xmath47 ) makes its occurrence exponentially unlikely .
the vertex locations of the remaining sequences of arrowheads are isomorphic to those of an ideal gas of point particles . the rare occurrence of ( @xmath47 )
, implies that ( @xmath46 ) is equally rare , as these pair sequences must alternate .
together , this implies that the system correlation length must become very large as @xmath84 .
this is verified by direct calculation as discussed below .
the two - point correlation function can be evaluated using the transfer matrix formalism . with free boundary conditions , we have @xmath85 where @xmath86 is the @xmath87 pauli matrix .
@xmath88 can be evaluated through a standard route .
assuming that @xmath27 and @xmath89 are both of order @xmath18 , we obtain @xmath90 = @xmath91 in the thermodynamic limit @xmath92 @xmath93 .
thus the correlation length @xmath94 is given by @xmath95 .
in the dense packing limit @xmath96 , we obtain @xmath97 .
it is also interesting to ask for the spatial persistence , namely the probability that the same value of the spin , say @xmath98 , occurs unbroken over a stretch of @xmath99 sites @xcite . in terms of the projection operator
@xmath100 = @xmath101 with eigenvalues @xmath100=0 or 1 this probability is given by @xmath102 = @xmath103 where the thermodynamic limit @xmath104 has been assumed .
given the form of the operator @xmath105 , the matrix element on the right hand side is found to be @xmath106 , independent of @xmath99 .
thus @xmath107 decays exponentially , as @xmath108 where @xmath109 $ ] . in the dense packing limit @xmath96
, we obtain @xmath110 .
thus in the limit of high density , both the correlation length @xmath94 and the persistence length @xmath111 diverge in similar ways , with prefactors differing by a factor 2 . recall that these ` lengths ' pertain to spin separations which correspond to arrowhead labels ,
so they need to be divided by the density in order to convert them to lengths on the line .
finally we investigate the effect of interaction asymmetry on the spatial separation of arrowheads .
suppose the orientations of the @xmath25 and @xmath112 arrowheads have been specified to be @xmath52 and @xmath113 , and their mean separation is @xmath114 . for a given sequence @xmath115 of intermediate spins , @xmath116 is the sum of separations of successive pairs @xmath117 .
from eq.([eq : potential ] ) it is easy to see that @xmath118 while @xmath119 .
thus @xmath120 here @xmath121 is the mean number of ( @xmath47 ) pairs in the stretch ( @xmath122 ) ; it can be calculated using the transfer matrix within the finite stretch , keeping track of each occurrence of ( @xmath47 ) .
the result is @xmath123 the mean numbers of @xmath14 @xmath13 pairs differ by exactly 1 for the two specifications of @xmath124 , implying that @xmath125 exceeds @xmath126 by @xmath127 .
this is a quantitative measure of the effects of asymmetry in the interaction .
in the one - dimensional model being considered any possible configuration is a sequence of domains each with all arrows in one direction ( @xmath14 or @xmath13 ) bounded by domain walls @xmath13@xmath14 or @xmath14 @xmath13 .
the dynamic moves are flips of arrows about their vertex with no overlap allowed between arrowheads .
allowed arrowhead flips like @xmath13@xmath13@xmath14 going to @xmath13@xmath14@xmath14 can generate hopping of the domain walls @xmath13 @xmath14 ( hereafter called `` walkers '' or mobile @xmath1 - particles in this section ) . here
we take the limit @xmath35 @xmath51 @xmath93 ( corresponding to @xmath84 in the spin model ) which makes any ( isolated ) one of the other type ( `` walls '' or @xmath2 - particles ) immobile .
further , since @xmath14@xmath13@xmath14 can go to @xmath14@xmath14@xmath14 ( etc ) we have a process which is a special form of @xmath129 in which a `` walker '' pair - annihilates with an otherwise fixed `` wall '' .
the no - overlap constraint prevents the annihilation of pairs of like particles .
furthermore , the two types of domain wall / particle , @xmath1 and @xmath2 , alternate , and continue to do so even after any allowed pair annihilation .
but the surviving @xmath2 s are fixed .
it will be seen that the general system dynamics is qualitatively different from that of @xmath130 . in this section
we investigate the survival probability for the general case of @xmath131 `` walls '' with @xmath128 intervening `` walkers '' .
the specific aim is to obtain the probability @xmath132 ( \{@xmath133 } , \{@xmath134 } ; \{@xmath135 } ) of survival to time @xmath5 of all @xmath128 walkers @xmath136 to @xmath128 , ( with initial positions @xmath134 ) to positions @xmath133 , and of all walls @xmath137 to @xmath131 ( fixed at @xmath135 ) .
this can be treated by an image method of the type used for vicious walkers and related systems @xcite despite the very different dynamics resulting .
figure [ fig : walkers ] ( a ) gives the simplest illustration of the image method , for a single walker . in the figure
the thick black line can represent a particular realization of the path of a walker shown by the black circle ( at its position @xmath138 at time @xmath5 , increasing to the right and downwards , respectively ) .
effects of an added fixed reflecting or absorbing wall , shown by the vertical straight line , can be represented by an image walker , the open circle , together with its path , a reflection ( in the wall ) of the original path .
this scheme is appropriate for walls of the type @xmath139 in the system being studied , which pair annihilate with the walker . for a single random walker with no wall
the `` free '' probability function ( over all possible paths starting at @xmath140 at @xmath141 ) is @xmath142\ ] ] ( taking diffusion constant unity ) .
the corresponding probability for an image resulting from a wall at @xmath143 in the manner just described , is @xmath144 @xmath145= @xmath146 .
the operation @xmath144 corresponds to a sort of reflection in the wall at @xmath147 .
then ( @xmath148)@xmath149 vanishes at @xmath147 and is , up to that point , the appropriate probability for obtaining survival properties for the system of the single walker with pair - annihilating wall at @xmath147 .
similarly for a single walker between two such walls at @xmath150 , @xmath151 an image method provides the survival probability distribution function @xmath152 ) .
however here the character of the procedure ( after the first stages ) and the consequences are very different from the trivial case just described , and representative of those for the general case . to satisfy the necessary boundary conditions ( vanishing of @xmath153 at the walls ) repeated application of the operations ( @xmath154 ) , ( @xmath155 ) to the free distribution function @xmath156 is required .
this gives rise to a proliferation of images whose positions are periodically related at any time , including @xmath141 .
figure [ fig : walkers ] ( b ) shows , for a particular realisation , the paths , and positions for a specific time @xmath5 , of the walker and of its images produced in this way .
here the black circle is the walker , and the dashed and open circles are respectively its positive and negative images ( having contributions to @xmath153 of opposite signs , coming from the negative sign in the factors ( @xmath154),(@xmath155 ) .
consequently the survival probability for the @xmath157 case being considered is @xmath158 .
\end{split}\ ] ] any distribution @xmath159 of initial positions
@xmath160 , which generalises the free distribution function @xmath156 to @xmath161 gets similarly proliferated , and superposition allows us to include that in what follows , by replacing @xmath149 in eq .
( [ eq : q12_1 ] ) by @xmath162 .
our discussion so far has not made apparent a crucially important property of the survival probability for the case being considered , namely its exponential time decay .
this property , shared with generalisations ( below ) to unlimited numbers of walkers , is a consequence of the walk s confinement .
the exponential decay is most easily quantified using the fourier representation of @xmath153 .
the fourier transform of @xmath163 is @xmath164 where @xmath165 .
it is easy to show that the non - vanishing fourier components of @xmath153 have discrete wave - vectors @xmath166 where @xmath167 is any non - zero integer ( corresponding to the period of the image structure , allowing for its alternation ) . the resulting form for @xmath153 is @xmath168 \end{split}\ ] ] so at long times @xmath169 @xmath170 @xmath171 @xmath172 the survival probability decays exponentially @xmath173 with decay time @xmath174 .
this @xmath172 applies also for the generalised case having distribution @xmath159 of initial positions , but here the replacement in ( [ eq : q12_1 ] ) of @xmath149 by @xmath162 takes ( [ eq : q12_2 ] ) to a generalised form involving the fourier transform of @xmath175(a ) .
the generalisation of the above development to @xmath131 fixed periodically located absorbing walls @xmath14 @xmath13 and @xmath128 intervening walkers @xmath13@xmath14 can be treated in a similar fashion .
this is true for both walkers with specific initial positions and for those with distributions of initial positions . for simplicity
we give just the development for the first , simpler , case which generalises trivially .
we treat @xmath131 walls @xmath14 @xmath13 ( denoted by @xmath176 to @xmath131 ) stuck at sites @xmath177 and @xmath128 intervening walkers @xmath13@xmath14 , position variables @xmath133 , @xmath136 to @xmath128 , initially at sites @xmath134 = @xmath178+@xmath179 which are distant @xmath179 and @xmath180 from the walls on their left and right respectively .
it can be verified that the image procedure used above to treat the case @xmath157 again closes for this case of general @xmath128 .
however , here each walker @xmath181 has its own image system arising from the two adjacent fixed walls on either side of it .
each such system is just like that for the single - walker - between - two walls case just discussed , and in particular is periodic with the same periodicity .
so , up to the time of the first walker annihilation , the probability of survival to time @xmath5 of all @xmath128 walkers to positions @xmath133 , and of all @xmath131 walls , is the following product of factors for each walker : @xmath182 here either of the two equivalent forms for @xmath153 ( given in ( [ eq : q12_1 ] ) and ( [ eq : q12_2 ] ) ) can be used .
it is easily verified ( using either form ) that @xmath132 vanishes for any @xmath133 equal to @xmath178 or @xmath183 ( walker at wall ) . until the first such event it provides the proper value for the survival probability distribution since all walkers are then between their " two walls .
the corresponding survival - everywhere " probability @xmath184(\{@xmath134};\{@xmath135 } ) is obtained from ( [ eq : qnn_1 ] ) by integrating each @xmath133 over all possible values ( between walls ) .
the late - time dependence of both types of survival probability for general @xmath128 is easily obtained from the product of the forms ( [ eq : q12_2 ] ) for the @xmath157 case , which involves the same discreteness of wave vectors , related by @xmath185 = @xmath186 to the wall spacing .
this again gives exponential decay at late times , the same for the two probabilities , but now for general @xmath128 the product of @xmath128 factors makes the decay time @xmath187 / n\ ] ] concerning early times , ( [ eq : qnn_1 ] ) shows using ( [ eq : q12_1 ] ) that then each survival probability is dominated by the contribution from the free - walk term in ( [ eq : q12_1 ] ) unless any walker starts near a wall , in which case the corresponding image term in ( [ eq : q12_1 ] ) gives an appreciable addition to the free - walk term after a time of order @xmath188 or @xmath189 .
the dominant early - time term , needed below , for the survival - everywhere probability for @xmath128 walkers starting midway between their walls is @xmath190^n\ ] ] among other quantities described in this paper , the analytical predictions for the survival - everywhere probability can be compared with numerical results from simulation .
simulations were carried out using an appropriate algorithm for the no - arrow - overlap dynamics , ie with mobile walkers and immobile walls , corresponding to the process @xmath12 , and also for the process @xmath4 . after the generation of appropriate initial states simulations were run up to times long enough to exceed expected ( and actual ) characteristic times and finite cut - offs from finite size .
averaging over many histories ( @xmath191 ) was exploited . for both types of process
a wide range of values of @xmath128 and of wall separations were used . for the first case , with no - overlap constraint , each mobile walker
was initially taken to be midway between its fixed confining walls @xmath192 .
the results for this case exhibited in figures [ fig : surv_a+b_tau_vs_n ] and [ fig : surv_a+b_tau_vs_a ] are all for the survival - everywhere probability of all @xmath131 walls and @xmath128 walkers .
the figures [ fig : surv_a+b_tau_vs_n ] , [ fig : surv_a+b_tau_vs_a ] both give clear evidence of exponential decay after a time roughly comparable to that , @xmath193 , predicted by the theory . : variation with @xmath128 .
the inset shows @xmath172 versus @xmath128 .
, width=294,height=219 ] the wide range of @xmath128 values covered by the results in figure [ fig : surv_a+b_tau_vs_n ] allow an accurate estimate of the @xmath128-dependence of the decay time .
the inset showing log @xmath172 versus log @xmath128 gives exponent 0.94 .
similarly in figure [ fig : surv_a+b_tau_vs_a ] the wide range of @xmath160 values used in the log - log plot in the inset allows the exponent estimate 2.22 .
the exponents appear to be converging towards the theoretical values ( 1 and 2 ) as the simulation runs get longer . :
variation with initial spacing @xmath160 .
the inset shows @xmath172 versus @xmath160.,width=294,height=219 ] figure [ fig : surv_a+b_tau_vs_a ] also exhibits , particularly for the largest @xmath160 s , the type of early - time behaviour theoretically predicted in ( [ eq : survival_prob_exp ] ) .
the late exponential decay in figures [ fig : surv_a+b_tau_vs_n ] and [ fig : surv_a+b_tau_vs_a ] is in stark contrast to the behaviour in figure [ fig : surv_a+a ] for the @xmath4 case . .
the dashed lines indicate the power law decays predicted in ref .
@xcite , width=294,height=219 ] here the results are consistent at large times with power - law decay , @xmath194 where @xmath195 as predicted by the vicious walkers " theory @xcite . as
just discussed , it has been found that for finite @xmath128 in the @xmath196 process ( i ) the probability that all walkers survive up to @xmath5 falls exponentially , ie much faster than ( the power law ) for the @xmath197 ( process ii ) .
however for persistence ( see v - b ) the fraction of walkers surviving falls slower for process i than for process ii .
such observations raise general questions about possible effects , including crossovers , of particular conditions considered , and of such things as finite number versus finite density of walkers , etc . a particular condition which affects the decay rate of the all - walker - survival prob @xmath198 ... is the initial gap distribution .
with all gaps starting at the same value , @xmath143 , eq([eq : survival_prob_exp ] ) gives @xmath199 decaying exponentially with rate @xmath200 .
for the same initial condition in process ii the corresponding decay starts similarly but the greater possibility of gaps increasing allows a crossover to later power - law decay @xcite .
for persistence the late time regime of interest is where a large fraction of the walkers have been annihilated , the @xmath143 s have become large , and the rare survivors are of interest .
all this occurs well after the crossover just referred to , which allows a less drastic diminution of the number of surviving walls in ii than in i. but simple qualitative arguments do not easily explain the faster decay of unflipped arrowheads in ii than i. different initial gap distributions can lead to different or no crossovers .
for example distributions giving weight to very small gaps allow much smaller characteristic and crossover times , and most initial distributions give rise to long - time distributions which are invariant except for a common inflation of gaps with time ( @xmath201 ) .
in this section , we discuss the dynamics of approach to equilibrium in large systems . every arrowhead configuration has alternating @xmath1 and @xmath2 domain walls and the approach to steady state involves the decrease of their number under annihilation kinetics @xmath12 .
we study the case in which @xmath1 walls diffuse while @xmath2 walls are stationary , as appropriate to the limit @xmath32 ; the effects of @xmath2 wall diffusion will be the subject of section vi .
we monitor domain wall densities and persistence properties of arrowheads as the system evolves towards equilibrium .
we use two types of initial conditions : a ) random : with random placement of arrowheads , leading to random locations of domain walls .
b ) periodic : with equally - long alternating stretches of ... @xmath38 ... and ... @xmath202 ... , implying a periodic arrangement of domain walls . in either case , @xmath1 and @xmath2 domain walls alternate in sequence .
as we will see , there are some similarities to , and some marked differences from , glauber - ising systems , whose dynamics is governed by the single - species annihilation process @xmath4 . evidently , arrowhead kinematics leads to alternating @xmath1 and @xmath2 walls in all configurations , including initial conditions .
this is very different from the situation in several studies of @xmath12 where @xmath1 and @xmath2 particles are placed at random ( @xcite ) . in the latter case ,
concentration fluctuations decay very slowly and dominate the late time dynamics . our study is closer to that of @xcite as will be discussed in section vi .
numerical results are obtained using monte carlo simulations where system size is typically @xmath20 = 10000 and averaging is done over few hundreds of histories .
as time passes , diffusing @xmath1 walls annihilate with stationary @xmath2 s . in section iv with a finite number of walls
, we saw that the all - walker survival probability decays exponentially in time , with a decay time which varies inversely with the number of walkers @xmath128 .
this time is vanishingly small in the present context , where the number of walls is macroscopic .
we focus , to start with , on the wall density and how it decays in time .
since the underlying dynamics of the mobile species is purely diffusive , we expect that the result would be qualitatively similar to that for @xmath4 for the glauber - ising case .
this is borne out by our numerical results which show the decay of wall density for the two systems , with initial conditions a ) and b ) for each case .
the density follows a @xmath203 decay with @xmath204 in all cases .
[ fig : domain_wall_decay ] this result agrees with that of @xcite where it was found that @xmath205 when the rate of the asep - like moves is set equal to zero .
persistence quantifies how much of particular properties of the initial configuration survive without change up to time @xmath5 @xcite .
these properties could either be local ( for example the orientation of individual arrowheads ) or global ( for instance the majority orientation of arrowheads ) . in spin models ,
it is customary to monitor the fraction @xmath206 of persistent spins , i.e. those which have not flipped up to time @xmath5 .
equivalently , @xmath206 can be thought of as the probability that a given spin has not flipped .
the persistence probability follows a power law decay , @xmath207 .
a closed form expression has been obtained for @xmath8 for the @xmath208-state potts models on a one - dimensional lattice ; for the ising model , the result is @xmath209 .
our data for the persistent fraction @xmath206 in the arrowhead problem are shown in fig .
[ fig : persist ] along - side the data for the ising model .
it is evident that the decay exponent is substantially different ; we find the value for @xmath8 is 0.245 , intriguingly close to @xmath210 .
note that the value of @xmath8 is independent of the initial condition a ) or b ) . the global persistence @xmath211 is defined as the fraction of histories in which a global variable has not changed sign .
an appropriate global variable is the ` order parameter ' @xmath212 where @xmath213 and @xmath214 are respectively the number of right and left pointing arrowheads . from our numerical results ( fig .
[ fig : global_persist ] we find @xmath215 where the value of @xmath216 is consistent with @xmath9 which agrees with the value for @xmath216 in the ising case . for the ising model ,
the result @xmath217 was derived in @xcite by observing that the scaled order parameter ( analogous to @xmath218 ) obeys random walk dynamics provided we redefine the time variable to be @xmath219 .
the only input required for this argument is that the density of surviving walkers falls as @xmath6 .
since this is true in our case as well , the result @xmath217 holds here as well .
since individual entities generally decay faster than global variables , we expect that the value of @xmath8 should be greater than or equal to @xmath216 .
thus we surmise that the value of the site persistence exponent is also @xmath9 . in all other systems we are aware of ,
@xmath220 ; this is the first instance where the equality @xmath221 seems to hold .
in order to investigate the universality of the results obtained in section v , we have studied the effects of letting the @xmath2 walls diffuse , by allowing the moves of eq .
( [ eq : moves_2 ] ) in addition to those of eq .
( [ eq : moves_1 ] ) . in the original arrowhead problem , movement of a @xmath2 wall would require a large enough gap to open up for it to occur without violating the hard - core constraint ; in the equilibrium state , this would be extremely rare if the density is large .
thus the moves considered ( eq.([eq : moves_2 ] ) ) are best viewed as put in ` by hand ' ; they would need a softening of the hard - core constraint between arrowheads at @xmath2 interfaces in order to happen .
our motivation in studying such moves is purely to investigate the theoretical question of universality with respect to allowing @xmath2 wall diffusion , partly motivated by the importance of the question for the @xmath12 problem , and partly by the observations of @xcite where the authors observed a continuous variation of decay exponents as some rates were changed in their model .
we first examined numerically the manner in which the density of walls decays .
we found that it follows @xmath222 where the value of @xmath29 appears to remain fixed at @xmath223 , as for the arrowhead and ising cases .
the data is not displayed , but falls in between the two limiting cases shown in fig.[fig : domain_wall_decay ] .
this is in contrast to the variation of @xmath29 observed in @xcite when the rate of the asep - like move in their model was varied .
coming to the question of persistence , namely the probability that a given spin has not flipped up to time @xmath5 , we note that the question may be posed in the context of a finite number of walls , as considered in section iv .
in fact , the simplest case is most illuminating : three walkers , with a diffusion constant @xmath224 for the outer two walkers , and @xmath225 for the central walker .
the survival problem can be solved exactly by a mapping to the motion of a composite particle in a wedge - shaped domain @xcite .
the result is @xmath226 the significant point is that @xmath8 depends explicitly on the ratio @xmath227 .
it is instructive to check limiting cases .
if the central particle is a stationary @xmath2 particle surrounded by two @xmath1 s , we obtain @xmath8 = 1 , implying a faster decay than with a single @xmath1 particle ( @xmath228 ) . on the other hand , if the central particle is an @xmath1 surrounded by two @xmath2 particles , we obtain @xmath229 , consistent with the exponential decay found in section iv . to see whether this dependence of the power law exponent on diffusion constants stays also for persistence properties in a thermodynamically large system , we numerically studied the fraction of persistent spins @xmath206 as a function of @xmath5 , and found a definite variation of the power law exponent @xmath8 as the ratio of diffusion constants @xmath45 is varied .
the data fig .
[ fig : persist_var ] indicate a smooth variation of @xmath8 from about 0.25 in the arrowhead model @xmath230 to 0.375 in the ising model @xmath231 .
however , the global persistence @xmath211 continues to show a @xmath232 decay as in the arrowhead and ising cases .
this is not unexpected , as following the argument in section v , the decay of @xmath211 is related to the manner in which the density of domain walls decays , and as we have seen above , this does not depend on @xmath45 .
in this paper we have studied a system of stochastically reorienting arrowheads in one dimension .
our study of the static properties and the dynamics of approach to the steady state has touched on several issues , and we discuss our results in that context .
the equilibrium state of our system approaches an orientationally ordered state as the density increases .
the origin of order is entropic , as follows from the fact that arrowheads may never overlap , implying that every allowed configuration is equally likely . for a fixed total length ,
it is evident that the entropy of translation is largest in an orientationally ordered state , as then there is no constraint on the locations of arrowhead vertices other than that they must maintain a sequence .
this contribution to the entropy dominates at large density ; at moderate values of @xmath35 , it must compete with the configurational entropy of the locations of interfaces .
the spatial extent of the order is quantified by the correlation length and the persistence length , and the calculations in section iii show that both diverge as @xmath233 as @xmath32 .
understanding the dynamics of approach to the ordered state is greatly facilitated by the observation that there are two types , @xmath1 ( @xmath234 ) and @xmath2(@xmath235 ) , of arrowhead interfaces obeying @xmath12 dynamics .
our study has a bearing on a couple of interconnected issues in this two - species annihilation problem . in this problem
, the influence of initial conditions on long - time decays has long been recognized . initial conditions in which the sequence of @xmath1 s and @xmath2 s is random give rise to long - lived concentration fluctuations in the number of @xmath1 and @xmath2 particles , which in turn give rise to multiple length scales and slow down the dynamics strongly @xcite .
by contrast , under initial conditions in which the imbalance between @xmath1 and @xmath2 particle numbers remains of the order of unity in every stretch , the two species are well mixed , and the decay exponent @xmath205 .
recent work @xcite has shown that when alternation follows the pattern @xmath236 there are multiplicative logarithms for even @xmath167 .
such initial conditions have been studied in the context of conserved lattice gas models , on identifying pairs of sites with @xmath1 and @xmath2 particles @xcite . in the arrowhead model , even a random placement of arrowheads results in strict alternation of @xmath1 and @xmath2 interfaces ; thus our result @xmath205 ( section v ) with no evidence of logarithms is fully consistent with the above .
another important issue concerns the fact that in our problem , the diffusion constants for @xmath1 and @xmath2 particles are not equal ; in fact the @xmath2 particles do not diffuse at all in the limit of infinite arrowhead density .
how pertinent is this for long time decays of survival probabilities under @xmath12 dynamics ? with a finite number @xmath128 of walkers , we showed that the all - walker survival probability falls exponentially rapidly in time , with the decay time being inversely proportional to @xmath128 .
this contrasts strongly with the power law decay with power @xmath237 found with equal diffusion constants , as appropriate to @xmath4 . in the thermodynamically large system
, we found the persistence probability falls as a power law for both @xmath12 and @xmath4 , but importantly , the powers differ , being close to @xmath9 in the former case , and @xmath17 in the latter .
we conclude with some comments on universality .
the studies of @xcite on a generalized model with asep - like moves together with annihilation indicate a violation of universality , for example in the power laws for domain wall density , as the asep rates are changed . in the reorienting arrowhead model
, there is no asep move .
we find that on varying the diffusion constant for @xmath2 particles from 0 to the value for @xmath1 particles , globally averaged properties such as the number of domain walls , or global persistence , remain universal .
but more delicate properties such as single site persistence are found to exhibit a continuous variation of power laws as the ratio of diffusion constants is varied .
this indicates a violation of universality , but at a weaker level than in the studies of @xcite .
mdk acknowledges institute of mathematical sciences ( imsc ) , chennai , india for selection in associateship programme for college teachers .
mb acknowledges the hospitality of the rudolf peierls centre for theoretical physics at the university of oxford where part of this work was done and the award of the j c bose fellowship by the department of science and technology , india .
he thanks satya majumdar for a useful discussion . @xmath238 ed .
v. privman ( cambridge university press , cambridge , 1997 ) p. l. krapivsky , s. redner and e. ben - naim , @xmath239 ( cambridge university press , cambridge , 2010 ) t. niori , t. sekine , j. watanabe , t. furikawa and h. takezoe , j. mater .
chem 6 , 1231 ( 1996 ) ; mol .
301 , 337 ( 1997 ) y. lansac , p. k. maiti , n. a. clark and m. a. glaser , phys .
e 67 , 011703 ( 2003 ) j. a. martinez - gonzalez , s. varga , p. gurin , and j. quintana - h , europhysics letters , 97 ( 2012 ) 26004 .
m. takahashi in @xmath240 edited by e. h. lieb and d. c. mattis ( academic press , new york , 1966 ) y. kantor and m. kardar , phys .
e 79 , 041109 ( 2009 ) m. kim , s - c park and j. d. noh , phys .
e 87 , 012129 ( 2013 ) m. kim , s - c park and j. d. noh , phys .
e 91 , 012132 ( 2015 ) m. e. fisher , j. stat .
, 34 , 667 ( 1984 ) b. derrida , a. j. bray and c. godrche , j. phys . a 27 , l357 ( 1994 ) b. derrida , v. hakim , and v. pasquier , phys .
75 , 751 ( 1995 ) , j. stat .
85 , 763 , ( 1996 ) a. j. bray , s. n. majumdar , and g. schehr , adv .
62 , 225 ( 2013 ) s. n. majumdar , a. j. bray , s. j. cornell and c. sire , phys .
77 , 3704 ( 1996 ) d. frenkel , physica a 263 , 26 ( 1999 ) j. l. lebowitz , j. k. percus and j. talbot , j. stat .
physics , 49 , 1221 ( 1987 ) p. gurin and s. varga , phys .
e 82 , 041713 ( 2010 ) s. n. majumdar and a. j. bray , phys .
86 , 3700 ( 2001 ) s. redner in @xcite d. toussaint and f. wilczek , j. chem .
78 , 2642 ( 1983 ) b. p. lee and j. cardy , j. stat .
80 , 971 ( 1995 ) ; j. stat .
87 , 951 ( 1997 ) s. j. odonoghue and a. j. bray , phys .
e 64 , 041105 ( 2001 ) g. odor , rev .
76 , 663 ( 2004 ) r. dandekar , arxiv : 1602.05483 s. bondyopadhyay , phys .
e 88 , 062125 ( 2013 ) s. kwon and j. m. kim , phys .
e 91 , 012149 ( 2015 ) s. b. lee , j. korean phys .
64 , 1636 ( 2014 ) s. redner , @xmath241 ( cambridge university press , cambridge , 2001 )
|
we demonstrate the large scale effects of the interplay between shape and hard core interactions in a system with left- and right - pointing arrowheads @xmath0 on a line , with reorientation dynamics .
this interplay leads to the formation of two types of domain wall , diffusive ( @xmath1 ) and static ( @xmath2 ) .
the correlation length in the equilibrium state diverges exponentially with increasing arrowhead density , with an ordered state of like orientations arising in the limit . in time , the approach to the ordered state is described by a coarsening process governed by the kinetics of domain wall annihilation @xmath3 , quite different from @xmath4 kinetics pertinent to the glauber - ising model .
the survival probability of a finite set of walls is shown to decay exponentially in time , in contrast to the power law decay known for @xmath4 . in the thermodynamic limit with a finite density of walls , coarsening as a function of time @xmath5 is studied by simulation . while the number of walls falls as @xmath6 , the fraction of persistent arrowheads decays as @xmath7 where @xmath8 is close to @xmath9 , quite different from the ising value .
the global persistence too has @xmath10 , as follows from a heuristic argument . in a generalization where the @xmath2 walls diffuse slowly ,
@xmath8 varies continuously , increasing with increasing diffusion constant .
|
recently , a considerable amount of attention has been devoted to the theoretical and experimental investigation of two - dimensional ( 2d ) quantum spin systems . in case of the ideal 2d heisenberg antiferromagnet ( af ) on a square lattice ,
the magnetic long - range order is suppressed by zero - point fluctuations at any finite temperature.@xcite however , the presence of an easy - axis anisotropy can induce a finite - temperature phase transition into the nel ordered state.@xcite on the other hand , for an easy - plane anisotropy , a transition of the berezinski - kosterlitz - thouless ( bkt ) type @xcite was proposed , where the high - temperature disordered phase can be described as a gas of vortices . below the bkt transition ( @xmath5 ) , the vortices are bound in vortex - antivortex pairs , and the spin - spin correlation decay changes from exponential to algebraic.@xcite a crossover from the isotropic to xy behavior in low fields has been predicted for 2d heisenberg af.@xcite interplane interactions ( which are always present in real materials ) can significantly modify the ground - state properties of quasi-2d systems , inducing for instance a phase transition with three - dimensional ( 3d ) long - range magnetic order . due to the onset of 3d ordering in most real materials ,
the critical behavior of diverging quantities as anticipated for ideal 2d systems at @xmath5 is hardly observable experimentally .
the quasi-2d heisenberg af model with interplane interactions can be described by the hamiltonian @xmath6 with @xmath7 , where @xmath8 is the bohr magneton , @xmath9 and @xmath10 are the intra- and interplane exchange interactions , respectively , and @xmath11 and @xmath12 correspond to nearest - neighbor spin pairs formed in the directions parallel and perpendicular to the @xmath13 planes , respectively .
the variation of the ratio @xmath14 allows to study nicely the rich phase diagram of quasi-2d magnetic systems , including the transition from a 2d heisenberg af ( @xmath15 ) to the 3d heisenberg af ( @xmath16 ) .
remarkably , even in presence of relatively strong interplane interactions , the 2d quantum fluctuations can significantly affect the magnetic properties of such systems,@xcite resulting , for instance , in a peculiar non - monotonic field dependence of the ordering temperature .
such a behavior has recently been observed in the quasi-2d square - lattice heisenberg af [ cu(c@xmath17h@xmath17n@xmath0)@xmath0(hf@xmath0)]bf@xmath17,@xcite and in the spatially anisotropic triangular magnet cu(tn)cl@xmath0,@xcite where tn is 1,3-diaminopropane ( c@xmath18h@xmath19n@xmath0 ) . here , we report on magnetic properties of single - crystalline samples of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 , which is regarded as a nearly perfect realization of a quasi-2d @xmath13 heisenberg af .
comprehensive specific - heat , high - field magnetization and electron spin resonance ( esr ) studies allow us to identify a long - range ordered nel state with easy - plane anisotropy and to accurately estimate the spin - hamiltonian parameters .
we show that in spite of the onset of the 3d ordered state below @xmath20 , the magnetic properties of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 are still significantly affected by 2d spin correlations .
( color online ) structure of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 viewed along the @xmath21 axis . ,
scaledwidth=35.0% ] the measurements have been performed at the dresden high magnetic field laboratory ( _ hochfeld - magnetlabor dresden _ , hld ) .
the pulsed - field magnetization was obtained by integrating the voltage induced in a compensated pick - up coil system containing a sample .
the low - field magnetization below 7 t was measured using a commercial squid magnetometer .
the specific - heat measurements were performed using a continuous relaxation - time technique.@xcite for the x - band esr measurements a bruker elexsys e500 spectrometer operating at a frequency of 9.4 ghz was used .
the esr experiments in the frequency range @xmath22 ghz were performed at temperatures down to 1.3 k using a tunable - frequency home - made esr spectrometer ( similar to that described in ref . )
equipped with a 16 t high - homogeneity superconducting magnet . in our experiments we used single crystals with a typical size of 4x3x1 mm@xmath23 , synthesized by aqueous reaction of stoichiometric amounts of ammonium bifluoride , pyrazine , and copper(ii ) hexafluorophosphate hydrate ( using the same procedure as reported in ref . )
[ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 , where pyz is pyrazine ( c@xmath17h@xmath17n@xmath0 ) , crystallizes in a tetragonal lattice ( space group @xmath24 ; @xmath25 ) . the structure consists of square - lattice planes of cu@xmath26 ions bridged by pyz groups in the @xmath27 plane ( fig . [ fig : structure ] ) .
the planes are connected by hf@xmath0 groups along the @xmath28 axis .
the planes of the pyz groups are nearly orthogonal to the cu@xmath26 planes . the non - coordinating pf@xmath1 groups pack in between the planes , in the center of almost cubic units .
@xcite ( color online ) ( a ) high - field magnetization of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 ( dashed and dotted lines correspond to @xmath29 and @xmath30 , respectively ) as a function of @xmath31 , where @xmath32 is the applied magnetic field and the @xmath33-factors are @xmath34 and @xmath35 , as determined by x - band esr measurements .
the calculated magnetization for an @xmath13 2d square - lattice heisenberg af is shown by the solid line .
( b ) the same data in an expanded scale close to saturation with additional model calculations assuming a finite interplane interaction.,scaledwidth=45.0% ] the magnetization of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 was measured at 1.5 k in magnetic fields up to 50 t applied parallel and perpendicular to the @xmath28 axis ( fig .
[ fig : magnet ] ) .
the fully spin - polarized state was observed at @xmath36 t and @xmath37 t for @xmath29 and @xmath30 , respectively . in the scaled plot of fig .
[ fig : magnet ] , it becomes obvious that the magnetizations for different magnetic - field orientations coincide nicely .
this means that the magnetization anisotropy is solely determined by the @xmath33-factor anisotropy .
the numerically calculated magnetization for an @xmath13 2d square - lattice heisenberg af with an intraplane exchange interaction @xmath38 k ( refs . ) is in very good agreement with our data ( fig .
[ fig : magnet ] ) . in order to estimate the ratio @xmath14 ,
calculations using this model with different interplane interactions have been made [ fig .
[ fig : magnet](b)].@xcite by comparison with our data it can be concluded that @xmath14 must be less than 1/16 .
( color online ) ( a ) temperature dependence of the specific heat of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 for @xmath29 ( blue symbols ) and @xmath30 ( red symbols ) .
the experimental data are shifted by 1 jmol@xmath39k@xmath39 per tesla .
( b ) the @xmath40 phase diagram of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 obtained from the specific - heat data ( full symbols ) shown together with the calculated phase diagram for the ideal 2d case with @xmath41 ( open symbols).@xcite lines are guides to the eye.,scaledwidth=45.0% ] the high - resolution specific - heat measurements allowed to clearly resolve the 3d af ordering temperature in zero and applied magnetic fields up to 14 t [ fig . [
fig : phasediag](a ) ] . using the local maximum of the specific - heat anomaly
, the @xmath40 phase diagram has been extracted .
figure [ fig : phasediag](b ) shows the result together with data calculated for an ideal 2d heisenberg af system ( @xmath41 ) with @xmath38 k.@xcite the theory @xcite predicts that an applied magnetic field suppresses fluctuations of the spin @xmath42 component . as a result
, an effective easy - plane anisotropy is induced in the system , accompanied by the rise of @xmath20 . on the other hand , for sufficiently strong magnetic field , the spin - canting effect prevails .
then , the critical temperature decreases and vanishes eventually at the saturation field , when the fully spin - polarized state is reached . the measured non - monotonic @xmath40 phase diagram of [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 can be nicely explained by modifying above model with a weak interplane interaction . as a threshold for the observation of a @xmath43-like anomaly in the specific heat
, @xmath44 has been given in ref . .
this ratio is very close to that , @xmath45 , estimated from previous experiments on powdered samples.@xcite in the range @xmath46 the interplane coupling can be estimated employing the following formula@xcite @xmath47 , \label{eq : ordtemp}\ ] ] where @xmath48 is the spin stiffness for a 2d system . using the exchange coupling , obtained from the high - field magnetization data , and @xmath49
k we find @xmath50 , which is in excellent agreement with our previous estimate .
( color online ) esr signal ( transmission ) measured at 90 ghz in a temperature range of @xmath51 k for @xmath30.,scaledwidth=45.0% ] esr is a powerful technique to determine the magnetic structure and the effective spin - hamiltonian parameters of in exchange - coupled spin systems . at room temperature ,
a single esr line was observed for all field orientations with a typical sinusoidal angular dependence of the @xmath33-factor expected for a cu@xmath26 ion in an octahedral surrounding with an axial symmetry due to the jahn teller effect .
the results confirm that the main exchange paths between the magnetic ions are formed in the @xmath27 planes by @xmath52 orbitals @xcite , which directly overlap with orbitals of pyrazine ligands , creating a quasi-2d network of exchange coupled cu@xmath26 ions .
the temperature evolution of the esr spectra in [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 measured at 90 ghz for @xmath30 is shown in fig .
[ fig : spectra ] .
the temperature dependence of the resonance field , the linewidth , and the integrated intensity of the esr lines in fields applied parallel and perpendicular to the @xmath27 plane are shown in fig . [
fig : eprtdep](a)-(b ) and fig .
[ fig : susc ] .
it is evident that in the vicinity of the ordering temperature ( @xmath53 k at 3 t ) , a region of enhanced short - range spin correlations is entered and both the resonance field and the linewidth change dramatically .
it is important to mention that the integrated esr intensity is proportional to the dynamic spin susceptibility . in fig .
[ fig : susc ] , the temperature dependence of @xmath54 ( measured using a squid magnetometer with applied magnetic field @xmath55 t ) , the integrated esr intensity , and the magnetic susceptibility calculated for a 2d heisenberg af system@xcite are presented .
the results of the calculations are in good agreement with our experimental data above @xmath20 for @xmath56 k , proving the crucial role of 2d spin correlations in [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 .
( color online ) temperature dependence of the resonance field ( a ) and linewidth ( b ) obtained from esr data measured at @xmath57 ghz for a magnetic field aligned parallel ( empty squares ) and perpendicular ( empty circles ) to the @xmath27 plane .
the transition temperature , @xmath58 k , as obtained from specific - heat measurements at 3 t for @xmath29 ( fig .
[ fig : phasediag ] ) is shown by the dotted line.,scaledwidth=45.0% ] the frequency - field diagram of the esr signals is presented in fig .
[ fig : epr ] .
well above the critical temperature ( @xmath59 k ) , we observe only one resonance mode for each orientation , @xmath30 and @xmath29 , as shown in the inset of fig .
[ fig : epr ] .
the frequency - field dependence of the these modes corresponds to transitions between the energy levels of the @xmath60 cu@xmath26 ions and can be described by @xmath61 with @xmath62 and @xmath63 .
the esr excitation spectrum changes dramatically below @xmath20 , as shown for @xmath64 k in fig .
[ fig : epr ] . two modes , @xmath65 and @xmath66 , are observed when the magnetic field is applied in the @xmath27 plane . for @xmath30 , a nonlinear dependence of the mode @xmath67
was observed .
most importantly , in the esr excitation spectrum the af resonance gap was determined by us directly , @xmath68 ghz ( @xmath69 k ) at @xmath70 .
such a frequency - field diagram is a textbook example for magnetic excitations in a system with collinear 3d af long - range order and easy - plane anisotropy.@xcite the frequency - field dependence of the antiferromagnetic resonance modes can be calculated using the mean - field approximation@xcite @xmath71 for magnetic fields applied parallel to the easy plane and @xmath72 for magnetic fields applied perpendicular to the easy plane , where @xmath73 is the gyromagnetic ratio , @xmath74 and @xmath75 are the exchange and anisotropy field , respectively .
our results clearly show that below @xmath20 the @xmath27 plane is the easy plane , along which the magnetic moments are aligned at @xmath70 .
the frequency - field diagram was analyzed using the eqs .
( [ eq : omegapar ] ) ( [ eq : omegaperp ] ) .
the best fit was obtained for @xmath76 t and @xmath77 t. the exchange field , @xmath74 , is related to the exchange coupling by the simple relation @xmath78 and yields @xmath79 k , which is in perfect agreement with the value of the intraplane exchange coupling obtained from our magnetization and specific - heat measurements .
noticeably , using esr data , the anisotropy constant ( @xmath80 of the exchange interaction ) can be estimated , revealing the validity of isotropic quasi-2d s=1/2 heisenberg model applied to [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 .
( color online ) temperature dependence of @xmath54 measured at @xmath55 t ( open symbols ) and the integrated intensity obtained from esr data measured at @xmath57 ghz ( full symbols ) .
the data are presented together with the calculated results ( solid line ) for a 2d square - lattice heisenberg af system with @xmath56 k.@xcite , scaledwidth=45.0% ] ( color online ) frequency - field diagram of the esr excitations measured at @xmath81 k for a magnetic field aligned parallel ( triangles ) and perpendicular ( circles ) to the @xmath27 plane .
the signals @xmath65 , @xmath66 , and @xmath67 correspond to af resonance modes of a 3d ordered af system with easy - plane anisotropy calculated as described in the text .
the inset shows the esr modes observed above the ordering temperature .
these modes can be described as transitions between the energy levels of cu@xmath26 ions by the equation @xmath61.,scaledwidth=45.0% ]
our comprehensive specific - heat , high - field magnetization , and esr studies revealed the important role of 2d spin correlations on the magnetic properties of the @xmath13 quasi-2d heisenberg af [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 .
the observation of two af resonance modes below @xmath20 confirms the presence of a 3d collinear af long - range ordered state with an energy gap of @xmath82 ghz in the magnetic excitation spectrum and @xmath83 for the anisotropy / exchange interaction ratio .
the presence of a field - induced xy behavior has unambiguously been identified by the observation of a peculiar non - monotonic field dependence of the ordering temperature .
this work has been supported by deutsche forschungsgemeinschaft , euromagnet ( eu contract no .
228043 ) , and uchicago argonne , llc , operator of argonne national laboratory ( `` argonne '' ) .
argonne , a u.s .
department of energy office of science laboratory , is operated under contract no .
de - ac02 - 06ch11357 .
is supported also by apvv - vvce-0058 - 07 and vega 1/0078/09 .
we would like to thank l. zviagina for preparations of pulse - field magnetization measurements .
99 n. mermin and h. wagner , phys .
* 17 * , 1133 ( 1966 ) .
a. cuccoli , t. roscilde , v. tognetti , r. vaia , and p. verrucchi , phys .
b * 67 * , 104414 ( 2003 ) .
berezinskii , zhur .
* 59 * , 907 ( 1970 ) ; v.l .
berezinskii , zhur .
. fiz . * 61 * , 1144 ( 1971 ) .
kosterlitz and d.j .
thouless , j. phys .
* c5 * , l124 ( 1972 ) ; j.m .
kosterlitz and d.j .
thouless , j. phys . *
c6 * , 1181 ( 1973 ) .
r. gupta and c. f. baillie , phys .
b * 45 * , 2883 ( 1992 ) .
pires , phys .
b * 50 * , 9592 ( 1994 ) .
a. cuccoli , t. roscilde , r. vaia , and p. verrucchi , phys .
b * 68 * , 060402(r ) ( 2003 ) .
p. sengupta , c.d .
batista , r.d .
mcdonald , s. cox , j. singleton , l. huang , t.p .
papageorgiou , o. ignatchik , t. herrmannsdrfer , j.l .
manson , j.a .
schlueter , k.a .
funk , and j. wosnitza , phys .
b * 79 * , 060409(r ) ( 2009 ) . j.l .
manson , m.m .
conner , j.a .
schlueter , t. lancaster , s.j .
blundell , m.l .
brooks , f.l .
pratt , t. papageorgiou , a.d .
bianchi , j. wosnitza , and m .- h .
whangbo , chem .
commun . , 4894 ( 2006 ) .
a. orendov , e. imr , l. sedlkov , j. hanko , m. kajakov , m. orend , a. feher , j. s. xia , l. yin , d. m. pajerowski , m. w. meisel , v. zelek , s. zvyagin , and j. wosnitza , phys .
b * 80 * , 144418 ( 2009 ) .
y. wang , t. plackowski , and a. junod , physica c * 355 * , 179 ( 2001 ) .
r. lortz , y. wang , a. demuer , p.h.m .
bttger , b. bergk , g. zwicknagl , y. nakazawa , and j. wosnitza , phys .
lett . * 99 * , 187002 ( 2007 ) .
zvyagin , j. krzystek , p.h.m .
van loosdrecht , g. dhalenne , and a. revcolevschi , physica b * 346 - 347 * , 1 ( 2004 ) .
manson , j.a .
schlueter , k.a .
funk , h.i .
southerland , b. twamley , t. lancaster , s.j .
blundell , p.j .
baker , f.l .
pratt , j. singleton , r.d .
mcdonald , p.a .
goddard , p. sengupta , c.d .
batista , l. ding , c. lee , m .- h .
whangbo , i. franke , s. cox , c. baines and d. trial , j. am .
soc . * 131 * , 6733 ( 2009 ) .
goddard , j. singleton , p. sengupta , r.d .
mcdonald , t. lancaster , s.j .
blundell , f.l .
pratt , s. cox , n. harrison , j.l .
manson , h.i .
southerland , and j.a .
schlueter , new j. phys . * 10 * , 083025 ( 2008 ) .
woodward , a.s .
albrecht , c.m .
wynn , c.p .
landee , and m.m .
turnbull , phys .
b * 65 * , 144412 ( 2002 ) . c. yasuda , s. todo , k. hukushima , f. alet , m. keller , m. troyer , and h. takayama , phys . rev .
lett . * 94 * , 217201 ( 2005 ) .
p. sengupta , a.w .
sandvik , and r.r.p .
singh , phys .
b * 68 * , 094423 ( 2003 ) .
a. abragam , b. bleaney , electron paramagnetic resonance of transition ions ( claredon press , oxford , 1970 ) .
kim and r.j .
birgeneau , phys . rev .
b * 62 * , 6378 ( 2000 )
turov , physical properties of magnetically ordered crystals ( academic press , 1965 ) .
battles and g.e .
everett , phys .
b * 1 * , 3021 ( 1970 ) .
|
we report on high - field magnetization , specific - heat and electron spin resonance ( esr ) studies of the quasi - two - dimensional spin-1/2 heisenberg antiferromagnet [ cu(pyz)@xmath0(hf@xmath0)]pf@xmath1 .
the frequency - field diagram of esr modes below @xmath2 k is described in the frame of the mean - field theory , confirming a collinear magnetic structure with an easy - plane anisotropy .
the obtained results allowed us to determine the anisotropy / exchange interaction ratio , @xmath3 , and the upper limit for the inter / intra - plane exchange - interaction ratio , @xmath4 .
it is argued that despite the onset of 3d long - range magnetic ordering the magnetic properties of this material ( including high - magnetic - field magnetization and non - monotonic field dependence of the nel temperature ) are strongly affected by two - dimensional spin correlations .
|
the goal of quantum key distribution ( qkd ) is to allow two distant parties , alice and bob , to share a secret key even in the presence of an eavesdropper , eve . since in quantum mechanics measurements
irremediably perturb the systems , it is impossible for an eavesdropper to extract useful information without being noticed . quantum key distribution protocols , like the bennett - brassard 1984 ( bb84 ) protocol @xcite ,
are proven to be unconditionally secure , when using single photons . however , due to the spread of laser systems and the difficulty of realizing true single photon sources , most qkd implementations use attenuated pulsed lasers as sources .
the need to avoid multi - photon pulses , that can leak information through the photon number splitting ( pns ) attack @xcite , requires a low mean photon number per pulse .
this , however , increments the incidence of pulses containing no photons , thus limiting the key generation rate . the incidence of the pns attack can be limited using the decoy state technique @xcite , at the expenses of the necessity to modulate the laser intensity .
these limitations have encouraged the research of sources emitting a single photon for each pulse .
one possible implementation of these sources is based on localized quantum structures , such as color centres @xcite , quantum dots @xcite , atoms @xcite or ions @xcite in a cavity .
these schemes , however , show some major drawbacks : the need of expensive equipment for their operation and limitations in wavelength and bandwidth selection @xcite .
an alternative scheme for single photon sources exploits the process of spontaneous parametric down conversion ( spdc ) in a non - linear crystal .
an intense laser pump on a non - linear crystal leads to the probabilistic emission of pairs of photons , called signal and idler .
the idler photon can be used to `` herald '' the presence of the signal photon , giving the so called heralded source ( hs ) . if the duration of the pulse is much greater than the reciprocal of the phase - matching bandwidth the statistics of the pairs is still poissonian @xcite .
there are some strategies for improving the single - photon character of the heralded source .
one strategy consists in using a single hs with photon number resolving detector on the idler channel and selecting the pulses where only one photon has been detected @xcite .
an alternative strategy uses parallel hs units and post - selection : each unit is pumped with low intensity to suppress multi - photon events , while keeping the overall rate at an acceptable level .
the scheme of multiple hs with post - selection ( mhps ) has been originally proposed by migdall _
et al . _
@xcite : it used an @xmath0-to-1 optical switch triggered by a detector on the idler photon of each hs .
_ took into account also the finite efficiency of single photon detectors but , as pointed out by shapiro and wong @xcite , the current low efficiency of @xmath0-to-1 optical switches strongly limits the performance of this architecture . to overcome this limitation , they proposed a symmetric scheme ( smhps ) using @xmath0 hs units linked by @xmath1 binary polarization - based photon switches in a tree structure @xcite , see fig .
[ fig : smhps ] .
they studied the probability of single - photon and multi - photon pulses , with real detectors and optical switches .
their scheme has been subsequently implemented using 4 crystals , both with bulk optics @xcite and using hybrid photonic circuits @xcite .
a thorough analysis of multiple heralded sources with post - selection has been performed in @xcite . from their work
, it emerges that , in the case of imperfect devices , the symmetric scheme suffers from a scalability issue , with a decrease in one photon probability when increasing the number of crystals . in @xcite , a new asymmetric scheme ( amhps ) that does not present this problem
was also proposed .
the scheme is shown in fig .
[ fig : amhps ] for completeness .
the smhps and amhps schemes were also referred respectively as _ log - tree _ and _ chained scheme _ in @xcite .
sub - poissonian sources have already been demonstrated to give an improvement over poissonian ones in the key rate of a qkd system @xcite .
this article specializes the analysis to the case of the multiple - crystal heralded photon sources studied in @xcite , parametrizing them using their design parameters instead of the more experimentally relevant efficiency and second - order correlation .
moreover , the security analysis is extended to the case of general attacks by eve @xcite . in this work
, we follow @xcite , and compare the performances of the different architectures and their efficacy in a practical qkd implementation .
an ideal mhps in the configuration proposed by migdall _
et al . _ , with perfect heralding efficiency and an @xmath0-to-1 optical switch with no losses , is studied first , in order to give an account of the maximum performance that can be obtained with this kind of sources .
next , two architectures for implementing a multiplexed source in a realistic scenario are analyzed , based on _ symmetric _ and _ asymmetric _ networks of 2-to-1 switches .
finite heralding efficiency and optical switch transmittance are considered in order to illustrate the potential of these source with state - of - the - art technology .
the different types of sources are inserted in a common model of qkd , based on the bb84 protocol and taking into account channel and detector inefficiencies @xcite .
the optimization maximizes the key generation rate for the different architectures , with different numbers of hs units , for a wide range of channel losses .
the obtained values are then compared with the key generation rate of both an ideal single photon source and an attenuated laser .
the symmetric and asymmetric schemes are also optimized when used with decoy states @xcite , both in an active @xcite and in a passive scheme @xcite .
the active scheme , where alice actively changes the output statistics of its apparatus , allows her to choose an arbitrary number of decoy states , giving her the ability to estimate channel parameters with arbitrary precision .
passive decoy , on the other hand , is implemented by exploiting photon number correlations of the two outputs of non - linear crystals @xcite .
this gives less flexibility in the choice of the decoy states , thus limiting the precision of parameter estimation , but allows to reach a key generation rate comparable with active decoy without introducing further complexity in alice s apparatus .
to evaluate the performances of the heralded single photon sources , we consider the bennett - brassard 1984 ( bb84 ) qkd protocol @xcite with polarization encoding .
alice , the transmitter , encodes the random bits into the @xmath2 or the @xmath3 basis , where @xmath4 , @xmath5 , @xmath6 and @xmath7 refer to the horizontal , vertical , diagonal and anti - diagonal polarization respectively .
bob randomly measures the incoming photon into the @xmath8 or @xmath9 basis .
after the transmission , alice and bob discard the events in which the sent and measured basis are different . on the remaining events , they evaluate the secret key rate , namely the amount of unconditional secret bits that they have produced . , with @xmath10 and @xmath11 is the transmittance of 2-to-1 optical switches .
the idler of each nlc is fed into a detector with quantum efficiency @xmath12 . ] .
the idler of each nlc is fed into a detector with quantum efficiency @xmath12 . ]
the single photon source held by alice is completely described by its output statistics .
we indicate the probability of emitting @xmath13 photon by @xmath14 for the mhps and @xmath15 and @xmath16 for the smhps and the amhps , respectively @xcite ( see appendix [ sec : source ] for a more detailed description of the different architectures and the explicit expressions of @xmath17 , @xmath18 and @xmath19 ) .
the source is pumped by a pulsed laser , with phase randomization between each pulse @xcite .
the variable @xmath20 is proportional to the intensity of the pump laser , which is related to the mean number of pairs generated by each hs per pulse .
the sources are parametrized by the number of non - linear crystals ( i.e. of hs units ) @xmath0 and , for the smhps and the amhps , by the efficiency of the heralding detectors @xmath12 and the transmittance of the optical switches @xmath11 . for @xmath21 ,
the scheme used in the switching network is no longer significant , therefore the output statistics of the two architectures coincide .
if the heralding efficiency @xmath22 as well , the two architectures become equivalent to the ideal mhps .
we used as model for the qkd channel a depolarizing lossy channel ( dlc ) , characterized by a transmittance @xmath23 , with @xmath24 the loss level in decibel , and a depolarization effect with visibility @xmath5 , that takes into account also alignment and stability issues .
the receiver consists of an optical apparatus , characterized by transmittance @xmath25 , and two single photon detectors , each with quantum efficiency @xmath26 and dark count probability @xmath27 .
they are threshold detectors , i.e. they can not discriminate the number of incident photons . assuming the effects of the channel on each photon of an @xmath13-photon pulse are independent
, the probability that a detector clicks , when an @xmath13-photon signal is sent , is @xmath28 the parameters used for the evaluation of the secret key rate are the _ gain _
@xmath29 , defined as the probability that a pulse gives a click in bob s measurement apparatus , and the _ quantum bit error rate _ ( qber ) @xmath30 , i.e. the error probability in bob s detection events .
these are the only measurable parameters during a qkd experiment . for the dlc
they can be estimated as follow ( see also @xcite ) .
the gain is defined as @xmath31 where @xmath32 is the probability of having @xmath13 photons in a pulse and @xmath33 is the yield of an @xmath13-photon signal , i.e. the conditional probability of a detection event at bob s side given that alice sends @xmath13 photons .
assuming independence between signal and background , the yield of an @xmath13-photon pulse for a dlc can be predicted to be @xmath34 where @xmath35 is the probability of a dark count event , which is @xmath36 in the case of two independent detectors and small dark count probability . from now on we use the convention of indicating with tilde the predicted parameters for a dlc . the negative term , coming from the fact that real detection events and dark counts are not mutually exclusive , can be neglected , since @xmath37 .
the qber is defined by @xmath38 where @xmath39 is the @xmath13-photon error rate , i.e. the probability of an error when alice sends a @xmath13-photon state . for a dlc
the @xmath13-photon error rate can be predicted to be @xmath40 where @xmath41 is the error probability of a dark count event , which is assumed to be random , and @xmath42 is the probability that a photon hits the wrong detector .
after the transmission , alice and bob use post - processing to extract a shared , secret key from the exchanged symbols . the secret key rate @xmath43 is defined as the fraction of pulses that produce a secret bit ( without counting those discarded in the sifting phase ) @xcite .
its high dependence on source statistics makes it the most suitable parameter for the comparison of the different configurations .
the rate of the bb84 protocol is limited by the fact that eve can , in principle , obtain full information from a multi - photon pulse through the photon number splitting ( pns ) attack , without introducing any error @xcite . in the asymptotic limit of infinite key
, the achievable key rate is @xmath44 - f_{ec } h(e ) \},\ ] ] where @xmath45 is the multi - photon rate , defined as @xmath46 @xmath47 is the error correction efficiency and @xmath48 is the binary shannon entropy @xcite . for true single photon sources @xmath49 and
the secret key rate is written as @xmath50 $ ] : the correction term @xmath51 in , indeed , takes into account the possible pns attack on the multi - photon pulses .
the decoy state technique has been introduced to counteract the pns attack @xcite .
it consists on randomly varying the source statistic , so that eve can no longer adapt her attack to alice s state .
after the transmission , alice communicates bob the state she used for every pulse , allowing them to estimate channel parameters conditioned to that knowledge .
the decoy state technique is active in the sense that alice chooses the output statistics using a random number generator and ( typically ) a variable attenuator after the source . in principle
, she can choose an arbitrary number of decoy states : however it has been shown that just using the vacuum and a weak decoy state gives tight bounds on the relevant parameters @xcite .
in the asymptotic limit of infinite key , the key rate is @xmath52 - q f_{ec } h(e),\ ] ] where @xmath53 and @xmath54 are given by the source statistics in the signal state and the parameters @xmath55 , @xmath35 and @xmath56 are the channel parameters estimated using decoy states @xcite .
following @xcite , we make the simplifying assumption that the parameters have been determined exactly . in passive decoy state qkd ,
the source statistics is not under alice s direct control , but is conditioned on some random event at alice s side .
a typical example consists in an attenuate coherent state passing through a 50/50 beam splitter ( bs ) with a single photon detector at the reflected output : the photon statistic at the transmitting output of the bs changes when alice detects or not a photon . in the case of heralded sources ,
we denote by @xmath57 or @xmath58 the output statistics if , respectively , at least one detector or no detector clicks ( see appendix [ sec : passive ] for the explicit form of these probabilities for the different architectures ) .
we note that @xmath58 is not trivial since in both schemes the post - selection mechanism outputs the first hs if no detector fires @xcite .
this feature can be used to implement a passive decoy state , since eve has no way of distinguishing the statistics of each pulse before it is publicly announced . assuming the post - processing is done separately for each statistics , the key rate is @xmath59 where @xmath60 and @xmath61 are the key rate for , respectively , the case of at least one detector and no detector clicking .
the key rate is , in the limit of infinite key , @xmath62 - q^{\xi } f_{ec } h(e^{\xi}),\ ] ] where @xmath63 , @xmath64 and @xmath65 are the parameters estimated from the pulses in the corresponding statistics and @xmath66 , @xmath67 , @xmath68 are the lower ( l ) and upper ( u ) bounds for the parameters estimated from @xmath69 and the known source statistics .
the explicit formulas for parameter estimation , derived from @xcite , are given in appendix [ sec : passive ] in eqs .
in the present section we compare the performances of the smhps and amhps sources for different values of @xmath0 , namely the number of hs units considered .
the parameter @xmath20 , related to the number of generated pairs per pulse , is the free parameter used to numerically maximize the rate . for the channel and detector parameters , we used typical values of present day fibre - based qkd systems @xcite .
we consider a channel with visibility @xmath70 and losses ranging from @xmath71 to @xmath72{db}$ ] ( we note that @xmath72{db}$ ] corresponds to @xmath73{km}$ ] if we consider the typical fibre attenuation of @xmath74{db / km}$ ] ) .
bob s apparatus is characterized by optical transmittance @xmath75 and detectors with quantum efficiency @xmath76 and dark count probability @xmath77 , corresponding to the state - of - the - art of infrared semiconductor single photon detectors @xcite .
the efficiency of the error correction code is @xmath78 @xcite .
all the simulated key rates are compared with the one obtained with an attenuated laser source ( wcs ) of poissonian output statistics @xmath79 , where @xmath20 is the mean number of photons per pulse , both without and with decoy state . in the passive scheme ,
the comparison is extended to the one decoy state with attenuated laser described in @xcite , where also the inefficiencies in parameter estimation are taken into account .
furthermore , all schemes are compared with the single photon case , representing an upper bound of the rate attainable in a given configuration . in the following subsections the performances of the heralded single photon source are compared in different cases .
the ideal mhps gives the maximum key rate attainable with this kind of sources @xcite .
this value is shown in figure [ fig : single ] . the key rate increases with the number of hs units and approaches the single photon case for @xmath80 , in the low loss regime . indeed , for @xmath81 , @xmath82 and @xmath83 , approximating the single photon case for @xmath84 .
at increasing losses , however , the contribution of multi - photon pulses increases and the source shows the same behaviour as the attenuated laser .
this had already been observed in @xcite , with the difference that , for low @xmath0 , not only the fraction of multi - photon pulses is higher , determining the lower maximum tolerable loss level , but also the incidence of pulses with zero photons is stronger , determining the lower key rate at @xmath85{db}$ ] .
while these limitations are proper of the multiple - crystal architecture itself , the implementation using finite efficiency devices further reduces the key generation rate . using @xmath86 and @xmath87 as , respectively , heralding efficiency and switch transmittance @xcite , the key generation rate of the smhps and the amhps has the values shown in figure [ fig : symm_asymm ] . both the key rate and the maximum tolerable losses are lower than for the ideal mhps , since the low switch transmittance requires a higher mean number of generated pairs per pulse in order to avoid zero - photon pulses , thus increasing also the incidence of multi - photon pulses .
this effect is particularly evident in the case of the smhps , where , independently from the hs unit triggered to output , the number of switches crossed scales as @xmath10 . for few hs units , where the number of crossed switches is low ,
the predominant effect is the suppression of multi - photon events , therefore the key rate increases with @xmath0 . for a higher number of hs units ( such as @xmath88 or @xmath80 )
, the source shows a low key rate for low losses ( not much higher than the attenuated laser one ) and a high maximum tolerable loss level . in the limit @xmath81 , any advantage over the attenuated laser is lost . on the other hand
, the amhps has a more stable behaviour , since its key rate never decreases by increasing @xmath0 , but reaches an optimal value and then remains unchanged ( the key rates for @xmath89 and @xmath80 are almost equal ) .
this is due to the fact that , when a certain number of hs units has been reached , the addition of further hs units does not yield significant improvement , since the probability that the rightmost hs units are triggered to the output is negligible and the output is given only by the leftmost hs units , whose configuration does not change ( see figure [ fig : amhps ] ) .
the different behaviour of the two architectures is even more evident when studying their key rate at the variation of the two relevant source parameters , the detection efficiency @xmath12 and the switch transmittance @xmath11 .
the key rate of the smhps and the amhps is shown , respectively , in figures [ fig : symm_hg ] and [ fig : asymm_hg ] , when fixing @xmath86 and changing @xmath11 on the left and with fixed @xmath87 and changing @xmath12 on the right .
the behaviour of the smhps is highly dependent on the switch transmittance . for low @xmath11
, the benefits deriving from multiple hs units do not compensate the higher absorption rate . as evident from figure [ fig : symm_hg ] , for @xmath90 the smhps does not perform much better than the attenuated laser .
this is consistent with the results from @xcite , where it has been shown that , in the asymptotic limit @xmath81 , the smhps performs better than the attenuated laser for @xmath91 .
the curve @xmath87 corresponds to the transition between the laser - like regime and the mhps - like one .
the effect of the detector efficiency @xmath12 is evident in the high loss regime , where the influence of the lower number of multi - photon pulses is more important .
this can be evinced from the right plot of figure [ fig : symm_hg ] , where the detection efficiency @xmath12 is varied for the case @xmath87 . in the low loss regime ,
the key rate is the same for all values of @xmath12 , showing that in this region the dominating effect is photon absorption in the optical routing . on the other hand
, the high loss regime shows an improvement in the maximum tolerable loss level from @xmath92{db}$ ] for @xmath93 to @xmath94{db}$ ] for @xmath95 .
the increased detection efficiency allows a better choice of the hs unit to route to output , thus allowing the hs units to be pumped with less intensity and decreasing the incidence of multi - photon pulses .
the key rate curves of the amhps , on the contrary , show the same trend for all tested combinations @xmath96 . in the asymmetric scheme , indeed , photons emitted by the leftmost hs units pass a low number of optical switches before being routed to output , therefore the effect of photon absorption never dominates over multi - photon pulses .
the key rate for active decoy is obtained by using equation ( [ eq : active_rate ] ) , with @xmath55 , @xmath35 and @xmath56 calculated by using the channel parameters given in ( [ eq : e_n ] ) and ( [ eq : y_n ] ) . the results of simulations are shown in figure [ fig : active ] .
the normalized mean number of generated pairs that maximizes the key rate is almost constant for both sources in the range @xmath97 , with a steep fall in the regime where dark counts become important .
the rate of the smhps decreases as the number of crystals increases , thus further underlining the detrimental effect of the increased absorption in optical switches .
the amhps , on the other hand , does not show any improvement for more than 4 hs units , because the probability of triggering the rightmost hs units is negligible ( see appendix [ sec : source ] ) . since in the proposed implementation of passive decoy the bound on channel parameters is no longer optimal , simulations have to take into account also this effect on the key rate .
therefore , the key rate is calculated by inserting the bounds for @xmath55 , @xmath35 and @xmath56 into ( [ eq : passive_partial_rate ] ) and taking into account the fraction of pulses using each statistics with equation ( [ eq : passive_rate ] ) .
simulation results are shown in figure [ fig : passive ] .
the use of just two different statistics gives a non - optimal parameter estimation , so a lower key rate than the schemes using active decoy . to account for this ,
the rates in figure [ fig : passive ] are compared also with a similar , one decoy scheme implemented with attenuated laser pulses @xcite .
the worse bound on the parameters gives a worse estimation of the information leaked to eve , thus requiring the sources to be pumped with lower intensity than in the case of active decoy ( the normalized mean number of generated pairs oscillates , in this case , between @xmath98 and @xmath99 ) .
the performance of the smhps is comparable , or slightly worse , to the one of decoy state with attenuated lasers , because of the already discussed detrimental effect on the source caused by optical switch attenuation . on the other hand ,
the key rate of the amhps is always higher than the one obtained with the attenuated laser in the one decoy scheme and almost reaches the maximum tolerable loss level of the attenuated laser with decoy . as in all previous schemes ,
the amhps shows no improvement after a certain threshold of hs units is reached ( in this case , @xmath100 ) .
the comparison of the two sources in the two different decoy schemes of figure [ fig : decoy ] directly shows the advantage of the amhps over the smhps . ,
@xmath86 and @xmath87 . ] indeed , the asymmetric scheme performs better than the passive one in both active and passive decoy .
furthermore , the key rate of the amhps in the passive scheme almost equals both the smhps and the attenuated laser with decoy , despite the worse parameter estimation caused by the use of just one decoy state .
the multiple - crystal heralded sources have shown better performance than the attenuated laser in all the studied cases , with the only exception of the smhps in the one decoy scheme , where the high absorption in optical switches , together with an imperfect parameter estimation , completely overrules the enhancement given by the multiple crystal configuration .
the amhps shows better scalability than the smhps , since in the latter the addition of hs units can degrade the performance while in the former it has , in the worst case scenario , no effect .
in addition to this , the key rate of the amhps is generally higher than for the smhps .
the asymmetric structure of the amhps , on the other hand , makes its implementation harder than the smhps , since each crystal much be fed a different pump intensity and the different hs units must be carefully synchronized .
both architectures have shown a stronger dependence of the key rate on optical switch transmittance than on detector efficiency .
the effect of multi - photon pulses , even if less important than for the attenuated laser , still limits the maximum tolerable loss level .
the use of the decoy state solves this problem also for multiple - crystal heralded sources .
the best results are given by active decoy , where a variable optical attenuator and a random number generator are used to change the output statistics .
the much more complex statistics of these sources , however , makes the calculations needed for the determination of the optimal decoy parameters impractical .
therefore , the simplifying assumption of exact determination of the parameters has been adopted and the obtained results just give a superior limit on the attainable key rate .
the optimal decoy parameters must be calculated on a case by case basis , once the characteristics of the source have been chosen .
these limitations might make it preferable the use of a passive decoy scheme , which can be easily implemented in both sources by exploiting the post - selection mechanism . in this case , the optimization has taken into account also the inefficiencies of parameter estimation , thus the obtained results give the effective key rate attainable with each configuration and not just a superior limit .
all the results here presented are subjected to the approximation of infinitely long key .
we leave finite key effects for future studies .
these effects are important especially for the schemes with decoy state , since each photon statistics requires a sufficiently high number of events . in the passive decoy scheme
the effect is still more important , since the need of considering also the relative frequency of the two photon statistics can drastically change the optimal source parameters .
the recent advances in integrated photonics suggest an increasing role of multiple - crystal heralded sources .
if built into a single chip , they might be a valid alternative to lasers in quantum key distribution and other quantum information tasks requiring single photons to work properly .
our analysis show the supremacy of the asymmetric scheme with respect to the symmetric one for qkd applications .
the authors acknowledge the strategic - research - project quintet of the department of information engineering , university of padova and the strategic- research - project quantumfuture of the univer- sity of padova .
ms acknowledges the ministry of research and the center of studies and activities for space ( cisas ) `` giuseppe colombo '' for financial support .
the multiple crystal heralded source with post - selection ( mhps ) consists on an array of @xmath0 hs units , simultaneously pumped with a laser pulse with intensity such that the mean number of generated pairs per pulse is @xmath20 @xcite . for each hs unit ,
labelled with index @xmath101 , the idler photon is used as a trigger for the signal one , which is injected into an optical switch , as shown in figure [ fig : mhps ] . .
the idler photon is fed into a detector ( apd ) , while the signal one is routed through an optical switch ( o.s . ) to the output . ] in the ideal case of perfect detector and optical switch , a post - selection mechanism that selects one of the channels whose detector has fired gives the output statistics @xmath102 where @xmath103 is the kronecker delta ( @xmath104 and @xmath105 ) @xcite .
taking into account real devices , the mhps would face the low efficiency of @xmath0-to-1 optical switches .
this can be avoided by replacing the @xmath0-to-1 switch with a tree structure of 2-to-1 switches , giving the symmetric mhps ( smhps ) @xcite shown in figure [ fig : smhps ] .
the structure of this scheme requires the number of hs units @xmath0 to be a power of 2 .
the idler photon of each hs unit is fed into a detector with quantum efficiency @xmath12 , while the signal photon is directed to a 2-to-1 optical switch of transmittance @xmath11 . since the photons produced by each hs unit pass @xmath10 switches before reaching the output , the crystals are pumped with an intensity such that the mean number of generated pairs per pulse is @xmath106 . the post - selection mechanism in each optical switch gives priority to the left hs unit and , in case no hs unit triggers , always outputs the left one @xcite .
the photon statistics at the output is @xmath107 in the asymmetric mhps ( amhps ) , the @xmath0 hs units are arranged following the scheme shown in figure [ fig : amhps ] @xcite .
each hs unit , whose idler photon is fed into a detector of efficiency @xmath12 , is pumped with an intensity such that the mean number number of generated pair per pulse is @xmath108 , with @xmath109 in order to compensate the different number of traversed 2-to-1 optical switches .
the post - selection mechanism of each optical switch is the same as for the smhps , i.e. it gives priority to the left hs unit and , if none triggers , outputs the left one .
differently from the smhps , this architecture requires delay lines to be introduced , in order to compensate the longer transmission time of the rightmost hs units @xcite .
the statistics at the output is @xmath110 e^{-(1-\eta ) \mu}}{n ! } e^{-\eta
\mu \frac{(2-\gamma)\gamma^{1-m } - 1}{1-\gamma } } \\ + \frac{\mu^n e^{-\mu}}{n ! } \sum_{i=1}^{m } e^{-\eta \mu \frac{\gamma^{1-i}-1}{1-\gamma } } [ 1 - ( 1-\eta)^n e^{\eta \mu } e^{- \frac{\eta \mu}{\gamma^{k_i } } } ] .\end{gathered}\ ] ]
the parameters @xmath35 , @xmath56 and @xmath55 , necessary in post - processing , are not directly measured during the key exchange session , but must be estimated from the experimental data @xmath29 and @xmath30 .
if alice registers , for each pulse , whether at least one or no detector has clicked , a different gain and qber for each case can be measured and these can be used for parameter estimation . since both statistics depend on the same @xmath20 , parameter estimation can be included in the general optimization process , thus giving the real key rate and not just a superior limit .
the probability that no detector clicks in a pulse is @xmath111 for the smhps and @xmath112 for the amhps , where @xmath10 .
the statistics in the case of no click is the same for both sources ( in both cases the first hs unit is routed to the output ) and is @xmath113^n}{n ! } e^{-\mu ( 1-\eta)},\ ] ] while the statistics for the case of at least a detector click is @xmath114 \frac{1}{1 - e^{\frac{-\mu \eta}{\gamma^k}}}\ ] ] for the smhps and @xmath115 \frac{e^{-\mu \eta \frac{\gamma^{1-i}-1}{1-\gamma}}}{1 - e^{-\mu \eta \frac{(2-\gamma)\gamma^{1-m } - 1}{1-\gamma}}}\end{gathered}\ ] ] for the amhps @xcite .
after the key exchange session , alice tells bob for which pulses at least one detector has clicked , so that they can estimate the gain and the qber separately for the two cases . from these values , referenced to as @xmath116 , and the known source statistics @xmath117 and @xmath118 , they can estimate the parameters of the channel using the method described in @xcite .
the first parameter to be estimated is @xmath35 .
its upper bound @xmath119 can be calculated starting from the relations @xmath120 since both inequalities must hold , the parameter @xmath35 is upper bounded by @xmath121 its lower bound @xmath66 can be calculated from @xmath122 that gives @xmath123 since , for both the smhps and the amhps @xmath124 \left\ { \begin{array}{lr } \leq 0 & \text{for } n \geq 2 \\
\geq 0 & \text{for }
n = 0 , \end{array } \right .
\end{gathered}\ ] ] with @xmath125 a positive constant .
the lower bound on the single photon yield @xmath35 is calculated starting from @xmath126 which leads to @xmath127 since @xmath128 \left\ { \begin{array}{lr } \leq 0 & \text{for } n \geq 2 \\
\geq 0 & \text{for } n \leq 1 \end{array } \right .
\end{gathered}\ ] ] with @xmath129 positive .
|
single photon sources ( sps ) are a fundamental building block for optical implementations of quantum information protocols . among spss ,
multiple crystal heralded single photon sources seem to give the best compromise between high pair production rate and low multiple photon events . in this work ,
we study their performance in a practical quantum key distribution experiment , by evaluating the achievable key rates .
the analysis focuses on the two different schemes , symmetric and asymmetric , proposed for the practical implementation of heralded single photon sources , with attention on the performance of their composing elements .
the analysis is based on the protocol proposed by bennett and brassard in 1984 and on its improvement exploiting decoy state technique . finally ,
a simple way of exploiting the post - selection mechanism for a passive , one decoy state scheme is evaluated .
|
faraday waves , which typically refer to complex patterns of standing waves on a fluid surface in an oscillating container , are among the classical problems of fluid mechanics @xcite .
the phenomenon has been a representative example of parametric instabilities @xcite . as is often the case with fluid mechanics , even classical phenomena are often not well understood in nonlinear regimes . indeed ,
continuing experiments on faraday waves beyond the linear regimes reveal intriguing new features , which include snake - like structures in drop - confined faraday waves @xcite , a turbulent state mediated by defects of the pattern @xcite and so - called oscillons @xcite . these surprising findings would probably be outside of the applicable range of the weakly nonlinear theories , such as the one developed by @xcite , on selection of various patterns of faraday waves .
hence , to understand these phenomena , fully nonlinear numerical simulations of the faraday systems , which can be complementary to laboratory experiments , play an indispensable role as discussed in @xcite .
perhaps the first such simulation , in which the motions of both the top and bottom fluids are simultaneously simulated , was performed recently by @xcite .
this motivates our present study . in these fully nonlinear simulations ,
the inevitable difficulties are how to represent numerically the interface between the two immiscible fluids , how to follow its motion and how to calculate its influence on the bulk fluids . in @xcite
, the interface is modelled by triangular elements whose apices are advected vertically and data on the interface elements are copied to the eulerian grids in the bulk of the fluids according to the recipe of the immersed boundary method @xcite . since numerical modelling of the interface involves various assumptions , its validation is required against laboratory experimental data or other data independent of the modelling . in the linear regime ,
the analytical result of the linear theory of faraday waves due to @xcite provides reliable data for comparison , as used in @xcite . in nonlinear regimes ,
comparison with experimental data is crucial .
remarkably , the simulation result by @xcite is in perfect agreement with the experimental data in the nonlinear regime conducted by @xcite , where the top fluid s motion can not be neglected .
the further task of such validated numerical simulations is to investigate data not easily accessible in experiments , such as nonlinear energy transfers between modes . for this purpose , in our opinion , it is important to have at least two validated numerical simulations with independent interface modelling and make sure that these simulations give consistent results .
the aim of this study is to develop another nonlinear simulation method of faraday waves with an alternative interface model to that devised by @xcite .
we here apply to the faraday wave problem the phase - field modelling with the cahn - hilliard equation for binary fluids @xcite .
the phase - field method , like front - tracking methods , volume - of - fluid methods and level - set methods , easily allows situations where the interface becomes a multivalued function of the horizontal coordinates .
see , for example , @xcite for recent application of the phase - field method to rayleigh - taylor instability and to other flows in the references therein . to our knowledge , the method hat not been applied to the faraday wave problem .
in this paper we focus on faraday waves in two spatial dimensions ( 2d ) to explore the capabilities of the phase - field method in a simple setting .
earlier numerical studies of 2d faraday waves include @xcite , @xcite and @xcite , where the fluid dynamical equations for the bottom fluid are solved but the top fluid s motion is neglected in contrast to the present simulation . another numerical approach using different formalisms ( the boundary integral and the vortex sheet ) is explored by @xcite .
here we proceed as follows . in the linear regime
, we compare quantitatively the phase - field simulation with the linear theory by @xcite for validation as in @xcite . in the nonlinear regime , unfortunately suitable experimental data in 2d for comparison with our simulation are not available .
however , we qualitatively make comparison with the simulation of 2d faraday waves with the vortex - sheet formulation by @xcite in the regime of plume formation , where the interface becomes a multivalued function . finally , we present simulation results concerning the period tripling state , where the oscillation period of the faraday waves becomes three times the basic period .
this state is found experimentally by @xcite in quasi - two dimensional faraday waves , by @xcite in axisymmetric three - dimensional faraday waves , and numerically also by @xcite .
its underlying physics is not yet understood .
the organization of the paper is the following . in 2 , we describe the basic fluid dynamical the equations and our phase - field model of the faraday waves .
our numerical method for solving the equations is explained in 2 .
we then make a quantitative comparison of its simulation results with the linear analysis in 3 . in
4 , we present results of the phase - field simulation in nonlinear regimes .
the first simulation shows plume formation , in which the interface overturns . the second one concerns the period tripling state .
concluding remarks are made in the last 5 .
we consider faraday waves between two immiscible fluids in two spatial dimensions .
the fluid dynamical equations are the incompressible navier - stokes equations @xmath0&= & - { \boldsymbol{\nabla}}p + \rho \bm{g } + { \boldsymbol{\nabla}}{\boldsymbol{\cdot}}\eta ( { \boldsymbol{\nabla}}\bm{u } + { \boldsymbol{\nabla}}\bm{u}^{t } ) + \bm{s},\label{ns}\\ { \boldsymbol{\nabla}}{\boldsymbol{\cdot}}\bm{u } & = & 0 , \label{incomp}\end{aligned}\ ] ] where @xmath1 is the pressure and @xmath2 is the velocity . here
@xmath3 is the surface tension force , which will be discussed below . the density @xmath4 and viscosity @xmath5 are distinct constants for each fluid , which are denoted as @xmath6 and @xmath7 respectively , for the top and bottom fluids ( @xmath8 ) . finally , @xmath9 is the gravitational acceleration , incorporating the vibration forcing with amplitude @xmath10 and angular frequency @xmath11 , which is written as @xmath12 here in two dimensions , @xmath13 .
the temporal period of the vibration forcing is denoted by @xmath14 , which will be used often in what follows .
the equations are solved with the boundary conditions of the faraday setting as depicted in figure [ f : setting ] . in the horizontal direction ( coordinate @xmath15 ) , we assume periodic boundary conditions with length @xmath16 . for the vertical direction ( coordinate @xmath17 ) , no - slip boundary conditions at the boundaries @xmath18 are assumed , that is , @xmath19 .
the interface position between the top and bottom fluids , denoted as @xmath20 , obeys the kinematic boundary condition . in terms of this @xmath21 ,
the density @xmath4 and viscosity @xmath5 in ( [ ns ] ) are written as function of @xmath17 : @xmath22 in this sharp interface formulation , the density and viscosity vary discontinuously as a function of @xmath17 , which makes numerical simulations in this form very difficult in practice . one standard way to overcome this in numerical simulations is to model the sharp interface by a diffuse one @xcite .
the phase - field modelling is one such diffuse - interface method , which we will explain below .
we adapt the phase - field modelling of binary fluids @xcite to faraday waves .
the top and bottom fluids are here indicated by the phase variable values @xmath23 and @xmath24 , respectively . the interface is modelled as the region where @xmath25 changes continuously from @xmath26 to @xmath27 .
the phase evolves to minimize the free energy functional of the phase @xmath25 : @xmath28 = \frac{\lambda}{2 } \int \left [ \frac{1}{2\epsilon^2}(\phi^2 - 1)^2 + |{\boldsymbol{\nabla}}\phi|^2 \right ] d\bm{x}.\end{aligned}\ ] ] accordingly , the equation of the phase variable is the cahn - hilliard equation with the advection term @xmath29 where @xmath30 is the chemical potential , @xmath31 is the magnitude of the free energy and @xmath32 is the mobility .
we assume that the mobility @xmath32 is constant .
the parameter @xmath33 controls the thickness of the smoothed interface .
these @xmath32 and @xmath33 are adjustable parameters of the phase - field modelling . how to determine them is discussed at the end of section [ s : linear ] . in terms of @xmath25 , the density and viscosity are expressed as @xmath34 in the framework of the phase - field modelling , the surface tension force written as @xmath3 in ( [ ns ] ) is given by @xmath35 @xcite .
the surface tension @xmath36 used in the sharp interface model is related to the phase - field simulation parameters as @xmath37 for planar interfaces . in the following
we assume that the correspondence ( [ sigma ] ) is valid for non - planar interfaces .
we use the following boundary conditions of the phase variable on the top and bottom walls @xmath18 : @xmath38 @xcite . in the horizontal direction ,
the boundary condition is periodic . for further details of the derivation of the phase - field model
, we refer the reader to @xcite . to summarise this subsection , the equations to be solved numerically are the incompressible navier - stokes equations ( [ ns ] ) and ( [ incomp ] ) and the cahn - hilliard equation ( [ ch ] ) .
the boundary conditions on the velocity are periodic in the horizontal direction and no - slip on the top and bottom walls in the two - dimensional space . for the phase variable ,
the boundary conditions are periodic in the @xmath15-direction and ( [ bphi ] ) in the @xmath17-direction .
we start with the discretization of the cahn - hilliard equation .
the computational mesh is a standard staggered arrangement as in @xcite .
we fourier - expand the phase variable @xmath25 in the periodic @xmath15-direction as @xmath39 the wavenumber is given by @xmath40 , where @xmath41 is the number of grid points in the @xmath15-coordinate .
then the cahn - hilliard equation in the @xmath42 space becomes @xmath43 \right\ } , \label{ch - hat } \end{aligned}\ ] ] where the cubic term @xmath44 is the fourier transform of the product @xmath45 calculated in the physical space @xmath46 .
no aliasing error is removed in this calculation of @xmath44 .
the advection term @xmath47 is calculated as follows .
first it is calculated in the physical space by using an elementary discretization in the staggered mesh as @xmath48_{l,\ , m } = \left ( u\frac{{\ensuremath{\partial}}\phi}{{\ensuremath{\partial}}x } + w\frac{{\ensuremath{\partial}}\phi}{{\ensuremath{\partial}}z } \right)_{l,\ , m } \nonumber\\ & & = \frac { u_{l - 1/2,\ , m } + u_{l + 1/2,\ , m } } { 2 } \,\,\frac{\phi_{l + 1,\ , m } - \phi_{l - 1 , \ , m } } { 2 \delta x } + \frac { w_{l,\ , m - 1/2 } + w_{l , \ , m + 1/2}}{2 } \,\,\frac{\phi_{l,\ , m + 1 } - \phi_{l , \ , m - 1 } } { 2 \delta z}\nonumber\\. \label{ch - ad - re}\end{aligned}\ ] ] in this notation @xmath49 denotes the velocity on the cell face @xmath50 and @xmath51 denote the phase variable on the cell centre @xmath52 .
the indices run as @xmath53 and @xmath54 ( @xmath55 is the number of grid points on the @xmath17-coordinates ) . in the denominator @xmath56 and @xmath57
are the grid spacings of the @xmath15 and @xmath17 coordinates .
the fourier transform of ( [ ch - ad - re ] ) gives @xmath47 .
again no aliasing error is removed .
equation ( [ ch - hat ] ) is discretized in time by using the semi - implicit stabilized scheme due to @xcite as @xmath58 \right\}. \nonumber\\ \label{plm}\end{aligned}\ ] ] here @xmath59 denotes the data at the @xmath60-th time step . on the right hand side ,
the fourth - order derivative term is treated fully implicitly .
the second - order term is handled semi - implicitly , which is split into two terms : one involves @xmath61 and the other @xmath62 as proposed by @xcite .
the @xmath17-derivatives in ( [ plm ] ) are evaluated as @xmath63 where @xmath64 denotes @xmath65 .
we use the bi - conjugate gradient stabilized ( bicgstab ) method @xcite to solve the equations ( [ plm ] ) .
the boundary conditions of @xmath25 in the @xmath17-direction are given in ( [ bphi ] ) .
now we turn to discretization of the incompressible navier - stokes equations ( [ ns ] ) and ( [ incomp ] ) .
we follow mostly the discretization method proposed by @xcite except for the surface tension force . in order to describe how we calculate the surface tension force in the present phase - field modelling , we repeat here some of the equations given in 3.3 of @xcite .
the method starts with the intermediate velocity @xmath66 from @xmath67 at the @xmath68-th time step , @xmath69 where @xmath70 and @xmath71 are here determined by the phase @xmath72 as ( [ rhoeta ] ) .
yet another intermediate velocity @xmath73 involves the gravity and the vibration forces @xmath9 and the surface tension force @xmath3 : @xmath74 since this @xmath3 is represented as ( [ s ] ) in the phase - field method , we discretize each component first as @xmath75 then , the vectors @xmath76 on the staggered grids are @xmath77 here the chemical potential ( [ mu ] ) is discretized as @xmath78 .
\nonumber\\\end{aligned}\ ] ] finally , the velocity at the @xmath79-th time step is obtained by @xmath80 where the pressure @xmath81 is determined , by demanding @xmath82 , as @xmath83 again the bicgstab method is employed to solve ( [ utilde ] ) and ( [ p ] ) . the boundary conditions for the velocity and the intermediate one @xmath66 are the same as in @xcite , namely , @xmath84 , @xmath85 and @xmath86 = 0 $ ] . as proposed by @xcite , in solving the equations ( [ utilde ] ) , we use the latest updated velocity component of @xmath66 to compute the other component . in this process , the order ( which component of @xmath66 is computed first and last ) is also interchanged at each time step to ensure symmetry .
we compare our simulation result with the linear theory of @xcite .
we follow here again the validation method proposed by @xcite . in the linear theory elaborated by @xcite ,
the faraday wave problem is formulated as a finite - depth , viscous binary fluid system with a sharp interface representation .
this provides the critical value of the vibration forcing amplitude @xmath87 as eigenvalues for a given perturbation @xmath88 .
the floquet modes @xmath89 of the interface position @xmath20 can be calculated as eigenvectors . in other words , the spatio - temporal variation reads@xcite @xmath90 recall that @xmath11 is the angular frequency of the vibration forcing ( [ g ] ) . here
the number of floquet modes @xmath91 is infinite in theory but it is finite for numerical calculations . the exponent @xmath92 is the floquet exponent . for the critical perturbations @xmath93 is zero .
the harmonic @xmath94 and subharmonic @xmath95 cases are considered here as in @xcite .
we numerically solve this eigenvalue problem ( for the precise form of the matrices , see @xcite ) .
the mathematica script that we use here is available at http://www.kyoryu.scphys.kyoto-u.ac.jp/~takeshi/kt94 .
the parameters here are the same as in @xcite in their validation with the linear theory .
the parameter values are @xmath96 , @xmath97 kg m@xmath98 , @xmath99 kg m@xmath98 , @xmath100 pa s , @xmath101 pa s , @xmath102 n m@xmath103 , @xmath104 m s@xmath105 , @xmath106 m and @xmath107 s@xmath103 .
the unperturbed interface is in the centre of the container .
the critical vibration forcing amplitude @xmath87 from the linear theory as a function of the perturbation wavenumber @xmath108 is plotted as a curve in figure [ f : ac ] .
the floquet modes @xmath109 are used later in figure [ f : floquet ] . with the phase - field simulation , critical values @xmath87 for four different @xmath108 s are determined , which are denoted as points in figure [ f : ac ] .
refer also to table [ t : ac ] for a precise comparison with the values from the linear theory calculation .
we see that agreement between the two is satisfactory .
the way to estimate @xmath87 in the phase - field simulation is as follows .
( i ) we consider the sharp interface location @xmath20 in the linear analysis to be the null point @xmath110 in the phase - field modelling .
numerically , such null points are calculated using the linear interpolation of @xmath25 data on the grid points ( we use this correspondence throughout this paper ) .
( ii ) for a given acceleration amplitude @xmath10 , we monitor the oscillating interface position at @xmath111 as a function of time .
we perform simulations by changing @xmath10 from the theoretical @xmath87 and estimate the critical value at which the temporal oscillation neither decays nor grows . more precisely , we take the absolute relative difference between the two peak values of the interface positions around @xmath112 and @xmath113 .
if the difference is smaller than @xmath114 , we regard this @xmath10 as the critical acceleration of the phase - field simulation . the largest error shown in table [ t : ac ] ( the case of @xmath115 )
would be due to a relatively large @xmath116 . here
we set the parameters of the phase - field simulation to be @xmath117 m , @xmath118 , @xmath119 ( see ( [ sigma ] ) ) and @xmath120 for all four cases .
the initial perturbed interface is written in terms of the phase variable as @xmath121}{\sqrt{2 } \epsilon}\right\ } , \label{e : tanh}\end{aligned}\ ] ] where @xmath122 corresponds to the perturbation amplitude of the interface position . here
we take @xmath123 .
we recall that in the unperturbed state the cahn - hilliard equation has the stationary solution @xmath124 $ ] ( see , e.g. @xcite ) .
the initial velocity in the phase - field simulation is set to zero .
.critical amplitude values @xmath87 of the linear theory calculation and the phase - field simulation . here
the horizontal length of the simulation @xmath16 is taken to be equal to the wavelength of the perturbation @xmath125 for every case .
the number of grid points of the simulation is @xmath126 [ cols="^,^,^,^,^ " , ] .
[ t : ac ] figure [ f : floquet ] shows temporal variations of the interface location obtained by the floquet analysis results ( [ floquet ] ) and the phase - field simulation results with critical @xmath87 values .
we here follow the representation used in @xcite .
the initial discrepancies between the linear theory and the simulation remain large until one vibrating period @xmath127 .
however , later than that , the phase - field solutions agree satisfactorily with the floquet analysis result as shown in figure [ f : floquet ] .
in contrast , the discrepancies between the simulation by @xcite and the linear theory vanish within a quarter of the period of the vibration . of the phase - field simulation ( circles ) and
the floquet analysis result ( curves ) .
the interface position is defined as the phase null points @xmath110 .
the thick arrows indicate the size of @xmath33 , which is the length scale of the diffuse interface .
( a ) perturbation wavenumber @xmath128 mm@xmath103 ( subharmonic case ) ; ( b ) @xmath129 mm@xmath103 ( harmonic case ) ; ( c ) @xmath130 mm@xmath103 ( subharmonic case).,width=288,height=192 ] of the phase - field simulation ( circles ) and the floquet analysis result ( curves ) .
the interface position is defined as the phase null points @xmath110 .
the thick arrows indicate the size of @xmath33 , which is the length scale of the diffuse interface .
( a ) perturbation wavenumber @xmath128 mm@xmath103 ( subharmonic case ) ; ( b ) @xmath129 mm@xmath103 ( harmonic case ) ; ( c ) @xmath130 mm@xmath103 ( subharmonic case).,width=288,height=192 ] of the phase - field simulation ( circles ) and the floquet analysis result ( curves ) .
the interface position is defined as the phase null points @xmath110 .
the thick arrows indicate the size of @xmath33 , which is the length scale of the diffuse interface .
( a ) perturbation wavenumber @xmath128 mm@xmath103 ( subharmonic case ) ; ( b ) @xmath129 mm@xmath103 ( harmonic case ) ; ( c ) @xmath130 mm@xmath103 ( subharmonic case).,width=288,height=192 ] finally , we comment on the values of the mobility @xmath32 and the thickness of the interface @xmath33 , which are adjustable numerical parameters of the phase - field simulation . the mobility @xmath32 is determined as follows .
the physical time scales of the various forces in ( [ ns ] ) can be dimensionally estimated for a given length scale @xmath131 as @xmath132 and @xmath133 , which are times scales of the gravity , viscosity and surface tension , respectively .
the time scale associated with the mobility is in a similar way given as @xmath134 by ( [ ch ] ) .
the mobility value is determined by requiring @xmath135 for a small length scale , say @xmath136 .
this means that the interface s relaxation to equilibrium occurs faster than other dynamics .
here we have @xmath137 and @xmath138s .
to have the condition @xmath135 , we set @xmath139 kg@xmath103m@xmath140s , which yields @xmath141s .
that is , @xmath142 here .
if this ratio is larger than @xmath143 , our simulation does not agree with the linear theory .
here we take the time step to be @xmath144s , which is about an order of magnitude smaller than @xmath145 . the thickness @xmath33 is determined as follows .
we first fix the perturbation amplitude to @xmath146 and then do simulations with @xmath147 .
good agreement with the @xmath87 of linear theory is obtained for @xmath148 .
this leads us to choose @xmath149 .
having validated the phase - field approach in the linear regime , we now move to nonlinear regimes .
two cases are considered . the first one concerns plume formation , which may eventually lead to droplet ejection at long time .
however , we do not follow the motion until the ejection or pinchoff occurs .
the second case is the period tripling state of the two - dimensional faraday waves .
plume formation of the faraday interfaces in two dimensions was studied numerically by @xcite using the vortex - sheet method , with which we compare the phase - field simulation .
the presence of a plume implies that the interface profile becomes a multivalued function of the horizontal coordinate .
such situations can be numerically handled by the phase - field method without _ ad hoc _ adjustments .
here the comparison is qualitative .
there are three reasons for this : the difference in atwood number @xmath150 , the difference in the boundary conditions , and with or without viscosity .
regarding the atwood number difference , the phase - field method with the discretization described in section [ ss : nm ] is not able to handle high atwood number @xmath151 .
our simulation works for at most @xmath152 in practice ( this issue is discussed in the last [ s : discon ] ) .
in contrast , the vortex - sheet calculation by @xcite can cope with up to unit atwood number .
concerning the difference in the boundary conditions , there is no rigid boundary in the vertical direction in their vortex - sheet calculation .
however , in our case fluids are contained between two rigid walls .
the third reason is that the vortex - sheet calculation is inviscid , while the phase - field method is viscous . in @xcite
, the overturning interface corresponding to formation of a plume at @xmath153 and @xmath143 is numerically studied .
we observe a similar overturning behaviour in the phase - field simulation even at as low as @xmath154 , which we show in figure [ f : ot ] .
the parameters of the phase simulation are : @xmath155 @xmath156 and @xmath157 .
the force @xmath9 in ( [ ns ] ) is here @xmath158 , oscillating around zero .
the initial interface is given by ( [ e : tanh ] ) with @xmath159 .
the time step and the number of grid points are @xmath160 and @xmath161 . here
we take the same parameters @xmath162 and the same vibration force @xmath9 setting ( oscillating around the origin ) as @xcite .
the differences between our simulations and theirs are the atwood number , the vibration amplitude value ( ours are nine times larger ) , the presence of the top and bottom rigid walls and the presence of viscosity . the time variation of the interface location at a fixed horizontal point is shown in figure [ f : ot](a ) .
the interface location first exhibits a rise and then a fall ( here at about @xmath163 and @xmath164 , respectively , where @xmath127 is the vibration period ) ; then it grows steeply ( at around @xmath165 ) and reaches a high plateau ( between @xmath166 and @xmath167 ) with a small oscillation .
this variation is similar to that with @xmath153 obtained with the vortex - sheet method ( figure 11(a ) of @xcite ) . in our simulation ,
the interface location later than @xmath168 ( continuation of figure[f : ot](a ) but not shown ) goes downwards and shows again a plateau with a small oscillation .
then it goes upwards and finally comes back to zero around @xmath169 .
this later variation can not be compared with the vortex - sheet method result of @xcite because their calculation is stopped before @xmath170 .
the phase field @xmath25 at @xmath171 is shown in figure [ f : ot](b ) , which clearly shows that the interface turns over and becomes a multivalued function of the horizontal coordinate @xmath15 .
qualitatively similar interface structures , called plumes , are shown in figure 11(b ) of @xcite . to check the boundary effect
, we also do the phase - field simulation with aspect ratio 2 ( @xmath172 ) .
the plume shape and its time evolution do not change qualitatively , indicating that the effect of the vertical boundary is small on the plume formation .
this agreement suggests that the phase - field simulation is valid , at least qualitatively , in the overturning regime although it can not handle cases with high atwood number . in figure
[ f : ot](c ) and ( e ) , we plot the vorticity field @xmath173 at the plume state .
indeed much of the vorticity concentrates inside the interface region .
the vorticity along the interface in the right half domain has three peaks as seen in the contour plot of figure [ f : ot](e ) .
in contrast , in the simulation of @xcite , the strength of the vortex sheet has at most two peaks in the developed plume state . in the phase - field simulation
we observe that the vorticity outside of the interface reaches about 40% of the maximum vorticity in the interface region .
in addition , thin vortex layers are attached near the top and bottom walls . in these layers ,
the vorticity is about 10% of the maximum in the interface region .
we now comment on the choice of the resolution @xmath174 used here .
we observe that the squared modulus of the fourier coefficients @xmath175 decreases exponentially for large @xmath108 such as @xmath176 $ ] .
we measure this factor @xmath177 and require @xmath178 \ge 2 \delta x = 2 l_x / n_x$ ] .
we take the smallest @xmath41 in powers of 2 satisfying this relation .
we then set @xmath57 to about the same value of @xmath56 . .
the interface position is defined from the phase field @xmath25 as described in section [ s : linear ] .
( b ) grey - scale coded phase field @xmath25 at time @xmath171 .
( c ) grey - scale coded vorticity field at the same instant .
( d ) interface profiles at four instants .
( e ) contours of the vorticity at @xmath171 .
contour values are from -20.0 to 20.0 by 4.0 .
the solid and dashed lines correspond to non - negative and negative vorticity values .
, width=288,height=192 ] .
the interface position is defined from the phase field @xmath25 as described in section [ s : linear ] .
( b ) grey - scale coded phase field @xmath25 at time @xmath171 .
( c ) grey - scale coded vorticity field at the same instant .
( d ) interface profiles at four instants .
( e ) contours of the vorticity at @xmath171 .
contour values are from -20.0 to 20.0 by 4.0 .
the solid and dashed lines correspond to non - negative and negative vorticity values .
, title="fig : " ] .
the interface position is defined from the phase field @xmath25 as described in section [ s : linear ] .
( b ) grey - scale coded phase field @xmath25 at time @xmath171 .
( c ) grey - scale coded vorticity field at the same instant .
( d ) interface profiles at four instants .
( e ) contours of the vorticity at @xmath171 .
contour values are from -20.0 to 20.0 by 4.0 .
the solid and dashed lines correspond to non - negative and negative vorticity values .
, title="fig : " ] .
the interface position is defined from the phase field @xmath25 as described in section [ s : linear ] .
( b ) grey - scale coded phase field @xmath25 at time @xmath171 .
( c ) grey - scale coded vorticity field at the same instant .
( d ) interface profiles at four instants .
( e ) contours of the vorticity at @xmath171 .
contour values are from -20.0 to 20.0 by 4.0 .
the solid and dashed lines correspond to non - negative and negative vorticity values .
, title="fig : " ] .
the interface position is defined from the phase field @xmath25 as described in section [ s : linear ] .
( b ) grey - scale coded phase field @xmath25 at time @xmath171 .
( c ) grey - scale coded vorticity field at the same instant .
( d ) interface profiles at four instants .
( e ) contours of the vorticity at @xmath171 .
contour values are from -20.0 to 20.0 by 4.0 .
the solid and dashed lines correspond to non - negative and negative vorticity values .
, title="fig : " ] the period tripling state of faraday waves , in which the period of the standing wave becomes three times the usual subharmonic period ( which is in turn twice the vibration period ) , is found in the quasi - two - dimensional experiment by @xcite ( see also @xcite ) and in the three - dimensional axisymmetric experiment by @xcite and also by @xcite . in the two - dimensional vortex - sheet simulation of @xcite , period tripling is also observed . from their numerical result ,
@xcite conclude that the period tripling is caused by the nonlinearity of the irrotational flow .
we here aim at reproducing the period tripling state with the phase - field modelling within our feasible atwood number range @xmath179 , although the above experiments and the vortex - sheet simulation are conducted in the case @xmath180 . by searching the parameter space , which will be discussed later ,
indeed we observe the period tripling state at low atwood number @xmath181 , as depicted in figure [ f : pt_fixed_x ] .
the interface shapes at the maxima are shown in figure [ f:3modes ] , some of which actually differ from the previous observations . in the experiment by @xcite , the interface shapes at the three peaks in the tripling state
are classified as sharp crest ( called mode a by them ) , flat or dimpled crest ( mode b ) and round crest ( mode c ) , respectively .
the corresponding shapes of the phase - field simulation shown in figure [ f:3modes](a c ) are qualitatively different except for the round crest ( figure [ f:3modes](c ) ) . in particular ,
the hourglass plume shown in figure [ f:3modes](a ) and waistless plume in figure [ f:3modes](b ) are different from the sharp crest ( mode a ) and the flat or dimpled crest ( mode b ) .
these sharp and flat crests are associated with wave breaking in @xcite but not necessarily so in @xcite . in the phase - field simulation with this parameter
set , wave breaking does not occur . in the period tripling state
observed numerically by @xcite , plume - type interfaces are not observed .
the vorticity contours at the three peaks are shown in figure [ f:3modes](d f ) .
the number of vortex peaks inside the interface region is three for the hourglass plume and one for the waistless plume and the round crest . outside of the interface region
the vorticity is weaker but not zero . to find the parameters of the period tripling state in practice , we use as a guide the phase diagram made by @xcite in the space of the detuning parameter @xmath1 and forcing parameter @xmath182 .
we search this @xmath183-space horizontally ( constant @xmath1 ) by going away from the neutral curve as shown in figure [ f : pqdiagram ] .
we take seven parameter sets along this line .
the period tripling state is found at the middle point . in the phase diagram in @xcite , the period tripling state extends for the large @xmath182 region .
but our case is very different .
the parameters of the tripling state are : @xmath184 m , @xmath106 m , @xmath185 , @xmath186 s@xmath103 , @xmath97 kg m@xmath98 , @xmath99 kg m@xmath98 , @xmath100 pa s , @xmath101 pa s , @xmath102 n m@xmath103 , @xmath104 m s@xmath105 , @xmath187 m , @xmath118 , @xmath119 and @xmath126 .
the spatial resolution is determined in the same way as described in subsection [ ss : pf ] .
now we discuss the behaviours at other parameter points shown in figure [ f : pqdiagram ] . for three @xmath182 values smaller than the period tripling value , we observe that the period of the standing wave remains twice the vibration period ( the subharmonic period ) .
the interface shape at the maximum and minimum surface elevations are hourglass - plume type .
in other words , the temporal symmetry is kept as we approach the period tripling state from the left while in the experiment and numerical simulation by @xcite the symmetry is broken before reaching the tripling state .
this discrepancy may be due to the difference in the atwood number or the difference in the dissipation caused by the sidewalls .
for the next larger @xmath182 parameter than that of the tripling state , we observe that the period of the standing wave returns to the subharmonic period ( @xmath188 ) .
we do not have an explanation for this .
the interface shape at the maximum and minimum elevation is also of the hourglass - plume type .
for the two largest @xmath182 parameters that we calculate , the oscillation becomes close to quasi - periodic .
the interface shapes at relative maximum elevations in this parameter range take the three forms shown in figure [ f:3modes](a - c ) almost randomly .
, showing the period tripling . here
@xmath127 is the period of the vibration forcing . ,
width=288,height=192 ] ; ( b ) @xmath189 ; ( c ) @xmath190 .
they correspond to the last three peaks shown in figure [ f : pt_fixed_x ] .
vorticity contours ( solid and dotted lines correspond to non - negative and negative values , respectively ) at the same instances : ( d ) @xmath191 , the contour values are from @xmath192 to @xmath193 by @xmath143 ; ( e ) @xmath189 , the values are from @xmath194 to @xmath195 by @xmath143 ; ( f ) @xmath190 , the values are the same as ( e ) . , title="fig : " ] ; ( b ) @xmath189 ; ( c ) @xmath190 .
they correspond to the last three peaks shown in figure [ f : pt_fixed_x ] .
vorticity contours ( solid and dotted lines correspond to non - negative and negative values , respectively ) at the same instances : ( d ) @xmath191 , the contour values are from @xmath192 to @xmath193 by @xmath143 ; ( e ) @xmath189 , the values are from @xmath194 to @xmath195 by @xmath143 ; ( f ) @xmath190 , the values are the same as ( e ) .
, title="fig : " ] ; ( b ) @xmath189 ; ( c ) @xmath190 .
they correspond to the last three peaks shown in figure [ f : pt_fixed_x ] .
vorticity contours ( solid and dotted lines correspond to non - negative and negative values , respectively ) at the same instances : ( d ) @xmath191 , the contour values are from @xmath192 to @xmath193 by @xmath143 ; ( e ) @xmath189 , the values are from @xmath194 to @xmath195 by @xmath143 ; ( f ) @xmath190 , the values are the same as ( e ) .
, title="fig : " ] ; ( b ) @xmath189 ; ( c ) @xmath190 .
they correspond to the last three peaks shown in figure [ f : pt_fixed_x ] .
vorticity contours ( solid and dotted lines correspond to non - negative and negative values , respectively ) at the same instances : ( d ) @xmath191 , the contour values are from @xmath192 to @xmath193 by @xmath143 ; ( e ) @xmath189 , the values are from @xmath194 to @xmath195 by @xmath143 ; ( f ) @xmath190 , the values are the same as ( e ) . , title="fig : " ] ; ( b ) @xmath189 ; ( c ) @xmath190 .
they correspond to the last three peaks shown in figure [ f : pt_fixed_x ] .
vorticity contours ( solid and dotted lines correspond to non - negative and negative values , respectively ) at the same instances : ( d ) @xmath191 , the contour values are from @xmath192 to @xmath193 by @xmath143 ; ( e ) @xmath189 , the values are from @xmath194 to @xmath195 by @xmath143 ; ( f ) @xmath190 , the values are the same as ( e ) . , title="fig : " ] ; ( b ) @xmath189 ; ( c ) @xmath190 .
they correspond to the last three peaks shown in figure [ f : pt_fixed_x ] .
vorticity contours ( solid and dotted lines correspond to non - negative and negative values , respectively ) at the same instances : ( d ) @xmath191 , the contour values are from @xmath192 to @xmath193 by @xmath143 ; ( e ) @xmath189 , the values are from @xmath194 to @xmath195 by @xmath143 ; ( f ) @xmath190 , the values are the same as ( e ) . , title="fig : " ] parameter space for seeking the period tripling regime . here
@xmath196 is the detuning parameter and @xmath197 is the forcing parameter , where @xmath198 $ ] .
the points are @xmath199 and constant @xmath200 .
the curve corresponds to the neutral stability line obtained by the linear analysis of @xcite ( the method described in [ s : linear ] is used ) .
, width=288,height=192 ]
we have applied the phase - field modelling of the binary fluids with the cahn - hilliard equation due to @xcite to numerical simulations of the faraday wave problem in two spatial dimensions .
here we have solved the navier - stokes equations for both the top and bottom fluids with rigid boundary conditions on the top and bottom walls and with periodic boundary condition for the side walls .
validation of this phase - field simulation is checked quantitatively in the linear regime and qualitatively in the nonlinear regime . in the nonlinear regime , we are not able to validate the present simulation against experimental results unlike @xcite .
nevertheless we have considered two cases . both cases involve plume formation where the two - fluid interface becomes a multivalued function of a horizontal coordinate .
such situations can be a good testing ground for the phase - field method since the method does not break down when the interface turns over .
the nonlinear case we considered first is the bursting plume formation which was studied by @xcite with the vortex - sheet method .
their system setting ( the boundary conditions , the dissipation and close - to - unity atwood number , especially ) is very different from the present binary fluid setting .
consequently the comparison is necessarily qualitative .
the phase - field method with the numerical scheme described in section 3 works only in the atwood number range @xmath179 .
nevertheless we found for very low atwood number @xmath154 that a similar plume is formed and the temporal variation of the interface agrees qualitatively with that in @xcite .
the second nonlinear case is the period tripling of the two - dimensional faraday waves first found experimentally by @xcite and numerically by @xcite .
again , in spite of the rather big system differences with these studies , we have observed the period tripling state in the phase - field simulation with low atwood number @xmath201 .
this result adds further evidence for the robustness of the period tripling as discussed in @xcite .
more detailed numerical study of the tripling state is currently under way and will be reported elsewhere
. it would be interesting to perform a laboratory experiment with the same physical parameters as in
[ s : pt ] to check whether or not the period tripling is observed . concerning the limitation on the atwood number @xmath179 in the present phase - field simulation
, we do not have an explanation for why it is so .
for @xmath202 , we observe that , if we put a larger number of grid points inside the interface region , we can postpone the breakdown of the simulation until somewhat later ( we can continue the simulation for a larger number of time steps ) . however , increasing the number of grid points there may cause a new problem . although @xcite successfully reaches atwood number @xmath203 , he reports that the interface becomes much wider than in reality .
it is not clear to us for the present whether or not we can overcome these problems by improving the temporal and spatial discretizations of the cahn - hilliard and navier - stokes equations . finally , we comment on extension of the phase - field method to three - dimensional faraday waves .
it is straightforward .
we believe that the three - dimensional phase - field method , once extended , can be complementary to the simulation method by @xcite since the two methods are different .
there is an interesting instance of three - dimensional faraday waves , where the interface becomes multivalued .
that is the propagating solitary state observed by @xcite .
our simulation code can easily study such a state .
we acknowledge the support by the grant - in - aid for the global coe program `` the next generation of physics , spun from universality and emergence '' from the mext of japan and by the grant - in - aid for young scientists ( b ) no .
2174290 from the jsps .
we thank the anonymous referees for constructive criticisms .
20 anderson , d. m. , mcfadden , g. b. & wheeler , a. a. 2000 diffuse - interface methods in fluid mechanics .
fluid mech .
_ * 30 * , 139165 .
arbell , h. & fineberg , j. 1999 temporally harmonic oscillons in newtonian fluids .
lett . _ * 85 * , 756759 celani , a. , mazzino , a. , muratore - ginanneschi , p. & vozella , l. 2009 phase - field model for the rayleigh - taylor instability of immiscible fluids .
_ j. fluid mech . _ * 622 * , 115134 .
chen , p. & vials , j. 1999 amplitude equation and pattern selection in faraday waves .
e. _ * 60 * , 559570 .
chen , p. & wu , k. 2000 subcritical bifurcations and nonlinear balloons in faraday waves .
_ * 85 * , 38133816 .
das , s. p. & hopfinger , e. j. 2008 parametrically forced gravity waves in a circular cylinder and finite - time singularity .
_ j. fluid mech . _ * 599 * , 205228 .
eyre , d. j. 1990 an unconditionally stable one - step scheme for gradient systems .
faraday , m. 1831 on a peculiar class of acoustical figures ; and on certain forms assumed by groups of particles upon vibrating elastic surfaces .
* 121 * , 299340 .
see also http://openlibrary.org/books/ol6976494m/experimental_researches_in_chemistry_and_physics , which is a scan of the book : faraday , m. 1859 _ experimental researches in chemistry and physics .
_ r. taylor and w. francis , london .
jacqmin , d. 1999 calculation of two - phase navier - stokes flows using phase - field modeling .
_ j. comput .
phys . _ * 155 * , 96127 .
jacqmin , d. 2000 contact - line dynamics of a diffuse fluid interface .
_ j. fluid mech . _ * 402 * , 5788 .
jiang , l. , perlin , m. & schultz , w. w. 1998 period tripling and energy dissipation of breaking standing waves .
_ j. fluid mech .
_ * 369 * , 273299 .
jiang , l. , ting , c. , perlin , m. & schultz , w. w. 1996 moderate and steep faraday waves : instabilities , modulation and temporal asymmetries .
_ j. fluid mech . _ * 329 * , 275307 .
kityk , a. v. , embs , j. , mekhonoshin , v. v. & wagner , c. 2005 spatiotemporal characterization of interfacial faraday waves by means of a light absorption technique .
e. _ * 72 * , 036209 .
kumar , k. & tuckerman , l. s. 1994 parametric instability of the interface between two fluids .
_ j. fluid mech . _ * 279 * , 4968 .
lioubashevski , o. , arbell , h. & fineberg , j. 1996 dissipative solitary states in driven surface waves .
lett . _ * 76 * , 39593962 .
miles , j. & henderson , d. 1990 parametrically forced surface waves .
fluid mech . _
* 22 * , 143165 .
murakami , y. & chikano , m. 2001 two - dimensional direct numerical simulation of parametrically excited surface waves in viscous fluid .
fluids _ * 13 * , 6574 .
perlin , m. & schultz , w. w. 2000 capillary effects on surface waves .
fluid mech . _ * 31 * , 241274 .
prinet , n. , juric , d. & tuckerman , l. s. 2009 numerical simulation of faraday waves .
_ j. fluid mech . _ * 635 * , 126 .
peskin , c. s. 1977 numerical analysis of blood flow in the heart .
_ j. comput .
* 25 * , 220252 .
pucci , g. , fort , e. , ben amar , m. & couder , y. 2011 mutual adaptation of a faraday instability pattern with its flexible boundaries in floating fluid drops .
lett . _ * 106 * , 024503 .
saad , y. 2003 _ iterative methods for sparse linear systems_. siam publishing .
shani , i. , cohen , g. , & fineberg , j. 2010 localized instability on the route to disorder in faraday waves .
lett . _ * 104 * , 184507 .
ubal , s. , giavedoni , m. d. & saita , f. a. 2003 a numerical analysis of the influence of the liquid depth on two - dimensional faraday waves .
fluids _ * 15 * , 30993113 .
wright , j. , yon , s. & pozrikidis , c. 2000 numerical studies of two - dimensional faraday oscillations of inviscid fluids .
_ j. fluid mech . _ * 400 * , 132 .
|
a fully nonlinear numerical simulation of two - dimensional faraday waves between two incompressible and immiscible fluids is performed by adopting the phase - field method with the cahn - hilliard equation due to jacqmin ( _ j .
comput .
phys .
_ , vol . 155 , 1999 , pp . 96 - 127 ) .
its validation is checked against the linear theory . in the nonlinear regime ,
qualitative comparison is made with an earlier vortex - sheet simulation of two dimensional faraday waves by wright , yon & pozrikidis ( _ j .
fluid mech .
_ , vol .
400 , 2000 , pp .
1 - 32 ) .
the vorticity outside the interface region is studied in this comparison .
the period tripling state , which is observed in the quasi - two dimensional experiment by jiang , perlin & schultz ( _ j .
fluid mech .
_ , vol .
369 , 1998 , pp . 273 - 299 ) , is successfully simulated with the present phase - field method .
faraday waves , multiphase flow , parameteric instability
|
after a long history in which a large number of experimental groups around the world contributed to the development of succesful methods to master stabilization and cooling of dilute bose gases , last year the aim of achieving bose - einstein condensation in such a system was finally reached .
indeed , a macroscopic occupation of the one - particle ground state was irrefutably observed in magnetically trapped and evaporatively cooled alkali gas samples of @xmath1rb and @xmath2na using relatively simple time - of - flight measurements @xcite .
the transition that was claimed to be seen in an experiment using @xmath3li was less convincing @xcite . in the latter case ,
the interatomic interaction is effectively attractive and the potential has a negative scattering length @xmath4 .
therefore , bose - einstein condensation in this system is preempted by a first order phase transition to a liquid or solid phase in the homogeneous case @xcite .
nevertheless , for inhomogeneous gas samples the trapping potential has a stabilizing influence and a macroscopic occupation of the ground state is possible in principle .
however , when the condensate contains more than some 1500 particles under the conditions of the @xmath3li experiment @xcite , the condensate is still expected to collapse @xcite .
after these first experiments , which were primarily aimed at proving the existence of a bose condensate , many experimental groups are now building or improving on their experimental setups to be able to perform much more precise measurements of various interesting properties of the gas in the degenerate regime .
superfluidity @xcite , the condensate density and its profile , the dynamics of condensate formation @xcite , the josephson effect @xcite , vortex dynamics , collective excitations @xcite and the precise value of the critical temperature are examples of phenomena and quantities of interest .
other types of experiments will presumably also study the properties of mixtures of atomic gases . in this respect
one might think of two bosonic species with a different sign of the scattering length , or mixtures of bosons and fermions , or mixtures uniting both aspects . in the case of a pure fermionic gas of @xmath5li atoms , a bcs transition to a superfluid state
is predicted to occur and should be within reach of the current experimental technology @xcite .
furthermore , in some cases the magnitude and even the sign of the scattering length can be changed by varying the applied magnetic field .
this opens the road to yet another type of experiment .
it is clear from these possibilities that a large number of experiments are expected to be performed in the near future , which clearly makes the degenerate dilute bose gas also a very interesting subject for theoretical studies .
indeed , this field of research has rapidly expanded during the last year . however , most approaches to the dilute bose gas use the bogoliubov ( or popov ) theory and are therefore of mean - field type and susceptible to improvements , both from a practical as well as a fundamental point of view . in these approaches
one mostly uses the so - called two - body @xmath6matrix , or equivalently the scattering length @xmath4 .
technically , this important quantity describes the collisions taking place in the dilute bose gas by summing all possible two - body scattering processes , i.e. all ladder diagrams , without taking into account the fact that the surrounding gaseous medium has an effect on these collisions .
however , we have recently shown that the many - body corrections arising from the surrounding gas may be important @xcite , and are even essential for solving the problem connected to the order of the phase transition which is found to be of first order when using the two - body @xmath6matrix @xcite . including quantatively the same many - body corrections in the case of a highly inhomogeneous gas sample
has at this point not yet been done . moreover , introducing the effect of the medium on two - particle collisions also in the condensed phase by means of the many - body @xmath6matrix leads to fundamental problems if we want to describe the physics at long wavelengths correctly as the ladder diagrams contain infrared divergencies in this case .
using renormalization group techniques , we expect in principle to be capable of resolving these infrared problems as with this method a correct resummation of diagrams is automatically performed , eliminating any potentially troublesome large distance behavior of the individual diagrams .
furthermore , a renormalization group calculation can be used to improve the usual mean - field approaches and the many - body @xmath6matrix theory in the critical region .
indeed , we recently predicted by these means for example that the critical temperature in the @xmath1rb and @xmath2na experiments can , due to interaction effects only , be raised with as much as 10 % compared to the ideal gas value found from the criterion @xmath7 @xcite , which is identical to the criterion also found using mean - field calculations @xcite . here
@xmath8 is the density and @xmath9 the thermal de broglie wavelength of the atoms in the gas .
therefore , it is very well conceivable that in the critical region also other properties of the dilute bose gas , such as the superfluid and condensate densities , will significantly change when going beyond the mean - field level or beyond the many - body @xmath6matrix theory .
the renormalization group method is a very powerful method which was in first instance developed by wilson @xcite to study the universal properties of second order phase transitions .
the basic idea is to perform the trace in the grand canonical partition function @xmath10 gradually , starting with the high momentum states . after each step one tries to find a new effective hamiltonian such that @xmath11 and the trace is limited to the low momentum states which have not been reached yet .
one proceeds until the complete sum has been performed . besides the partition function one in this manner also ends up with the effective hamiltonian describing the long distance properties of the system .
the renormalization group method has been applied to the dilute bose gas @xcite before , but without actually performing an extensive quantitative study of this system .
this is due to the fact that in general the quantities of nonuniversal nature , such as the critical temperature and the superfluid and condensate densities , depend on the microscopic details of the system considered .
put differently , they depend on the ultraviolet cutoff @xmath12 of the theory and this quantity is usually unknown .
however , due to the diluteness of the gas the nonuniversal properties are in the present and forthcoming experiments the most interesting ones , and therefore we are in this paper mainly interested in these aspects .
the reason that it is nevertheless possible to perform a quantitative study of the dilute bose gas using the renormalization group method , is that for this system we do have sufficient information about the microscopic details to calculate and predict the nonuniversal properties .
thus we can , by correctly applying this knowledge , use the renormalization group method and eliminate the cutoff dependence at the same time .
we will come back to this point in sec .
ii . furthermore
, we will show that , in contrast to regular perturbation theory , the problems related to the infrared divergencies are in principle indeed resolved , but lead nevertheless to some problems whose solution requires further investigation . however , these problems are only of importance when the interaction energy is no longer negligible compared to the kinetic energy of the particles .
the dimensionless parameter reflecting this aspect is @xmath13 .
therefore , we will in this paper first concentrate on the regime where @xmath0 .
we treat here only the homogeneous bose gas with effectively repulsive interactions , i.e. with a positive scattering length .
however , as in all experiments up till now the number of particles @xmath14 is so large that the critical temperature @xmath15 is much larger than the energy splitting @xmath16 between subsequent levels , one can practically for all temperatures use a local density approximation to describe the gas in the trap .
the criterion for this description to be valid is that the correlation length should be smaller than the typical length scale on which the atomic density varies .
therefore , a local density approximation breaks down close to the edge of the gas cloud , which is for most practical purposes an unimportant region , but also in the center of the trap if the temperature approaches the critical temperature and the diverging correlation length starts to exceed the typical dimensions of the trap . the temperature interval where this occurs has a width of @xmath17 around the critical temperature . as this region is very small ,
we conclude that the results we find in this paper for the homogeneous gas are essentially also valid for the inner part of the trapped bose gases , and in particular pertain to the @xmath1rb and @xmath2na experiments .
furthermore , we want to remark here that it would in principle also be possible to set up a renormalization group calculation for the inhomogeneous bose gas . of course , there is no real second order phase transition present in this system because the correlation length @xmath18 can never become infinite , but the techniques of renormalization group as presented in this article can still be used to calculate the ( nonuniversal ) properties also in this case .
the paper is organized as follows . in sec .
ii we briefly discuss the theoretical description of the dilute bose gas and the renormalization group method . in sec .
iii we first apply the renormalization group to the uncondensed phase of the bose gas because the flow equations are relatively simple and easy to understand in this case . in sec .
iv we then go over to the degenerate bose gas and again describe the gas properties following from the renormalization group approach . finally , in sec .
v we end with some concluding remarks . in the numerical calculations we always use @xmath2na as an example , because the experiment with this atomic species is closest to the conditions of homogeneity @xcite .
we take in these cases the most up - to - date value of @xmath19 for the two - body scattering length @xcite .
the renormalization group equations are most easily obtained using the functional integral formulation of the grand canonical partition function @xcite . we thus write @xmath20d[\psi]\exp\left\{-\frac{1}{\hbar}s[\psi^*,\psi]\right\}~.\ ] ] this functional integral is over @xmath21-number fields @xmath22 and @xmath23 periodic in imaginary time over @xmath24 .
the so - called euclidian action for the dilute bose gas is given by @xmath25&=&\int_{0}^{\hbar\beta}d\tau \left ( \int d{\bf x}~ \psi^*({\bf x},\tau)\left [ \hbar\frac{\partial}{\partial\tau}- \frac{\hbar^2{\bf \nabla}^2}{2m}-\mu \right ] \psi({\bf x},\tau)\right .
\nonumber \\ & & + \left.\frac{1}{2}\int d{\bf x}\int d{\bf x'}~\psi^*({\bf x},\tau)\psi^*({\bf x'},\tau)v({\bf x}-{\bf x'})\psi({\bf x'},\tau)\psi({\bf x},\tau ) \right)~,\end{aligned}\ ] ] with @xmath26 the chemical potential and @xmath27 the effectively repulsive interaction potential . in principle , the action also contains a term describing three - particle interactions , i.e.@xmath28 and terms describing interactions between four and more particles due to the finite extent of the electron clouds of the atoms .
however , as we are describing the _ dilute _ bose gas these terms are expected to make in general no significant contribution to the thermodynamic properties to be calculated .
indeed , in usual approaches to the dilute bose gas these terms are always neglected for this very reason .
one aspect of the renormalization group calculation is that it is possible , and even rather straightforward , to include the three - body interaction in the calculations . as its influence
will turn out to be extremely small , except in the critical region where it becomes somewhat larger , we will in the rest of this section omit this term for reasons of brevity .
however , in sec .
iii we perform some calculations including the three - body term to show its effect quantatively , and at that point introduce it again . expanding the fields in fourier modes through @xmath29 and the complex conjugate expression for @xmath22
, we can write the action in momentum space as @xmath30=\sum_{{\bf k},n}(-i\hbar\omega_n+\epsilon_{{\bf k}}-\mu ) a^*_{{\bf k},n}a_{{\bf k},n}+\frac{1}{2}\frac{1}{\hbar\beta v } \sum_{\stackrel{\bf k , k',q}{n , n',m } } v_{\bf q}a^*_{{\bf k+q},n+m}a^*_{{\bf k'-q},n'-m } a_{{\bf k'},n'}a_{{\bf k},n}~.\end{aligned}\ ] ] in this equation @xmath31 is the kinetic energy , @xmath32 is the fourier transform of the interaction potential , @xmath33 is volume of the system and the bosonic matsubara frequencies @xmath34 reflect the periodicity of the fields in imaginary time .
the renormalization group equations now follow from repeatedly applying the renormalization group transformation to this action .
it consists of three different stages @xcite which we summarize here .
the first step is to perform the functional integral in eq .
( 1 ) for the most rapidly oscillating fields . in other words ,
we eliminate the highest fourier modes from the action in eq .
( [ basisactie ] ) . in practice
, we therefore have to split the fields in ` slow modes ' and ` fast modes ' , i.e. we define @xmath35 where @xmath12 is the ultraviolet cutoff in the theory , and d@xmath12 is the thickness of the shell in momentum space which is going to be integrated out .
we will dwell on the issue of the ultraviolet cutoff shortly . with the definitions in eq .
( [ modesplitting ] ) , we can write the action as @xmath25=s_0[\psi^*_<,\psi_<]+s_0[\psi^*_>,\psi_>]+s_i[\psi^*_<,\psi _ < , \psi^*_>,\psi _ > ] , \end{aligned}\ ] ] where @xmath36 is quadratic in the fields , and @xmath37 contains the remaining terms , and in particular the terms that couple the slow and fast modes .
the partition function can now be written as @xmath38d[\psi_<]\exp\left\{-\frac{1}{\hbar}s_0[\psi^*_<,\psi_<]\right\ } \times\int d[\psi^*_>]d[\psi_>]\exp\left\{-\frac{1}{\hbar}(s_0[\psi^*_>,\psi _ > ] + s_i[\psi^*_<,\psi_<,\psi^*_>,\psi_>]\right\}\nonumber \\ & & \equiv\int d[\psi^*_<]d[\psi_<]\exp\left\{-\frac{1}{\hbar}s_0[\psi^ * _ < , \psi_<]\right\}\times\exp \left\{-\frac{1}{\hbar}\delta s[\psi^*_<,\psi _ < ] \right\}~.\end{aligned}\ ] ] after performing the last integral , we obtain an expression for the partition function containing the new effective action @xmath39=s_0 [ \psi^*_<,\psi_<]+\delta s[\psi^*_<,\psi_<]$ ] for the slow modes .
the effect of this mode elimination is to change the value of the existing coupling constants and to introduce new vertices in the original action .
the discussion of the rules on how we can obtain the equations describing this change of the coupling constants in general will be deferred to secs .
iii and iv where we will also apply these rules to find the flow equations for the bose system .
the two remaining stages of the renormalization group transformation consist of a rescaling of the momenta such that the new cutoff , which is @xmath40 , is restored to its initial value @xmath12 , and a rescaling of the frequency and fields in the action such that there is no effect from this momentum rescaling on some appropriate terms in the quadratic part of the action .
if we neglect the renormalizations from the first step , this procedure yields the so - called trivial scaling of the coupling constants and reveals the relevance of the various vertices as it tells whether the value of the corresponding coupling constant grows ( termed relevant ) , remains the same ( termed marginal ) , or shrinks ( termed irrelevant ) as we scale the ultraviolet cutoff . having completed these final steps of the renormalization group transformation , we can read off the new coupling constants from the action .
thus , the renormalization group equations have two contributions .
one is from the rescaling , the other is from the actual integrating out of high momenta from the action . concerning the trivial scaling , there are two cases to distinguish .
for the normal renormalization group equations this scaling is found using the full quadratic action .
the other situation occurs exactly at the critical temperature . at that point the correlation length @xmath18 of the system
is infinite , and the system looks the same on every length scale .
hence , the coupling constants remain the same when performing the renormalization group transformation and we are at a fixed point . to find this fixed point from the set of renormalization group equations
, we generally have to use a different scaling from the one found above , and this is the scaling at the critical temperature .
it can easily be obtained by realizing that the correlation time @xmath41 also diverges as we approach the critical temperature , behaving like @xmath42 . at temperatures such that @xmath43 , we can neglect the time dependence in the action and the @xmath44term in the quadratic part of the action is unimportant . as a result , we then only need to consider the kinetic energy term to find the trivial scalings .
this yields the scaling at the critical temperature . in renormalization group studies ,
one normally restricts oneself to couplings which are relevant or marginal at the critical temperature , and this we will in first instance also do .
however , it turns out that even the marginal coupling constant @xmath45 from the three - body interaction term is quite irrelevant to the nonuniversal properties to be calculated .
this is of course due to the fact that three - body interactions are unimportant for a _ dilute _ system .
however , if one is interested in the universal properties of the phase transition , such as e.g.the critical exponent @xmath46 connected to the divergence of the correlation length when approaching the critical temperature , marginal and also irrelevant couplings can have a considerable effect .
we will explicitly encounter this fact in the next section .
also the position of the fixed point is changed when including irrelevant couplings .
the renormalization group equations describe the change of the coupling constants as we integrate out momentum shells from the action .
having arrived at the long wavelength effective action , we find whether or not we are in the condensed phase , and we can calculate universal properties connected to this phase transition .
however , our aim is to use the renormalization group method to obtain information on other properties as well . in particular
, we want to calculate the equation of state , the superfluid density , and the pressure of the gas .
this can also straightforwardly be done by noting that e.g. for the total density we have @xmath47 .
this equation can be cast into a differential equation describing the building up of the density as we proceed with the elimination of the fast fourier modes . for that purpose
, we have to use the right value of the chemical potential @xmath26 in each subsequent momentum shell as found from the renormalization group equations . in the same fashion
one can also determine e.g. the superfluid density and the thermodynamic potential . before we can start with the derivation of the flow equations
, we first have to pay some attention to the high momentum limit of the action . in principle
, there is no real sharp ultraviolet cutoff @xmath12 in our problem .
however , the typical behavior of the fourier transform of the two - body interaction potential , depicted in fig . 1 , is such that there is an effective ultraviolet cutoff around the momentum scale set by the scattering length @xmath4 of this potential . below this value ,
the fourier transform is practically momentum independent and equal to @xmath48 . as in the bose systems considered here and realized
experimentally the temperatures are so low that @xmath49 , the particles in the gas reside in a momentum range well below this ultraviolet cutoff .
thus , we can represent the interaction potential by the momentum independent value @xmath48 for momenta below a cutoff @xmath50 of @xmath51 , and zero for larger momenta .
modelling the potential as such implies that the nonuniversal properties we find from a renormalization group calculation will be sensitive to the specific value of the cutoff @xmath12 taken in the calculations .
however , at this point our knowledge about the microscopic details of the bose gas comes in to resolve this potential problem .
in particular , we know that the two - body interaction potential @xmath52 has to renormalize to the two - body @xmath6matrix @xmath53 when we include all possible two - body scattering processes in the vacuum @xcite .
the two - body @xmath6matrix roughly has the same momentum dependence as @xmath54 , cf .
fig . 1 , and is in particular constant and equal to @xmath55 in the range of thermal momenta and energies .
thus , given an ultraviolet cutoff @xmath12 we can fix the renormalization group equations by demanding that for the two - body problem , @xmath48 indeed correctly renormalizes to @xmath56 .
since this value is , due to the inequality @xmath57 , already attained before we enter the thermal regime as we integrate out more and more momentum shells from the action , this indeed leads to a correct description of the properties of the bose gas which is independent of the ultraviolet cutoff @xmath12 .
having eliminated the cutoff dependence , we are then in a position to determine also the nonuniversal properties of the dilute bose gas .
furthermore , we can perform the calculation for any ( positive ) value of the scattering length , thus being able to describe any atomic species with effectively repulsive @xmath58wave scattering .
the results we find are therefore relevant to the current experiments using @xmath1rb and @xmath2na , but also to future experiments using atomic hydrogen or other atoms with a positive scattering length .
we now turn to the application of the renormalization group method to the dilute bose gas .
first , we derive the renormalization group equations when the chemical potential is negative , starting from the action in eq.([basisactie ] ) .
these equations however , do not describe the bose condensed phase for then it is required that the chemical potential be positive . as a result
, we have to rederive the renormalization group equations for that case and this is carried out in sec .
the derivation is now much more involved as the space and time independent part of the effective action , i.e. @xmath59 , has for @xmath60 a mexican hat shape and we first have to expand the action around the correct extremum by performing the shift @xmath61 and introducing the condensate density @xmath62 . only after that
can we proceed to find the contributions to the renormalization of the vertices .
in this section we first concentrate on the renormalization group equations valid for negative chemical potential .
the reasons for this are threefold .
the equations correctly describe the bose gas in the uncondensed phase , and they offer an easy way to determine the influence of three - body effects on the quantities of interest .
moreover , it is best to start with this relatively simple set of equations because it illuminates most clearly our procedure to eliminate the cutoff dependence of the theory . to calculate the change of the couplings after each step of the renormalization group transformation we can technically proceed in two different , but equivalent ways .
the first and probably most familiar method is to expand the integrand in the partition function in powers of @xmath63 $ ] , and then perform the integrals over the fast modes by evaluating the appropriate feynman diagrams contributing to the renormalization or flow of the vertices of interest .
the renormalization of the gradient and time - derivative terms of interest are found by performing a taylor expansion in external momenta and frequencies of these diagrams .
however , the associated couplings become more and more irrelevant for higher order terms in this expansion .
the integration over the internal momenta of the diagrams is restricted to the afore - mentioned momentum shell , which can in principle be chosen to have any thickness d@xmath12 .
taking it infinitesimally small , possible in the thermodynamic limit only , leads to a continuous renormalization group transformation described by a coupled set of first order differential equations .
a much more important reason to take d@xmath12 infinitesimal is that in this case the type of diagrams which contribute to the renormalization are the so - called one - loop diagrams only .
this is a point which is not always explicit in the literature , but it follows from the fact that each extra loop , i.e. each extra momentum integral , introduces an extra factor of d@xmath12 @xcite , and therefore will vanish when d@xmath12 is infinitesimal .
clearly , this is not so when the thickness d@xmath12 has a finite value .
note furthermore that the above holds irrespective of the magnitude of the interactions , in contrast to what is sometimes mentioned in the literature @xcite . in the case of e.g. a large two - body interaction vertex
, the infinitesimal thickness of the momentum shell still ascertaines the validity of a perturbation expansion with the use of one - loop diagrams only .
this most notably implies that in this formulation the restrictions of the renormalization group method lie in the number of vertices considered , and not in the type of diagrams taken into account .
the second way to obtain the renormalization group equations does not explicitly make use of feynman diagrams , and is therefore very useful and efficient when the number of diagrams is large and/or the associated combinatorics is complicated .
we here use the diagrammatic method as in the unbroken phase the number of diagrams is limited and the combinatorics is simple .
we use the other method in sec .
iv . moreover ,
using feynman graphs also gives a transparent way to find the trivial scaling of the vertices in a somewhat different fashion than from the rescaling procedure described in sec.ii , which brings out more clearly the physics of this procedure .
as mentioned before , we include in first instance also the three - body interaction term written down in the previous section into our considerations , and take it , as the two - body interaction , to be momentum independent below the cutoff @xmath12 .
the one - loop diagrams to be calculated are depicted in fig.2 .
we find that in the thermodynamic limit the total contribution to the chemical potential from integrating out an infinitesimal momentum shell in the hartree and fock diagrams is @xmath64 where @xmath65 denotes the radius of the shell in momentum space .
the ladder and bubble diagrams renormalizing the two - body interaction potential give , together with the diagram containing the three - body term , @xmath66 + \nonumber \end{aligned}\ ] ] @xmath67 and for the vertex @xmath45 we find @xmath68[2n(\epsilon_{\bf k}-\mu)+1]+ \nonumber\end{aligned}\ ] ] @xmath69- \nonumber\end{aligned}\ ] ] @xmath70~.\end{aligned}\ ] ] in these expressions @xmath71 is the bose - einstein distribution function which result from the summation over the matsubara frequencies @xmath72 . to derive the renormalization group equations and the trivial scalings of the coupling constants we focus on the first of these equations renormalizing the chemical potential , and cast it into a differential equation .
using @xmath73 and performing the angular integrals , we can write eq .
( [ mu1loop ] ) as @xmath74 next , we remove all explicit @xmath75dependencies from the bose - einstein distribution function by letting both the temperature and the chemical potential scale with exponent 2 , i.e. we put @xmath76 and @xmath77 . hence , both temperature and chemical potential scale trivially with exponent 2 and we find for the differential equation describing the change of the chemical potential when integrating out a momentum shell @xmath78 finally , we now also absorb the factor @xmath79 into @xmath48 in order to remove the remaining explicit @xmath75dependence . as a result
, @xmath48 scales trivially with exponent @xmath80 . by considering in eq .
( [ vnul1loop ] ) the term containing @xmath45 , we can analogously show that this vertex trivially scales with an exponent @xmath81 .
note that to find the real physical quantities we should always remove the trivial scalings again .
we see that the trivial scaling can , in a very simple way , be found from the one - loop expressions .
moreover , it shows that introducing the trivial scalings does not have an essential effect on the renormalization of the vertices ; it is merely a rewriting of the differential equations . in the case of a negative chemical potential
we eventually obtain the following coupled set of renormalization group equations for the coupling constants @xmath82 and @xmath45 , @xmath83 @xmath84\right]\nonumber \\ & & + \frac{3\lambda^3}{2\pi^2}u_{\bf 0}n(\epsilon_\lambda-\mu)~,\end{aligned}\ ] ] @xmath85[2n(\epsilon _ \lambda-\mu)+1]+\right . \nonumber \\ & & \left.\frac{3}{(\epsilon_\lambda-\mu)^2}[1 + 2n(\epsilon_\lambda-\mu)+2\beta ( \epsilon_\lambda-\mu)n(\epsilon_\lambda-\mu)[n(\epsilon_\lambda-\mu)+1 ] ] \right]- \nonumber \\ & & \frac{\lambda^3}{2\pi^2}u_{\bf 0}v_{\bf 0}\left[\frac{3(1 + 2n(\epsilon_\lambda -\mu))}{\epsilon_\lambda-\mu}+18\beta n(\epsilon_\lambda-\mu)[n(\epsilon _ \lambda-\mu)+1]\right]~.\end{aligned}\ ] ] to argue that these equations are the only ones we need to consider , we still have to determine the trivial scaling at the critical temperature .
this different scaling comes about because in the limit @xmath86 the bose - einstein distribution function behaves as @xmath87 as we are effectively at very high temperatures due to @xmath88 . to remove again all explicit @xmath75dependencies after substituting this behavior
, we clearly need a different trivial scaling of the vertices , and this is precisely the trivial scaling at the critical temperature , since putting @xmath89 is equivalent to neglecting the time - derivative in the action . in this manner
we can straightforwardly show that the scaling of the chemical potential remains the same , i.e. we have @xmath90 , that the scaling of the two - body interaction becomes @xmath91 instead of @xmath92 , and that the three - body interaction does not scale , i.e. @xmath93 instead of @xmath94 . from this
we conclude that @xmath26 and @xmath48 are relevant and @xmath45 is marginal at the critical temperature .
furthermore , the four - body interaction is indeed irrelevant and therefore not included in the calculations .
the coefficients of the gradient and time derivative terms in the quadratic part of the action are , like @xmath45 , marginal and would in principle also have to be included in the renormalization group equations .
however , as we are in the regime @xmath57 , the interactions are expected to be independent of momentum and energy ( but see below ) and there is no renormalization of the @xmath95 and @xmath96 terms .
moreover , we know from the @xmath97expansion that the anomalous dimension @xmath98 , indicating the importance of the @xmath96 renormalization at the critical temperature , is very small , namely @xmath99 @xcite .
therefore , the @xmath95 and @xmath96 renormalizations will be neglected and eq .
( 13 ) describes the renormalization of the vertices we will consider . to calculate the partition function of the gas
, we need , next to the flow equations for the above quantities , also the correct boundary conditions .
the first one for the chemical potential @xmath26 is just the bare value in the action eq .
( [ basisactie ] ) .
for the two - body interaction potential @xmath48 we need to be more careful , as this vertex has to correctly fix the renormalization group equations as described in the previous section . from eq.([vnulrgboven ] ) we recognize that in a vacuum , i.e.@xmath100 , the renormalization of the interaction between two particles is governed by @xmath101 this is just the differential form of the lippmann - schwinger equation @xcite for the two - body @xmath6matrix at energy @xmath102 , i.e. @xmath103 .
as the two - body @xmath6matrix is energy independent for low energies , we can neglect @xmath26 and use @xmath104 . as the solution of eq.([lipschwing
] ) is also practically independent of the chemical potential , we can there also neglect it . as a result ,
the requirement is now that , given an ultraviolet cutoff @xmath12 , @xmath48 flows for @xmath86 to the value @xmath55 .
this can be ascertained by choosing the right initial condition for @xmath48 , and more precisely we find from analytically solving eq .
( [ lipschwing ] ) that @xmath105 leads to the correct result .
note that we can describe different atomic species by only changing the value of the scattering length @xmath4 used in this equation .
finally , we in principle also need a boundary condition for the three - body interaction @xmath45 . for this interaction we can , analogous to the scattering length @xmath4 for the two - body interaction , introduce a length scale @xmath106 , and again fix @xmath107 such that the renormalization group calculation gives the correct result @xmath108 for elastic three - body scattering in a vacuum .
however , in general not much is known about the microscopic details of the three - body interaction in a dilute bose gas , and in particular about the value of the ` three - body scattering length ' @xmath106 .
but , as can be expected from the fact that the three - body interactions are in the renormalization group sense irrelevant at large momenta , the results are practically insensitive to the boundary value of @xmath45 used , and @xmath107 is hardly of any importance .
this is shown explicitly in sec .
iiib where we analyze the results from the renormalization group approach for @xmath2na .
note , that taking @xmath109 is equivalent to assuming that three - particle scattering is solely due to the sum of pair interactions .
this is a standard approximation in atomic three - body calculations . finally , to describe the dilute bose gas we still need to derive expressions for the total and superfluid density , and the equation giving the thermodynamic potential @xmath110 , and thus the pressure @xmath111 .
to do so , we make use of the following well - known one - loop expressions for the density @xmath112 the superfluid density @xmath113~,\end{aligned}\ ] ] where @xmath114 is the normal density given by the momentum - momentum correlation function , and the thermodynamic potential @xmath115 casting these equations into a differential form by performing the integration shell by shell leads to @xmath116 @xmath117e^{-3l}\end{aligned}\ ] ] @xmath118 where the inverse temperature again scales as @xmath88 and the chemical potential is found from eq .
( [ murgboven ] ) at each step of the integration .
these equations describe the building up of these quantities as we integrate out more and more momentum shells from the action . for convenience
, the explicit @xmath75dependence is not removed from these equations , and they thus immediately yield the physical quantities .
note furthermore that these equations have no influence on the renormalization of @xmath82 , and @xmath45 as they are decoupled from the renormalization group equations ( 13 ) .
we start our analysis by first focussing on the critical properties of eq.(13 ) and in particular on the critical exponent @xmath46 pertaining to the divergence of the correlation length , i.e. the correlation length behaves as @xmath119 when approaching the critical temperature . for that purpose
we have to find the fixed point of the renormalization group equations , linearize the flow equations around this fixed point and identify the largest eigenvalue @xmath120 which is related to this critical exponent via @xmath121 @xcite .
we in first instance omit the three - body interaction , but in a subsequent calculation include it again to determine the influence of this marginal vertex on the value of @xmath46 . only for the set \{@xmath122 } , when @xmath123 , do we perform the calculation of the fixed point explicitly . with the remarks made in sec .
iiia we have that the fixed point is found from @xmath124 @xmath125 yielding @xmath126 . from linearization of the eqs .
( [ fipomu ] ) and ( [ fipovnul ] ) around this value we find for the largest eigenvalue @xmath127 , implying that @xmath128 . repeating the calculation including the equation for @xmath129 , the fixed point
is shifted and the critical exponent is found to be @xmath130 .
thus , we see that the marginal operator @xmath45 has a rather large effect .
moreover , we can conclude from this result that also irrelevant coupling constants must have a considerable effect as it is known , from the @xmath97expansion of the @xmath131model @xcite and from measurements in @xmath132he experiments , that the true critical exponent of the bose gas should have the value @xmath133 .
this discrepancy should be alleviated by including more and more irrelevant vertices .
however , as we are in particular interested in the nonuniversal properties of the dilute bose gas , we now turn to the influence of the three - body interaction term @xmath45 on these quantities .
the influence of this term will of course be largest when we start close to the critical chemical potential , because at the critical point in principle all fluctuations are of importance .
starting with a chemical potential near the critical value leads to a trajectory that almost flows into the fixed point of the renormalization group equations , and the momentum interval in which irrelevant vertices can have a significant contribution to the flow , and also to the building up of the density , is then largest .
the bare chemical potential yielding a flow into the fixed point is positive , and the physical chemical potential ( i.e. with the trivial scaling removed ) renormalizes to zero . for a bare chemical potential larger than this critical value
the flow is no longer defined as at some value of @xmath134 we have that @xmath135 becomes zero and the bose - einstein distribution function diverges .
we will come back to this point later on and restrict ourselves here to the accessible regime , which physically implies that @xmath136 .
the first aspect connected to the three - body interaction concerns the initial value problem for @xmath45 .
indeed , as alluded to before , an explicit calculation shows that changing the boundary condition for @xmath45 from 0 to one corresponding with a ` three - body scattering length ' @xmath137 , which is extremely large in general , changes the total density and the pressure in the system with less than 0.1 % .
thus , the results we obtain are practically insensitive to this boundary condition and henceforth we simply use @xmath138 .
the second aspect we want to consider is the influence of the three - body interaction term itself on the outcome of the renormalization group flow .
this we do by alternatively including and excluding this vertex .
that is , we solve the set @xmath139 and the set @xmath140 and compare the results we find . for that purpose ,
we plot in fig . 3 the @xmath141diagram near the critical density @xmath142 , where the influence of @xmath45 is largest .
from this figure we see that the change in density and pressure is about 1% at maximum . far away from the critical conditions ,
i.e.at large negative chemical potential , the system becomes more and more dilute , and the influence of the @xmath45 term vanishes , consistent with expectations . in principle , we could choose to maintain the three - body interaction term in the renormalization group equations .
however , as its effect is very small we will from now on neglect @xmath45 altogether .
thus , the dilute bose gas , and more in particular the nonuniversal properties we are interested in , is accurately described by only following the renormalization of the chemical potential and the two - body interaction . having concluded this , we restrict ourselves from now on to the coupled eqs .
( [ murgboven ] ) and ( [ vnulrgboven ] ) . however , before we analyze some physical implications of these equations , we want to remark that the dependence on the ultraviolet cutoff @xmath12 is indeed eliminated from the theory .
the influence on e.g. the density can be shown to be completely absent , of course with the limitations that @xmath50 should be larger than the thermal momentum @xmath143 and that @xmath48 is properly renormalized to @xmath55 when we enter the thermal regime .
we are going to compare the results from the renormalization group calculation with known results for the weakly - interacting bose gas as found from the many - body @xmath6matrix theory @xcite .
far from the critical temperature we expect the results of the mean - field and renormalization group calculations to be identical .
however , close to the critical temperature the renormalization group method will clearly deviate from mean - field theory .
we will here focus on the behavior of the effective two - body interaction , and defer the discussions concerning the equation of state and other thermodynamic quantities to sec .
iv when we are able to describe also the condensed phase of the gas . in the many - body @xmath6matrix theory ,
the chemical potential is renormalized to @xmath144 @xcite .
this is one of the results of including all two - body scattering processes , but also incorporating the effect of the medium on the scattering . including the latter effect on the collisions in the gas , which was first carried out explicitly in refs.@xcite , @xcite and @xcite , is an important step forward in the correct mean - field treatment of the dilute bose gas , since including many - body effects causes the effective interaction to go to zero at the critical temperature .
this resolves a number of problems found in previous approaches using just the two - body scattering length @xmath4 @xcite . with our renormalization group approach
we can corroborate this result and even go somewhat further than that . in our previous work we included only the many - body effects coming from the ladder diagrams .
however , a class of diagrams that in principle also affects the two - body interaction are the bubble diagrams . with our current set of renormalization group equations we can precisely study the effect of these bubble diagrams on the effective interaction .
this is straightforward because we can pinpoint the ladder and bubble contributions in the equation describing the renormalization of the interaction . by alternatively including and excluding the bubble diagrams and then solving the renormalization group equations we can study the relative importance of the bubble diagrams on the effective ` many - body ' scattering length @xmath145 . to avoid any confusion we will adopt the following notation for the scattering length .
the normal ` bare ' two - body scattering length as found e.g. from analysis of the appropriate association spectra is denoted by @xmath4 , as usual .
the effective scattering length @xmath145 includes also effects of the medium on two - body scattering , and therefore depends on the specific approximation used to calculate this effect .
here we consider two such approximations and denote the corresponding scattering lengths by @xmath146 and @xmath147 . in the first case it is the result of a many - body @xmath6matrix calculation and
is defined through @xmath148 . in the second case
it is the result of a renormalization group calculation and is defined through @xmath149 , irrespective of the fact if bubbles are or are not included in this calculation . in fig .
4 we depict the ratio of the effective scattering length resulting from the renormalization group approach to the simple two - body scattering length @xmath4 when we include and exclude the bubble diagrams , as a function of @xmath150 and at a density of @xmath151 @xmath152 .
we conclude that the effect of the bubbles can be rather large and is in particular important near the critical temperature . further away from the critical temperature the importance of the bubbles rapidly decreases .
this implies that calculating the influence of many - body effects by means of the ladder diagrams ( i.e. doing the full many - body @xmath6matrix calculation ) can quantatively give a poor estimate for @xmath145 .
however , qualitatively there is clearly good agreement as we find from the renormalization group calculation that the effective scattering length indeed goes to zero at the critical temperature . as already mentioned before
, a bare chemical potential larger than the critical one , which is positive and yields a flow into the fixed point , corresponds to an inaccesible density regime .
therefore , we are not able to penetrate the region with a bose condensate and can not describe the bose gas below the critical temperature with the renormalization group equations in eq .
moreover , the critical exponent @xmath46 found from this set is not a very good approximation to the true value @xmath133 , even if we would include three - body interactions .
these aspects are intimately related and due to the fact that in the case of a positive chemical potential , the @xmath153 part of the effective action has a mexican hat shape .
therefore , we must explicitly break the symmetry and introduce the condensate density into the action by expanding the action around its minimum , and not around @xmath154 as was done in this section . for a negative chemical potential ,
the above approach is of course correct as then @xmath155 is equal to zero .
breaking the symmetry allows us to describe the dilute bose gas below the critical temperature .
moreover , also above the critical temperature we find considerable improvement .
this is due to the fact that a positive bare chemical potential can renormalize to negative values .
thus , starting out in the broken phase , the fluctuations can restore the symmetry and we end up in the unbroken phase above the critical temperature .
the new set of renormalization group equations explicitly takes this broken symmetry into account and therefore gives a much better description than the one resulting from the set of equations in eq .
( 13 ) . in order to expand the action around the correct extremum we have to perform the shift @xmath156 which introduces the condensate density @xmath62 . substituting this in the action eq .
( [ basisactie ] ) leads to the familiar expression @xcite @xmath157&=&-\hbar\beta v(\mu n_{{\bf 0}}-\frac{1}{2}n_{{\bf 0}}^2v_{{\bf 0}})+(-\mu\sqrt{n_{\bf 0}}+n_{\bf 0}\sqrt{n_{\bf 0}}v_{\bf 0})(a^*_{{\bf 0},0}+a_{{\bf 0},0})\nonumber \\ & + & \sum_{{\bf k},n}(-i\hbar\omega_n+\epsilon_{{\bf k}}-\mu+2n_{\bf 0}v_{\bf 0})a^*_{{\bf k},n}a_{{\bf k},n}+\frac{1}{2}n_{\bf
0}v_{\bf 0}\sum_{{\bf k},n}(a^*_{{\bf k},n}a^*_{{\bf -k},-n}+a_{{\bf k},n}a_{{\bf - k},-n})\nonumber \\ & + & \sqrt{\frac{n_{{\bf 0}}}{\hbar\beta v}}\sum_{\stackrel{\bf k , q}{n , m}}v_{{\bf 0}}(a^*_{{\bf q},m}a^*_{{\bf k - q},n - m}a_{{\bf k},n}+a^*_{{\bf
k+q},n+m}a_{{\bf q},m}a_{{\bf k},n})\nonumber \\ & + & \frac{1}{2}\frac{1}{\hbar\beta v}\sum_{\stackrel{\bf k , k',q}{n , n',m}}v_{{\bf 0 } } a^*_{{\bf k+q},n+m}a^*_{{\bf k'-q},n'-m } a_{{\bf k'},n'}a_{{\bf k},n}~,\end{aligned}\ ] ] for the action . the magnitude of the condensate is determined by eliminating the linear term from the action . in first instance
we thus find @xmath158 . as a result ,
we can write the action as @xmath159&=&-\hbar\beta v\omega_0+\sum_{{\bf k},n}(-i\hbar\omega_n+\epsilon_{{\bf k}}+\gamma_{11})a^*_{{\bf k},n}a_{{\bf k},n}+\frac{1}{2}\gamma_{12}\sum_{{\bf k},n}(a^*_{{\bf k},n}a^*_{{\bf -k},-n}+a_{{\bf k},n}a_{{\bf - k},-n})\nonumber \\ & + & \frac{\gamma_3}{\sqrt{\hbar\beta v}}\sum_{\stackrel{\bf k , q}{n , m}}(a^*_{{\bf q},m}a^*_{{\bf k - q},n - m}a_{{\bf k},n}+a^*_{{\bf k+q},n+m}a_{{\bf q},m}a_{{\bf k},n})\nonumber \\ & + & \frac{1}{2}\frac{v_{\bf 0}}{\hbar\beta v}\sum_{\stackrel{\bf k , k',q}{n , n',m}}a^*_{{\bf k+q},n+m}a^*_{{\bf k'-q},n'-m } a_{{\bf k'},n'}a_{{\bf k},n}~,\end{aligned}\ ] ] introducing @xmath160 as the lowest order approximation to the thermodynamic potential density , and defining the vertices @xmath161 and @xmath162 for which we have in lowest order that @xmath163 and @xmath164 .
these are the boundary conditions for the flow equations we derive next .
deriving the renormalization group equations using feynman diagrams is in this case more involved than in the unbroken phase as the number of vertices is larger , but more so because we now also can have anomalous propagators @xmath165 and @xmath166 in these diagrams .
the number of diagrams is therefore much larger and the combinatorics is more complicated .
therefore , we will here use a different method to obtain the renormalization group equations which does not explicitly make use of feynman diagrams , and is therefore very useful and efficient in this case .
it relies on the fact that in eq .
( 7 ) @xmath167d[\psi_>]\exp\left\{-\frac{1}{\hbar}(s_0[\psi^*_>,\psi_>]+s_i[\psi^ * _ < , \psi_<,\psi^*_>,\psi_>])\right\ } \equiv \exp\left\{-tr[\ln ( -g_>^{-1 } ) ] \right\}~,\end{aligned}\ ] ] where the trace is over fourier modes in the shell d@xmath12 only and @xmath168 is the greens function for these fast modes . taking the shell infinitesimal again , we can simply calculate ln@xmath169 because we then only need the part of the total action which is quadratic in the fast fourier modes ( yielding all one - loop contributions ) .
the coefficients of this quadratic part also contain the slow modes because the interaction term @xmath170 $ ] couples the slow and fast modes . by simply taylor expanding ln@xmath169
we straightforwardly find the new effective action for the slow modes and the renormalization group equations of any vertex we would like to consider .
thus , we split the fields in slow modes and fast modes as in eq.([modesplitting ] ) and find that the part from the action which is quadratic in the fast modes and only leads to the renormalization of coupling constants reads @xmath171&=&{\sum_{{\bf k},n}}'(-i\hbar\omega_n+\epsilon_{{\bf k}}+\gamma_{11}+2\gamma_3(\psi^*_<+\psi_<)+2v_{\bf 0}\mid \psi_<\mid^2)a^*_{{\bf k},n}a_{{\bf k},n}\nonumber \\ & + & ( \frac{1}{2}\gamma_{12}+\gamma_3\psi_<+\frac{1}{2}v_{\bf 0}\psi^2 _ < ) { \sum_{{\bf k},n}}'a^*_{{\bf k},n}a^*_{{\bf -k},-n}\nonumber \\ & + & ( \frac{1}{2}\gamma_{12}+\gamma_3\psi^*_<+\frac{1}{2}v_{\bf 0}\psi^{*2 } _ < ) { \sum_{{\bf k},n}}'a_{{\bf k},n}a_{{\bf -k},-n}~,\end{aligned}\ ] ] where the prime denotes that the sum over momenta is restricted to an infinitesimal momentum shell d@xmath12 at the cutoff . evaluating the functional integral over these fast modes leads to adding tr(ln(@xmath172 ) ) to the action for the slow fields , and thus changes the vertices
. this quantity is most easily evaluated by performing a bogoliubov transformation to diagonalize eq .
( [ fastquadratic ] ) @xcite , and we find that @xmath173 d\lambda~,\end{aligned}\ ] ] where the second term originates from the diagonalization procedure , and the ` dispersion ' @xmath174 is found from @xmath175 in zeroth order in @xmath176 and @xmath177 we retrieve the well - known bogoliubov dispersion @xmath178 , equal to @xmath179 and thus gapless at @xmath180 due to the equality @xmath181 .
this corresponds to the hugenholtz - pines theorem @xcite .
performing a taylor expansion in terms of the slow modes we find the new action .
thus , integrating out a momentum shell renormalizes the existing vertices , but also generates new terms in the action , and in particular a linear term . to eliminate this term and remain in the minimum of the action we again have to perform a small shift in @xmath182
this implies that the magnitude of the condensate changes as we are integrating out momentum shells and we also have a flow equation for the condensate density .
since we omit three - body interactions containing six fields , we also have to neglect terms containing five fields since they correspond , together with the condensate field , with a three - body term . as a result ,
the action remains of the form written down in eq .
( [ actiebreakdef ] ) and the renormalization group equations can now be obtained for all the vertices of interest . however , due to the @xmath183symmetry of our problem , we can relate some of these vertices and thereby limit the number of flow equations we actually need to describe the complete renormalization of the action in eq .
( [ actiebreakdef ] ) . as this @xmath183symmetry can not be broken during the process of renormalization , the action can , at any time , be recast in the explicitly @xmath184 symmetric form @xmath185&=&\int_{0}^{\hbar\beta}d\tau\int d{\bf x}\left(~ \psi^*_<({\bf x},\tau)\left [ \hbar\frac{\partial}{\partial\tau}- \frac{\hbar^2{\bf \nabla}^2}{2m}-\mu(l)\right ] \psi_<({\bf x},\tau)+\frac{1}{2}v_{\bf 0}(l)\mid\psi_<({\bf x},\tau)\mid^4 \right)~.\end{aligned}\ ] ] from this it is then first of all easy to see that the hugenholtz - pines theorem @xcite holds , implying that @xmath186 .
( see appendix a for an explicit derivation of this important relation . ) next , also due to the neglect of three - body interactions , we have that @xmath187 and @xmath188 .
thus , we only need to know the flow equations for e.g. @xmath189 and @xmath190 and then the other renormalization group equations can be determined . in the following
we can therefore restrict ourselves to the renormalization of the linear term and the term proportional to @xmath191 as these determine the change of the condensate density and the anomalous selfenergy @xmath192 respectively .
note that the equality @xmath193 ensures that the dispersion @xmath194 is gapless at any point during renormalization , as it should .
to arrive at the flow equations for @xmath189 and @xmath190 , we first of all need the linear term @xmath195 that is generated by integrating out a momentum shell . we find that @xmath196 analogously we find a change @xmath197 in the anomalous selfenergy .
( see appendix a for details .
) however , before we know the full renormalization of this vertex we have to determine the shift needed to eliminate the linear term that is generated .
substituting @xmath198 and retaining only the term linear in @xmath199 we find that @xmath200 which influences the renormalization of @xmath192 because we have for the total renormalization of the anomalous selfenergy @xmath201 due to this shift .
using the above mentioned relations between the vertices implicate from the @xmath183symmetry we then make contact with the renormalization group equations for the unbroken phase by determining the flow equations for the chemical potential @xmath202 and the two - body interaction @xmath203 .
after some algebra we ultimately find @xmath204\right]\end{aligned}\ ] ] @xmath205\right]~,\end{aligned}\ ] ] and the condensate density follows from @xmath206 . breaking the symmetry is irrelevant to the trivial scaling of the vertices , and thus these are identical to what was found in the previous section . comparing these flow equations to eqs .
( [ murgboven ] ) and ( [ vnulrgboven ] ) omitting @xmath45 , we see that both sets coincide when @xmath26 is taken equal to zero .
thus , the renormalization group equations for negative and positive chemical potential yield a flow which is continuously differentiable , also at @xmath207 .
we now know the renormalization of the chemical potential and the two - body interaction .
finally , we again have to find the equations describing the building up of the total and superfluid densities and the thermodynamic potential as we are integrating out momentum shells .
analogous to the procedure followed in the previous section we find for the total density , being the sum of the condensate density and the above condensate density , @xmath208 with the boundary condition @xmath209 .
the superfluid density follows from @xmath210e^{-3l}~,\end{aligned}\ ] ] and the thermodynamic potential from @xmath211 ( cf .
( [ traceresult ] ) ) with the boundary condition @xmath212 . with the boundary conditions mentioned above we
can again numerically integrate the renormalization group equations . for a fixed temperature ,
we vary the value of the ( positive ) bare chemical potential , and calculate e.g. density and pressure .
the physical effective chemical potential , i.e. with the trivial scaling removed , decreases as we perform the integration , and depending on the starting value remains positive , renormalizes exactly to zero , or becomes negative at some value of the integration parameter @xmath134 . in the first case
we start out and remain in the broken phase and are below the critical temperature of bose - einstein condensation , i.e. we have a finite condensate density . in the second case ,
the condensate density @xmath206 renormalizes exactly to zero for @xmath86 , and we are at the critical conditions for bose - einstein condensation . in the third case we started out in the broken phase , but the fluctuations restore the symmetry . at the value of @xmath134
for which the chemical potential becomes negative we have to continue the integration with the set eqs .
( [ murgboven ] ) and ( [ vnulrgboven ] ) , and we are thus in the uncondensed phase . hence , to be able to describe the dilute bose gas above , but not too far from the critical temperature , we need the renormalization group equations for both signs of @xmath26 . in fig . 5 we depict the trajectories resulting from the integration of eqs .
( [ murgonder ] ) and ( [ vnulrgonder ] ) .
it is evident from this figure that the critical properties of the bose gas are determined by this set of equations .
by linearizing the flow equations around the fixed point we can identify the largest eigenvalue @xmath120 , and determine the critical exponent @xmath46 following from this set .
we find @xmath213 , which gives a much better approximation to the critical exponent than the renormalization group equations studied in sec .
iii and is to be compared with the value @xmath133 found from the @xmath214expansion of the @xmath215 model @xcite and measured in @xmath132he experiments .
the agreement is surprisingly good , and together with the fact that we explicitly showed that three - body effects are negligible , this indicates that we are indeed accurately describing the bose gas with the derived renormalization group equations , also in the critical region .
the cause of this good agreement is that although we only consider the renormalization of @xmath26 and @xmath48 , the type of scattering processes in terms of real ( bare ) particles we are actually taking into account are very elaborate .
the propagator for the the bogoliubov quasiparticles is namely the result of dressing the original bare propagator @xmath216 with interactions with the condensate as we use the terms @xmath217 and @xmath218 in the zeroth order quadratic part of the action .
thus , the diagrams we calculate actually contain an infinite number of scattering processes with the condensate .
therefore , we are describing the system much better than in sec .
iii already at this level of renormalization group .
the first nonuniversal property we concentrate on is the change in the critical temperature of bose - einstein condensation due to interaction effects .
this result is presented also elsewhere @xcite , but we will recapitulate it here because of its experimental significance . at fixed temperatures we vary the ( bare ) chemical potential to find the trajectories flowing into the fixed point . using eq .
( [ dichtheidonder ] ) , this yields the critical densities for these specific temperatures and gives us the @xmath219 relation at which bose - einstein condensation occurs .
we repeat this for different values of the scattering length to obtain the dependence of the critical temperature on the strength of the interaction . in fig .
6 we show the degeneracy parameter @xmath220 found from the renormalization group calculation as a function of @xmath221 .
as seen from this figure we conclude that the critical temperature is _ raised _ with respect to the ideal gas value .
this is in qualitative agreement with the recent experiments @xcite and also preliminary quantum monte carlo calculations seem to confirm this result @xcite .
an indication of an upward shift was also found some time ago by one of us studying the nucleation of bose - einstein condensation @xcite . from our calculations
we predict that for the @xmath1rb and @xmath2na experiments the critical temperature can be raised with as much as 10 % , which appears to be a very promising result because one might expect that an effect of this magnitude can very well be measured in future , more accurate , experiments .
it is important to note that this shift in @xmath222 can be observed in magnetically trapped atomic gases if one directly measures the density in the center of the trap at the critical temperature .
one should in particular _ not _ measure the total number of particles , because this involves the density profile in the trap and due to the repulsive nature of the interactions thus tends to obscure the effect @xcite .
the reason for a higher critical temperature , or more precisely , a lower critical density is the following .
the effective chemical potential renormalizes from a positive initial value exactly to zero .
consequently , we have the bogoliubov dispersion in the equation for the density , and this depresses the occupation of the non - zero momentum states relative to the ideal gas case , where we would just have @xmath223 in the bose - einstein distribution function .
the magnitude of the effect is related to the behavior of the chemical potential when renormalizing to the fixed point value @xmath224 .
suppose we effectively have @xmath225 independent of @xmath134 , with some positive @xmath226 smaller than 1 .
we can then translate the differential equation for the density into an ordinary integral over momentum space following the inverse procedure from which we found the flow equations in sec .
iii . doing so , we find that the bogoliubov dispersion effectively becomes @xmath227 .
thus , this essentially boils down to a mass renormalization , and the change in the critical temperature is directly related to the magnitude of this renormalization . starting with the equation describing the building up of the above condensate density ( cf .
eq.([dgam11a ] ) from appendix a ) one can easily show that we approximately have @xmath228 .
for an increase of 10 % in the critical temperature , we find from this result that @xmath226 must be equal to 0.2 . indeed , this turns out to be the typical value of @xmath226 in the thermal regime where the contributions to the density are largest . to obtain also information on the properties of the bose gas below the critical temperature we have to start with a chemical potential which always remains positive under renormalization .
we are then always in the condensed phase and have to use the renormalization group equations ( [ murgonder ] ) and ( [ vnulrgonder ] ) . written as ordinary one - loop integrals
, these equations contain infrared divergencies and a straightforward perturbative analysis is not possible .
this is a well known problem in the theory of the interacting bose gas @xcite .
however , doing the calculation by means of the renormalization group approach , this problem is in principle resolved due to the resummation which is automatically performed .
the presence of infrared divergencies causes both the physical chemical potential and two - body interaction in the renormalization group approach to renormalize to zero , instead of becoming infinite as they would in a regular one - loop calculation .
indeed , one can show from eqs .
( [ murgonder ] ) and ( [ vnulrgonder ] ) that for @xmath230 we have @xmath231 and @xmath232 .
these scalings are different from the trivial scalings @xmath233 and @xmath234 , and are therefore termed anomalous .
physically , the anomalous scaling implies an effective energy and momentum dependence of the coupling constants , as we show explicitly in appendix b. in particular , the anomalous scaling we find leads to the result that the chemical potential , and thus the normal and anomalous selfenergies , as well as the two - body interaction , behave linearly with @xmath235 for low momenta , i.e. our renormalization group approach reveals that @xmath236 and @xmath237 .
this behavior , implying that @xmath238 , is an exact result for low momenta @xcite which we explicitly recover here .
the consequences for the application of renormalization group are however twofold .
first , it implies that the bogoliubov dispersion @xmath239 does not possess a sound mode anymore since the low momentum behavior is not linear , but instead we have @xmath240 .
this is of course an incorrect result @xcite indicating that eqs .
( [ murgonder ] ) and ( [ vnulrgonder ] ) are insufficient to describe the sound mode . indeed , following the argument in sec .
iii we neglected the renormalization of the ( marginal ) @xmath95 and @xmath96 terms since the interactions were anticipated to be momentum and energy independent .
however , below the critical temperature we see that the momentum and energy dependence of the selfenergies does become important , and should therefore be included in the renormalization group calculation by taylor expanding the selfenergies in terms of the external momentum and frequency . put differently
, the anomalous dimension @xmath98 is no longer small below the critical temperature .
the extra renormalization group equations obtained in this manner may change the particular anomalous scaling found above , but since @xmath241 is an exact result we still expect to have anomalous scaling of the coupling constants .
the effect of the extra renormalization group equations will be to change the dispersion relation in such a way that the linear sound mode is recovered .
we will not pursue this calculation here , but postpone it to future work .
the reason for the anomalous scaling and consequently the disappearance of the sound mode , is caused by the infrared divergence in the one - loop expressions for the selfenergies and two - body interaction .
this can be traced back to the behavior of the coherence factors @xmath242 and @xmath243 of the bogoliubov transformation diagonalizing the quadratic part of the action .
the bogoliubov transformation is given by [ bogtrafo ] @xmath244 @xmath245 where @xmath246 and @xmath247 are the fourier components of the fields creating respectively annihilating a bogoliubov quasiparticle .
the coherence factors are given by [ coherence ] @xmath248 @xmath249 and satisfy the requirement that @xmath250 because the bogoliubov transformation is unitary . as the bogoliubov dispersion is given by @xmath251 , it is easy to see that @xmath252 and @xmath253 for large momenta , and both factors behave as @xmath254 as @xmath235 tends to zero .
the region where the crossover occurs between these two regimes is determined by the parameter @xmath255 , found from comparing the chemical potential with the thermal energy , and also determining the crossover from quadratic to linear behavior of the bogoliubov dispersion . when @xmath256 the linear regime of the dispersion is extremely important in determining the properties of the bose gas .
conversely , when @xmath229 , the bogoliubov dispersion can be approximated by the normal dispersion @xmath257 , except for a very small region at low momenta , which is then not very important and can for most practical purposes be neglected . in the following ,
we therefore concentrate on the regime @xmath229 , covering a large temperature interval below the critical temperature , and use @xmath258 and @xmath259 in contributions to the renormalization of the vertices that contain infrared divergencies .
otherwise , we will use the full expression eq .
( [ coherence ] ) .
this procedure turns out to be necessary to ensure the hugenholtz - pines theorem to be satisfied , as is shown in appendix a. in the method used in the previous section it is not possible to pinpoint the bogoliubov coherence factors at any stage of the derivation of the renormalization group equations .
therefore , we really have to go through the calculation of the feynman diagrams relevant to the various vertices we are interested in . in our case
these are the linear term of the action and the anomalous selfenergy .
the one - loop diagrams of interest are depicted in fig . 7 and contain also the anomalous propagators @xmath260 and @xmath261 . using the bogoliubov transformation ,
these diagrams are straightforward to calculate , and we find that for the renormalization of the condensate density both diagrams for @xmath262 contribute , but that for the renormalization of the anomalous selfenergy only diagrams @xmath263 and @xmath264 give contributions that do not contain infrared divergencies . the explicit calculation of these diagrams is presented in appendix a. using the expressions obtained there , we find that the flow equations become @xmath265\right)~,\end{aligned}\ ] ] @xmath266\right)~,\end{aligned}\ ] ] where we again made use of the relations @xmath267 due to @xmath183symmetry .
note that also this set coincides with eqs .
( [ murgboven ] ) and ( [ vnulrgboven ] ) when the chemical potential is equal to zero , so the flow is also in this case everywhere continuous and continuously differentiable .
moreover , the equations for the density , superfluid density and thermodynamic potential do not contain an infrared divergency , so the flow equations for these quantities remain the same and are given by eqs.([dichtheidonder ] ) , ( [ superdhonder ] ) and ( [ omegaonder ] ) , respectively . at this point
we can describe any point in the phase diagram of the dilute bose gas . for negative chemical potential
we have eqs .
( [ murgboven ] ) and ( [ vnulrgboven ] ) , for positive chemical potential we must use eqs.([muzonderir ] ) and ( [ vnulzonderir ] ) .
moreover , we have to combine both sets when we are not too far below the critical density when the chemical potential changes sign during application of the renormalization group transformation . using eq .
( 37 ) , we find that the critical temperature of bose - einstein condensation changes with less than 0.1 % compared to the more accurate result found using eq .
therefore , this shows that using the set not containing the infrared divergencies essentially leads to the same results , implying that it is indeed correct to neglect the linear part of the bogoliubov dispersion .
having proven this explicitly , we will in the following present the results of the renormalization group calculation of the effective two - body interaction , i.e. the many - body scattering length @xmath145 , the condensate and superfluid densities and the @xmath141diagram below as well as above the critical temperature , and compare them in all cases with the results from the many - body @xmath6matrix calculation .
we start with the scattering length @xmath145 . above the critical temperature
it is straightforward to take into account only the contributions of the ladder diagrams , or to include the effect of the bubble diagrams as well . below the critical temperature
there are no clear ladder and bubble diagrams , so the various contributions are in principle not clearly associated with ladders or bubbles .
however , comparing the differential equations for positive and negative @xmath26 we can conclude that the first nontrivial term on the righthand side of eq .
( [ vnulzonderir ] ) is effectively related to a ladder contribution , and that the second nontrivial term is effectively related to a bubble diagram .
we depict in fig .
8 the scattering length @xmath145 , normalized to the two - body scattering length @xmath268 for @xmath2na , when including and excluding the bubble contributions ( @xmath269 ) , and also the result found from the many - body @xmath6matrix calculation ( @xmath270 ) @xcite .
as our approach is restricted to the regime @xmath271 , we present no results for @xmath255 larger than one .
the renormalization of the effective two - body interaction to zero at the critical temperature , a result already found in the many - body @xmath6matrix approach , turns out to be correct .
indeed , this can be easily understood from the renormalization group equations . a fixed point @xmath272 is present in the set @xmath273 , which means that for @xmath86 the physical two - body interaction , i.e. with the trivial scaling removed , behaves as @xmath274 and thus goes to zero .
the depression in the scattering length around the critical temperature occurs in a fairly large temperature interval .
note furthermore that applying the renormalization group equations ( [ murgonder ] ) and ( [ vnulrgonder ] ) would lead to @xmath275 everywhere below the critical temperature due to the anomalous scaling found from this set .
this property is actually expected to hold true even in a more elaborate renormalization group calculation and is an issue worthwile studying as it may have important consequences for e.g. the exact form of the gross - pitaevskii equation describing the condensate .
next , we turn to the equation of state . for the ideal bose gas we have in general that @xmath276 where @xmath277 is a bose - function defined as @xmath278 and @xmath279 the fugacity . in the many - body @xmath6matrix theory above the critical temperature
we have that the chemical potential is renormalized to @xmath144 @xcite . below the critical temperature
the dispersion changes to the bogoliubov dispersion and the equation for the density is in essence given by eq .
( [ dichtheidonder ] ) , recast in the form of an ordinary one - loop integral . in fig .
9 we plot @xmath280 for the ideal bose gas , following from the many - body @xmath6matrix calculation , and from the renormalization group calculation . in the latter case
, we numerically integrate the flow equations for a fixed temperature , and vary the value of the bare chemical potential .
this changes the total density in the system , and therefore the value of @xmath281 . for the ideal bose gas , positive values of the chemical potential
are not allowed . with the repulsive interactions taken into account , a positive value is possible and
the fugacity can be larger than one .
the renormalization group calculation yields the same density for a chemical potential slightly above and slightly below the critical chemical potential .
this double valuedness in the density occurs in an extremely small region around the critical density and the effect is smaller than 1 promille in the case of @xmath2na . for @xmath282h , with a scattering length @xmath283 , the effect is even far below the promille level .
however , it is important to note that this effect is much smaller than the change in density we obtain when we include three - body interactions , being approximately 1 % ( cf .
therefore , the double valuedness we find can not be trusted physically and should be neglected .
it presumably disappears when we include three - body effects or extend the renormalization group calculation otherwise .
when the bare chemical potential decreases the gas becomes more and more dilute and the influence of the interactions can better and better be accounted for using mean - field theory .
this fact is actually evident from the renormalization group equations in eq .
when the chemical potential is large and negative , the bose - factors @xmath284 are strongly depressed and there is hardly any many - body effect on the renormalization of the two - body interaction so we will find that it is just renormalized to @xmath285 .
indeed , the same is true in the many - body @xmath6matrix calculation @xcite , since we have that @xmath286 when the system is extremely dilute , i.e. there is no effect of the medium in this regime .
in addition , the differential equations for @xmath26 and @xmath48 are now almost decoupled , and consequently the chemical potential will renormalize approximately to @xmath287 , which is only a small effect in this limit .
therefore , the renormalization group equations can practically be recast in the regular one - loop expressions one encounters in the many - body @xmath6matrix theory and the results we find in this regime are approximately the same . in figs .
10(a ) and 10(b ) we depict the condensate and superfluid fractions as a function of temperature for a density of 1.5 10@xmath288 @xmath152 both from the renormalization group and many - body @xmath6matrix calculation .
again , not too close to the critical temperature we have good agreement .
note that the superfluid density has not yet become zero at the temperature where the condensate density vanishes . to explain this aspect we will focus on the equation for the superfluid density in the unbroken phase .
when we interpret the fact that the chemical potential is renormalized as we integrate out momentum shells as a chemical potential depending on momentum , we can express the superfluid density as a regular integral over momentum space , i.e. @xmath289 writing @xmath290 and performing a taylor expansion , we find to first order in @xmath291 that @xmath292[1-\frac{2}{3}\beta\epsilon_{\bf k}(2n(\epsilon_{\bf k}-\mu_0)+1)]~\delta\mu ( k ) , \end{aligned}\ ] ] where the lowest order terms drop out .
this is due to the fact that these terms yield the expression for the superfluid density in an ideal bose gas above the critical temperature which can be shown to be exactly equal to zero .
it is clear from eq .
( [ superfluidrg ] ) that the superfluid density will in general not be equal to zero .
the behavior of @xmath293 would have to be very special to give a superfluid density exactly equal to zero .
it is therefore not surprising that we find from our limited set of renormalization group equations a different temperature for which the condensate and superfluid densities vanish .
for the situation depicted in fig .
10 we have however that @xmath294 is only about @xmath295 .
extending the renormalization group calculation would in principle lead to a superfluid density which vanishes at the same temperature as where the condensate density becomes zero .
finally , we also present the pressure of the dilute bose gas as a function of the inverse density . using eq .
( [ omegaonder ] ) , we depict this behavior in fig . 11 for a @xmath2na gas at 1 @xmath26k , together with the result of the many - body @xmath6matrix calculation .
although the critical densities are different , we find a fairly good agreement between the two curves and the difference between the renormalization group calculation and the many - body @xmath6matrix calculation is small when it concerns the pressure of the gas .
however , concerning the other nonuniversal properties discussed in this article , the difference can be substantial near the critical temperature as we have seen .
in summary , we have derived the renormalization group equations for a dilute bose gas . to obtain these equations we have to make a distinction between the case of a negative and a positive chemical potential . in the latter case
we have to take the presence of a condensate into account , and the renormalization group equations are different from the ones valid in the unbroken phase .
our philosophy of using the renormalization group approach is not to study universal properties of the dilute bose gas , which are the same as for the @xmath131model in three dimensions , but to make quantative predictions about various nonuniversal properties of this system . to achieve this goal one has to find a method to eliminate the ultraviolet cutoff dependence inherent to the application of the renormalization group .
our knowledge about the two - body scattering problem is sufficient in this respect and we can fix the renormalization group equations in this manner .
we compared results of the renormalization group with mean - field calculations and showed the difference to vanish in appropriate limits .
we also checked the influence of three - body effects , which turned out to be unimportant even in the critical region .
furthermore , we showed that the influence of bubble diagrams on the effective interaction can be fairly large , and that the effective scattering length @xmath145 vanishes when approaching the critical temperature .
this confirms earlier results @xcite and is of importance for a lot of current work concerning condensate properties in trapped bose gases .
it implies that there can in principle be an important change in the results of calculations that make use of the nonlinear schrdinger equation when determining these properties . for that purpose
the effective scattering length @xmath145 is of importance , and not the two - body scattering length @xmath4 .
next , we derived the renormalization group equations in the broken phase , i.e.for positive chemical potential , and showed that the transition to the bose - einstein condensed phase is contained in this set of renormalization group equations .
we found a critical exponent @xmath213 , which agrees very well with @xmath133 found in studies of the @xmath131model , and furthermore calculated the effect of the interactions on the critical temperature of the phase transition .
the change can be as much as 10 % in the current type of experiments and may be measured if one can improve the precision in determining the temperature and the central density in the trap . as these latter renormalization group equations lead to a dispersion relation which has no linear part ,
i.e. there is no sound mode , we have to restrict ourselves to the regime @xmath229 , where the linear part of the dispersion is unimportant for determining properties such as density and pressure .
the region where @xmath229 is currently still the most interesting one from the experimental point of view although the other region is certainly within reach . including the renormalization of the time derivative and gradient terms in the action
is expected to resolve the problem of the disappearing of the sound mode , but we postponed this to future work .
a resolution of this problem would correspond to a resolution of the long standing problem of the infrared divergencies in the perturbation expansion around the bogoliubov theory .
we concluded this work with using the renormalization group to calculate superfluid and condensate densities as a function of temperature and also a @xmath141diagram is presented . of course the renormalization group can be used to find out many more things about the dilute bose gas .
note e.g. that it is in principle also possible to use the renormalization group method to find the quantative form of the correlation function as it is the fourier transform of the occupation number @xmath296 .
this can also be translated into a differential equation as in eq .
( 19 ) where we now have the distance @xmath297 as a free parameter . also the specific heat may be calculated . in principle
, our results pertain to a homogeneous bose gas , but in situations where the application of a local density approximation is allowed , they are also applicable to trapped bose gases
. moreover , we indicated that it is in principle also possible to set up a renormalization group calculation for the inhomogeneous case .
finally , we want to note that the procedure of renormalization group as described in this article can in principle also be used to study the kosterlitz - thouless transition to the superfluid phase in a two - dimensional bose gas . however , there are difficulties in this case connected with the fact that all coupling constants are relevant at the critical temperature .
nonetheless , work along these lines is in progress since a number of experiments are currently under construction which aim at reaching the kosterlitz - thouless phase in doubly spin - polarized atomic hydrogen adsorbed on a superfluid helium film .
we acknowledge helpful discussions with eric cornell , wolfgang ketterle and steve girvin .
in this appendix we go through some of the technicalities of the calculation of the one - loop feynman diagrams for the coupling constants @xmath298 and @xmath192 , i.e. essentially the normal and anomalous selfenergies , and explicitly show the hugenholtz - pines theorem to hold in our renormalization group approach .
this theorem states that @xcite @xmath299 where @xmath300 and @xmath301 are the irreducible normal and anomalous selfenergies respectively . in our notation
this relation reads @xmath302 fig . 7 in the main text
contains the diagrams renormalizing @xmath192 and in fig .
12 we depict the one - loop diagrams renormalizing @xmath298 .
using the designation of the diagrams as in these figures one can show , after going through the combinatorics , that the ( infinitesimal ) change of the vertices after integrating out an infinitesimal momentum shell is given by @xmath303 @xmath304 the last term in both expressions originates from the shift in @xmath182 required to eliminate the linear term d@xmath305 from the action . from figs . 7 and 12 it is clear that a number of diagrams contributing to @xmath298 or @xmath192 are mathematically identical .
most notably we have @xmath306 , @xmath307 , and @xmath308 .
therefore , there are only six independent diagrams , namely @xmath309 , and @xmath310 . using the bogoliubov transformation from eq .
( [ bogtrafo ] ) , it follows that @xmath311 @xmath312 with @xmath242 and @xmath243 given in eq .
( [ coherence ] ) , and these diagrams are now straightforward to calculate by applying the usual feynman rules @xcite .
we have @xmath313 @xmath314\right ) e^{-3l}dl~,\end{aligned}\ ] ] @xmath315\right ) e^{-3l}dl , \end{aligned}\ ] ] @xmath316\right)e^{-3l}dl~,\end{aligned}\ ] ] @xmath317\right ) e^{-3l}dl~,\end{aligned}\ ] ] @xmath318 introducing @xmath319 and @xmath320 with which one can easily verify that @xmath321 combining these equations leads to the flow equations for @xmath298 and @xmath192 .
however , we here concentrate on the hugenholtz - pines relation , which is satisfied at @xmath180 , and therefore remains valid if also @xmath322 at @xmath180 . using the above mentioned equalities of the various diagrams ,
this equality reduces in first instance to @xmath323 however , due to the @xmath183symmetry of the action @xmath324 $ ] we have at any value of @xmath134 that @xmath325 and at @xmath180 , when @xmath326 , eq .
( [ hupistep1 ] ) reduces with the help of eq .
( [ dnnulap ] ) to @xmath327 or equivalently @xmath328 note that , due to the fact that @xmath329 , the left - hand side of this equation actually does not contain bogoliubov coherence factors . from the definition of the coherence factors in eq .
( [ coherence ] ) it is now easy to show that indeed , at @xmath180 , we have @xmath322 and as a consequence that the hugenholtz - pines relation remains valid at any point during the integration .
next , we briefly turn to the regime @xmath229 where we put @xmath258 and @xmath259 in contributions containing infrared divergencies . in this case , only the first term in the left - hand side of eq.([hupistep3 ] ) contributes , but now with a factor 1 instead of @xmath330 .
thus , nothing changes with respect to the previous situation , and the hugenholtz - pines relation is again valid at any point during the renormalization if we still use the exact expressions for @xmath242 and @xmath243 in contributions not containing infrared divergencies , and most importantly in @xmath331 .
in this appendix we discuss the implications of anomalous scaling , focussing on the situation at the critical temperature where the coupling constants flow into the fixed point @xmath332 .
we start with the fundamental statement from renormalization group theory that the exact @xmath333point vertex function remains identical during renormalization , i.e. in our case @xmath334 where @xmath335 .
the coupling constants obey the derived renormalization group equations ( without trivial scaling ) and @xmath76 .
when we are at the critical temperature , we can take @xmath336 and we find after performing the trivial rescaling that @xmath337 a dimensional analysis then shows that @xmath338 combining these two equations , and taking the limit @xmath86 in which we approach the fixed point , this implies @xmath339 the right - hand side has , just like the left - hand side , to be independent of @xmath134 , and this leads to the conclusion that @xmath340 or that @xmath341 has to be a homogeneous function of degree @xmath342 .
thus @xmath343 and we can conclude that anomalous scaling reveals information about the momentum dependence of the @xmath333point vertex function . in particular , we have e.g. for the selfenergy at the critical temperature that @xmath344 and most importantly for the four - point function that @xmath345 which shows that the effective interaction at long wavelengths has to vanish at the critical temperature . clearly , the above reasoning is in principle not restricted to the critical temperature and a similar argument can be set up also for arbitrary temperatures . as a result ,
the anomalous scaling we find from the set of renormalization group equations derived in sec .
iva for the symmetry broken phase also implies a nontrivial momentum dependence of the coupling constants .
anderson , j.r .
ensher , m.r .
matthews , c.e . wieman and e.a .
cornell , science * 269 * , 198 ( 1995 ) .
davis , m.o.mewes , m.r .
andrews , n.j .
van druten , d.s .
durfee , d.m .
kurn and w. ketterle , phys .
. lett . * 75 * , 3969 ( 1995 ) .
bradley , c.a .
sackett , j.j .
tollett and r.g .
hulet , phys .
lett . * 75 * , 1687 ( 1995 ) .
stoof , phys .
a * 49 * , 3824 ( 1994 ) .
ruprecht , m.j .
holland , k. burnett , and m. edwards , phys .
a * 51 * , 4704 ( 1995 ) .
stoof , ( unpublished , cond - mat/9601150 ) .
s. stringari , phys .
* 76 * , 1405 ( 1996 ) .
stoof , phys .
lett . * 66 * , 3148 ( 1991 ) ; phys .
a * 45 * , 8398 ( 1992 ) .
f. dalfovo , l. pitaevskii , s. stringari , ( unpublished , cond - mat/9604069 ) .
m. edwards and k. burnett , phys . rev .
a * 51 * , 1382 ( 1995 ) .
s. stringari , ( unpublished , cond - mat/9603126 ) .
fetter , phys .
a * 53 * , 4245 ( 1996 ) .
m. edwards , p.a .
ruprecht , k. burnett , r.j .
dodd and c.w .
clark , ( unpublished , cond - mat/9605170 ) .
stoof , m. houbiers , c.a . sackett and r.g .
hulet , phys .
lett . * 76 * , 10 ( 1996 ) .
m. bijlsma and h.t.c .
stoof , ( unpublished , cond - mat/9507134 ) .
a. griffin , physica c * 156 * , 12 ( 1988 ) .
h. shi , g. verechaka and a. griffin , phys .
b * 50 * , 1119 ( 1994 ) .
t.d . lee and c.n .
yang , phys . rev . * 112 * , 1419 ( 1958 ) .
l. reatto and j.p .
straley , phys . rev . * 183 * , 321 ( 1969 ) .
m. bijlsma and h.t.c .
stoof , ( unpublished , cond - mat/9605120 ) .
k.g . wilson and j. kogut , phys .
* 12 * , 75 ( 1974 ) .
d.s . fisher and p.c .
hohenberg , phys .
b * 37 * , 4936 ( 1988 )
weichmann , phys .
b * 38 * , 8739 ( 1988 ) . e.b . kolomeisky and j.p .
straley , phys .
b * 46 * , 11749 ( 1992 ) .
kolomeisky and j.p .
straley , phys .
b * 46 * , 13942 ( 1992 ) .
e. tiesinga , c.j .
williams , p.s .
julienne , k.m .
jones , p.d .
lett and w.d .
phillips , ( unpublished ) .
see for example j.w .
negele and h. orland , _ quantum many - particle systems _ ( addison - wesley , new york , 1988 ) .
r. shankar , rev .
phys . * 66 * , 129 ( 1994 ) .
w. glckle , _ the quantum mechanical few - body problem _ ( springer , berlin , 1983 ) . k. halpern and k. huang , phys .
lett . * 74 * , 3526 ( 1995 ) .
j. zinn - justin , _ quantum field theory and critical phenomena _
( oxford university press inc .
, new york , 1989 ) .
stoof , m. bijlsma and m. houbiers , ( accepted for publication in the nist journal of physics ) .
popov , _ functional integrals in quantum field theory and statistical physics _
( reidel , dordrecht , 1983 ) , chap . 6 . a.l .
fetter and j.d .
walecka , _ quantum theory of many - particle systems _ ( mcgraw - hill , new york , 1971 ) , chapter 10 .
hugenholtz and d. pines , phys . rev . *
116 * , 489 ( 1958 ) .
p. grter , d. ceperley , m. dewing and f. lalo , ( private communication ) . s. giorgini , l.p . pitaevskii and s. stringari , ( unpublished , cond - mat/9607117 ) .
a. griffin , _ excitations in a bose - condensed liquid _ , ( cambridge university press , new york , 1993 ) .
a. nepomnyashchi@xmath346 and a.a .
nepomnyashchi@xmath346 , zh . eksp .
. fiz . * 75 * , 976 ( 1978 ) , [ sov .
jetp * 48 * , 493 ( 1979 ) ] .
fig . 1 typical behavior of the fourier transform of the interatomic interaction potential @xmath347 .
+ + fig . 2 the one - loop feynman graphs contributing to the renormalization of ( a ) the chemical potential , ( b ) the two - body interaction , and ( c ) the three - body interaction .
the dot represents the two - body vertex @xmath48 and the square the three - body vertex @xmath45 .
+ + fig . 3 the @xmath141diagram including ( solid line ) and excluding ( dashed line ) a three - body interaction term for atomic @xmath2na at a temperature of @xmath348 .
the influence of @xmath45 is approximately 1.5 % near the critical density .
+ + fig . 4 the ratio of the many - body scattering lengths @xmath147 including ( solid line ) and excluding ( dashed line ) the bubble diagrams in the flow equations as a function of @xmath150 for atomic @xmath2na at a density of @xmath151 @xmath152 .
+ + fig . 5 flow diagram resulting from the renormalization group equations .
the fixed point is indicated with an asterix .
+ + fig . 6
the critical degeneracy parameter @xmath222 found from the renormalization group calculation as a function of @xmath221 .
+ + fig . 7
the one - loop feynman diagrams contributing to the renormalization of ( a ) the linear term and ( b ) the anomalous selfenergy in the broken phase . the filled circle represents the vertex @xmath48 and the open circle the vertex @xmath162 .
+ + fig . 8
the ratio of the effective scattering length @xmath145 and the two - body scattering length @xmath4 for atomic @xmath2na at a density of @xmath151 @xmath152 .
the dashed line corresponds to the result from the many - body @xmath349-matrix calculation , i.e. @xmath350 .
the solid lines correspond to the result from the renormalization group approach , i.e. @xmath351 , with bubble diagrams included and excluded .
the effect of the bubbles ( lower solid line ) is seen to be considerable , certainly below the critical temperature .
+ + fig . 9
the degeneracy parameter @xmath281 as a function of the fugacity @xmath352 for the ideal gas ( dotted line ) , from the many - body @xmath6matrix theory ( dashed line ) and as found from the renormalization group calculation ( solid line )
. + + fig .
10 ( a ) the condensate fraction and ( b ) the superfluid fraction as a function of temperature for a density of @xmath151 sodium atoms per cubic centimeter , both from the many - body @xmath6matrix calculation ( dashed line ) as from the renormalization group calculation ( solid line ) .
+ + fig .
11 the pressure as a function of inverse density for atomic @xmath2na at a temperature of @xmath348 .
the dashed line is the result from the many - body @xmath6matrix calculation , the solid line from the renormalization group calculation .
+ + fig .
12 the one - loop feynman diagrams contributing to the renormalization of the normal selfenergy in the broken phase . the filled circle represents the vertex @xmath48 and the open circle the vertex @xmath162 .
|
we study the three - dimensional atomic bose gas using renormalization group techniques . using our knowledge of the microscopic details of the interatomic interaction
, we determine the correct initial values of our renormalization group equations and thus obtain also information on nonuniversal properties . as a result
, we can predict for instance the critical temperature of the gas and the superfluid and condensate density of the bose - einstein condensed phase in the regime @xmath0 .
|
classical t tauri stars ( ctts ) are young ( @xmath3 yr ) , low mass @xmath4 stars at the stage of gravitational contraction towards the main - sequence , activity of which is caused by magnetospheric accretion of protoplanetary disc matter @xcite .
inner regions of the accretion disc are truncated by the stellar magnetic field and disc s matter slides down toward the star along the field lines .
having reached the dense layers of the stellar atmosphere , matter is decelerated in the accretion shock , behind the front of which the most part of its kinetic energy converts into the short - wavelength radiation .
one half of the radiation flux irradiates the star , producing the so - called hot spot on its surface .
simulations , performed by dodin & lamzin @xcite , confirm the hypothesis , suggested by batalha et al .
@xcite , that the so - called narrow ( @xmath5 kms@xmath6 ) components of emission lines in spectra of ctts are formed in this region .
hence , if the narrow components of the emission lines are used for the magnetic field measurements , then the obtained field strength characterizes the field in the accretion zone on the stellar surface .
the second half of the short - wavelength radiation flux of the shock from the cooling zone escapes upward , heating and ionizing the pre - shock gas .
broad @xmath7 kms@xmath6 ) components of the emission lines are formed in this region as well as in a more extended region at the truncation radius , where the disc interacts with the field , and a magnetospheric wind are launched .
the relation between the intensity of the narrow and broad components varies from line to line , changes over time , and for the same lines varies from star to star .
because the broad components are formed in the region at a significant distance from the stellar surface , the field strength , measured on these components , is , as a rule , in several times smaller than that , measured on the narrow ones ( see , however , @xcite ) .
the magnetic field of young stars determines an activity of ctts and plays a crucial role in an evolution of angular momentum of these objects , therefore a question about strength and topology of the magnetic field is one of the fundamental questions of physics of young stars . at the moment ,
the magnetic field has been found in about 40 ctts .
these measurements are significantly differ in the methods , and therefore , the field strength , measured in different works , can relate to different spatial regions .
maps of the magnetic field at the photospheric level have been reconstructed by analysis of polarized light for the following stars : cv cha and cr cha @xcite , tw hya @xcite , v2129 oph @xcite , v2247 oph @xcite , aa tau @xcite , bp tau @xcite .
the field has been detected by means of zeeman broadening in unpolarized light in 26 stars @xcite .
the field on the narrow emission components , which are formed in the accretion zone , has been measured only in 10 stars : rw aur @xcite , tw hya @xcite , v2129 oph @xcite , aa tau @xcite , bp tau @xcite , gm aur , df tau , dn tau , gg tau @xcite , t tau @xcite . to measure a magnetic field , spectra with a high quality
are needed , therefore the observations carry out on large telescopes for bright stars with numerous strong lines . however , the number of such stars is small , and practically , it is limited by the list , presented above . if we do not aim to determine the magnetic field in the photosphere of the star , and want to determine the field only in the accretion spot at the stellar surface , by measuring the narrow components of the emission lines , then the success of the measurement is determined by the flux and shape of the profile of the lines , rather than by the brightness of the star .
having considered a few dozen spectra of ctts , obtained at the vlt and keck observatories , we chose 8 relatively faint stars with the strong and narrow emission line of hei 5876 , the magnetic field of which had never been measured before . in our paper
the measurements for three of these eight stars are presented : do tau ( spectral type m0 , range of variability in @xmath8 @xcite ) , dr tau ( k5 , @xmath9 @xcite ) , ds tau ( k5 , @xmath10 @xcite ) .
the method we use to measure magnetic field is based on the fact that so - called @xmath11-components resulted in zeeman splitting are polarized circularly such as oppositely polarized components are located on different sides of the central wavelength @xmath12 . if magnetic field in a line formation region has a longitudinal component @xmath13 then the right- and left - hand polarized components will be shifted relative to each other to @xcite : @xmath14 where @xmath15 is the lande @xmath15-factor of the considered line , @xmath16 and @xmath12 are in and @xmath13 is in kg .
this relation allows to find the longitudinal magnetic field component @xmath17 averaged over the line formation region , by measuring @xmath16 from two spectra observed in the right- and left - hand polarized light .
observations were carried out on 2012 october 26 - 27 with the main stellar spectrograph @xcite of 6-m telescope of special astrophysical observatory equipped with a polarization @xmath18-plate and a double slicer @xcite .
the spectrograph slit width was @xmath19 the observations were carried out in the spectral range 5640 - 6480 by using ccd array ( e2v ccd42 - 90 ) , the size of which along the dispertion was 4600 pixels that corresponds to the inverse linear dispersion of 0.183 / pixel .
spectra were processed as follows @xcite .
night sky emission and detector bias as well as cosmic ray traces were removed in a standard way , using routines from the midas software package .
a spectrum of a thorium - argon lamp was used for a wavelength calibration .
each observed spectrum was transformed into the stellar rest frame by shifting it as a whole ( by radial velocity ) until a coincidence of photospheric lines in the observed and simulated spectra @xcite .
the log of the observations is presented in table [ tab - zhurnal ] , which contains julian date jd of the middle of an observation , an exposure time and a signal - to - noise ratio s / n .
c | c | c | c | c n & jd 2456220 + & star & @xmath20 sec & s / n + 1 & 7.181 & @xmath21 equ & 180 & 360 + 2 & 7.186 & @xmath21 equ & 180 & 360 + 3 & 7.299 & hd216228 & 100 & 180 + 4 & 7.324 & ds tau & 1200 & 62 + 5 & 7.354 & ds tau & 1200 & 64 + 6 & 7.383 & ds tau & 1200 & 60 + 7 & 7.420 & dr tau & 1200 & 123 + 8 & 7.449 & dr tau & 1200 & 119 + 9 & 7.479 & dr tau & 1200 & 113 + 10 & 7.510 & do tau & 1200 & 83 + 11 & 7.540 & do tau & 1200 & 78 + 12 & 7.628 & hd31398 & 100 & 110 + 13 & 7.638 & 53 cam & 300 & 370 + 14 & 8.212 & @xmath21 equ & 180 & 310 + 15 & 8.336 & ds tau & 1200 & 72 + 16 & 8.365 & ds tau & 1200 & 70 + 17 & 8.394 & ds tau & 1200 & 78 + 18 & 8.429 & do tau & 1200 & 49 + 19 & 8.458 & do tau & 1200 & 49 + 20 & 8.488 & do tau & 1200 & 44 + 21 & 8.521 & dr tau & 1200 & 88 + 22 & 8.550 & dr tau & 1200 & 89 + 23 & 8.579 & dr tau & 1200 & 84 + 24 & 8.639 & 53 cam & 300 & 350 + + to exclude systematic instrumental errors , our observations were organized as follows .
one measurement of the field needed two exposures , between which the superachromatic quarter - wave phase plate was rotated in such a way that the right- and left - hand polarized spectra changed places on the ccd array .
thus , we get four spectra of the star : right- and left - hand polarized at the first exposure @xmath22 , @xmath23 and similar spectra of @xmath24 , @xmath25 at the second exposure .
further we calculated the difference between the positions of the lines in the spectra @xmath22 and @xmath23 , which we denoted as @xmath26 , between @xmath24 and @xmath25 , which we denoted as @xmath27 and between @xmath28 and @xmath29 which we denoted as @xmath30 here @xmath31 denotes the sum of simultaneously obtained spectra .
these differences were calculated by using the crosscorrelation method @xcite for the confidence level @xmath32 that corresponds to 1@xmath11 error .
the shift due to zeeman effect is equal to @xmath33 a mean systematic shift between @xmath34 and @xmath35 : @xmath36 and a systematic shift between two exposures : @xmath37 all shifts were calculated in pixels , and then , if necessary , were translated in .
we used this method to measure @xmath13 on the lines hei 5876 , nai d and [ oi ] 6300 .
however , in the case of the field measurements on many lines , it is appropriate to measure the shifts all considered lines at once . in this case
, the procedure of measurements @xmath16 should be refined .
the shifts @xmath38 @xmath39 , calculated by the crosscorrelation method , include two shifts of different nature : @xmath16 and @xmath40 on the one hand , the shift @xmath16 is different for each spectral line , therefore to find @xmath38 by calculating the maximum of the correlation function , we should choose the value of @xmath13 as an independent variable , because it is the same for all lines , while the shift of each line should be calculated by the formula ( 1 ) .
on the other hand , the systematic shift @xmath41 is assumed to be constant for the entire spectrum , therefore to find it , we should choose a shift of the spectrum as a whole as the independent variable . to combine these contradictory requirements
, we had found an average systematic shift @xmath42 , assuming that it behaves like @xmath16 and then we corrected the spectrum for this shift @xmath43 because @xmath44 this correction does not eliminate the systematic error , but only reduces it . however , 3 - 4 iterations of the procedure allow to eliminate the systematic shift almost entirely .
an uncertainty of the value @xmath16 , defined in this manner , equals to @xmath45 lande factors of photospheric lines were adopted from the vald database @xcite . in the case of blended photospheric lines @xmath15-factors were averaged with weights of a central depth : @xmath46 where @xmath47 is a gaussian profile of an @xmath48-th line in the blend .
the width of the gaussian was found from the observed spectrum , and the amplitude was equated to the line depth from the vald database .
we take into account all photospheric lines with known lande factors @xmath49 which fall in our spectral range and have a central depth greater than 0.1 .
lande factors for emission lines , considered in our paper : hei 5876 ( excitation potential @xmath50 ev ) , [ oi ] 6300 ( @xmath51 ev ) and nai 5890 , 5896 ( @xmath52 ev ) were assumed to be 1.1 , 1.0 , 1.33 , correspondingly . to test the discribed method ,
we have observed stars with a known magnetic field : 53 cam and @xmath21 equ , as well as giants : hd216228 and hd31398 , the field of which should be close to zero .
the results of the measurements @xmath13 are presented in the table[standart ] .
it can be seen from the table that the values of @xmath13 are equal to zero within its uncertanties in case of hd216228 and hd31398 and , in case of @xmath21 equ , @xmath13 agree with the results , obtained by kudryavtsev & romanyuk @xcite , who have found that for this star @xmath13 varies from @xmath53 to @xmath54 kg . these results convince us of the correctness of the chosen method . in case of 53 cam , the result of second measurement coincides within the error with the expected value , calculated from the formulas from hill s paper @xcite , but @xmath13 of first measurement differs noticeably from the predicted value .
note that similar and even greater deviations from the predicted values were also observed in 53 cam by other authors @xcite .
|c c| c n & jd 2456220 + & star & @xmath13 & @xmath55 & @xmath56 + 1 & 7.181 & @xmath21 equ & @xmath57 & @xmath58 & @xmath59 + 2 & 7.186 & @xmath21 equ & @xmath60 & @xmath58 & @xmath59 + 3 & 7.299 & hd 216228 & @xmath61 & @xmath62 & @xmath63 + 12 & 7.628 & hd 31398 & @xmath64 & @xmath62 & @xmath63 + 13 & 7.638 & 53 cam & @xmath65 & @xmath58 & @xmath66 + 14 & 8.212 & @xmath21 equ & @xmath67 & @xmath58 & @xmath59 + 24 & 8.639 & 53 cam & @xmath68 & @xmath69 & @xmath70 + +
ctts usually have a period of axial rotation about a week @xcite , therefore it can be expected that the field could not change significantly during the time of the observation of every star within each night ( around 2 hours ) . hence , there are reasons to calculate an average @xmath13 for each night : @xmath71 an error @xmath72 of an individual observation was estimated from the uncertanties of @xmath73 @xmath27 , as it was described in the previous section .
a scatter of the individual observations @xmath74 during one night may be caused by the noise of the spectrum . in this case
an uncertainty of the average can be estimated as : @xmath75^{-1/2}.\ ] ] however , we can not exclude that the scatter may be caused by some other random process , for which the standard deviation can be estimated as follows : @xmath76\right)\times\sqrt{\frac{1}{(n-1 ) \sum(1/\sigma_{i}^{2 } ) } \sum{\frac{(b_{zi}-\overline{b_z})^2}{\sigma_{i}^{2}}}},\ ] ] where @xmath77 is the confidence level corresponding to 1@xmath78 @xmath79 is the number of measurements being averaged , @xmath80 the inverse of student s t cumulative distribution function .
if there are both processes , and they are independent , then a total error of the average over night can be estimated as @xmath81 the values of @xmath82 and @xmath83 are given in the table.[tts ] .
c | c | c c| c c | c c | c c star & n n & @xmath82 & @xmath83 & @xmath82 & @xmath83 & @xmath82&@xmath83 & @xmath82&@xmath83 + & & & & & + do tau & 1011 & @xmath84 & @xmath85 & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 + & 1820 & @xmath92 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & @xmath99
+ dr tau & 79 & @xmath100 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath93 & @xmath105 & @xmath106 + & 2123 & @xmath107 & @xmath108 & @xmath109 & @xmath87 & @xmath64 & @xmath89 & @xmath110 & @xmath111 + ds tau & 46 & @xmath112 & @xmath113 & @xmath114 & @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 + & 1517 & @xmath119 & @xmath120 & @xmath114 & @xmath114 & @xmath121 & @xmath118 & @xmath122 & @xmath123 + it follows from the table [ tts ] that the fields have been detected in the formation region of the narrow component of the line hei 5876 in the stars do tau and ds tau at a level @xmath124 and in the star dr tau at a level @xmath125 .
examples of @xmath126profiles for the given line for these stars are presented on the fig.[example ] .
vertical dashed lines mark the part of the profile ( 21 pixels ) , which was used in the measurement .
a variation of the equivalent width of the line from night to night can be caused either by the stellar rotation or by changes of physical conditions in the accretion spot .
it manifested most clearly in the case of ds tau : the line of hei 5876 in first night was approximately six times smaller than in second night . and
@xmath127profiles near the line hei 5876 for the stars do tau , dr tau , ds tau .
the left and right panels correspond to the observation on october 26 and 27 , correspondingly .
the number in the upper right corner of each panel corresponds to the number in the table[tab - zhurnal ] .
the thick curve indicates the observed @xmath126profile .
the dots indicate the observed @xmath127profile .
the thin curve indicates the predicted @xmath127profile for the field value from the table[tts ] .
both @xmath128-profiles were multiplied by 10 times and shifted to 0.8 for readability . dashed lines
mark the continuum level and a spectral range , on which we measure the field .
an abscissa is the velocity in kms@xmath6 from the line center @xmath129 in the stellar reference frame . ]
dots on the figure correspond to an `` observed '' @xmath128-profile , which was calculated by using the left- and right - polarized spectra , summarized pixel by pixel over the two exposures and corrected for @xmath41 and @xmath130 @xmath131 @xmath132 @xmath133 the shifts @xmath41 and @xmath134 are about tenths of a pixel , therefore to calculate the values of @xmath24 and @xmath25 for fractional pixel values , we use a linear interpolation , which leads to a slight smoothing of the spectra .
a thin line on the figure shows a `` theoretical '' @xmath128-profile , which was calculated as a relative difference of two @xmath135-profiles , the first of which was shifted by @xmath136 to the left and the second was shifted by the same amount to the right . in case of
do tau and dr tau the broad emission component ( see fig.[nai ] ) dominated in nai d lines , the field on which was measured on both line simultaneously on the spectral range 5885.6 5900.3 .
results of the measurements are collected in the the table[tts ] and they show that the value @xmath82 in the broad emission components of nai d is equal to zero with accuracy better than @xmath137 that is typical for the broad components of emission lines in ctts ( see the introduction ) .
profiles near the line nai d for the stars do tau , dr tau , ds tau at the first night ( the left column ) and at the second night ( the right corner ) .
the number in the upper right corner of each panel corresponds to the number in the table[tab - zhurnal ] .
dashed lines mark the continuum level and a spectral range , on which we measure the field .
an abscissa is the velocity in kms@xmath6 from the line center @xmath138 in the stellar reference frame . ]
the broad emission component of nai d lines is absent in ds tau and the lines consist from broad photospheric absorption wings and the narrow emission component .
the value of @xmath17 measured on the narrow components and averaged over first night ( numbers 4 - 6 from the table.[tab - zhurnal ] ) , was equal to @xmath139 kg .
in other words , the field in a line formation region of nai d as well as hei 5876 has not been detected in this night . however , similar measurements in the next night ( the observations 15 - 17 ) gave the values @xmath140 , which well coincided with @xmath141 measured on hei 5876 : @xmath142 kg . the field measured on the forbidden line of [ oi ]
6300 is always inside its uncertainty , because the line are formed at low gas densities in a large volume , where the magnetic field is close to zero .
the measurements of this line were carried out to check the absence of large systematic errors in the method of processing the spectrum .
the field , measured on photospheric lines , in all considered stars is equal to zero within its uncertainty .
this seems to be due to two effects .
first , we observed faint ctts , and the signal - to - noise ratio for them was usually much lower than , for example , in case of the test stars , moreover , the radiation of the accretion spot leads to a significant shallowing the absorption lines in ctts spectra ( so - called veiling ) , that increases the error of @xmath143 second , we measure only the longitudinal component of the field . hence ,
even if the field in the photosphere is the same as in the line formation region of hei 5876 , then the longitudinal component of the field , averaged over the entire visible surface of the star , is several times lower than in the line hei 5876 , for which the field is averaged only over the accretion spot .
the possibility of determining the magnetic field in the accretion spot in a relatively faint ctts by measuring the zeeman splitting of the narrow component of strong emission lines in their spectra has been demonstated on examples of three stars . following the tradition , we have measured the field on the line hei 5876 , but lines of other elements , which as well as hei 5876 are formed in the accretion spot ,
can be also used in the future .
the measured values of @xmath13 are the values of the longitudinal component of the field , which is averaged over the accretion spot with weights of the line intensity in each point of the spot .
dodin et al .
@xcite have shown that there are lines in spectra of ctts , the intensities of which vary in different ways depending on the accretion flux @xmath144 where @xmath145 and @xmath146 are a pre - shock gas density and velocity , correspondingly .
for instance , the intensity of cai lines is increased with increasing of @xmath147 and vice versa caii lines get smaller .
it gives a possibility to use such lines for the measurements of the magnetic field in parts of the accretion spot with different values of the accretion flux @xmath148 we wish to thank the large telescopes program committee for providing time for the observations on 6-m telescope .
the observations on the 6-m telescope are held with a financial support of ministry of science and education of the russian federation .
the work was supported by the program for support of leading scientific schools ( nsh-5440.2012.2 ) .
99 c. bertout , ann .
astron . astrophys .
* 27 * , 351 ( 1989 ) .
dodin , s.a .
lamzin , astron . lett . * 38 * , 649 ( 2012 ) . c.c .
batalha , n.m .
stout - batalha , g. basri , m.a.o .
nerra , astrophys .
. ser . * 103 * , 211 ( 1996 ) .
dodin , s.a .
lamzin , g.a .
chountonov , astron .
lett . * 38 * , 167 ( 2012 ) .
hussain , a.c .
cameron , m.m .
jardine et al . , mon . not .
. soc . * 398 * , 189 ( 2009 ) .
donati , s.g .
gregory , s.h.p .
alencar et al . , mon . not .
. soc . * 417 * , 472 ( 2011b ) .
donati , j. bouvier , f.m .
walter et al . , mon . not .
412 * , 2454 ( 2011a ) .
donati , m.b .
skelly , j. bouvier et al . ,
mon . not .
. soc . * 402 * , 1426 ( 2010a ) .
donati , m.b .
skelly , j. bouvier et al . ,
mon . not .
. soc . * 409 * , 1347 ( 2010b ) .
donati , m.m .
jardine , s.g .
gregory et al . , mon . not .
. soc . * 386 * , 1234 ( 2008 ) c.m .
johns - krull , astrophys .
j. * 664 * 975 , ( 2007 ) .
h. yang , c.m .
johns - krull , astrophys .
j. , * 729 * , 83 ( 2011 ) .
n.h . symington , t.j .
harries , r. kurosawa , t. naylor , mon . not .
. soc . * 358 * , 977 ( 2005 ) .
chuntonov , d.a .
smirnov , s.a .
lamzin , astron . lett . * 33 * , 38 ( 2007 ) .
smirnov , s.a .
lamzin , s.n .
fabrika , g.a .
chuntonov astron .
lett . * 30 * , 456 ( 2004 ) .
w. herbst , d.k .
herbst , e.j .
grossman , d. weinstein , astron .
j. * 108 * , 1906 ( 1994 ) .
grankin , s.yu .
melnikov , j. bouvier , w. herbst , v.s .
shevchenko , astron .
astrophys . * 461 * , 183 ( 2007 ) .
babcock , astrophys .
j. suppl . ser . * 3 * , 141 ( 1958 ) .
v.e panchuk , preprint sao ras n 154 ( 2001 ) g.a .
chuntonov , _ magnetic stars : proc.of the intern.conf._ , n.arkhyz , eds .
glagolevskij , d.o .
kudryavtsev , i.i .
romanyuk , moscow , 286 ( 2004 ) .
johnstone , m.v .
penston , mon . not .
r. astron .
219 * , 927 ( 1986 ) .
f. kupka , n. piskunov , t.a .
ryabchikova et al . , astron .
. ser . * 138 * , 119 ( 1999 ) .
d.o . kudryavtsev and i.i .
romanyuk , astron . nachr .
* 333 * , 41 ( 2012 ) .
hill , d.a .
bohlender , j.d .
landstreet et al . , mon . not .
r. astron .
297 * , 236 ( 1998 ) .
artemenko , k.n .
grankin , p.p .
petrov , astron .
* 38 * , 783 ( 2012 ) .
johns - krull , j.a .
valenti , a.p .
hatzes , a. kanaan , astrophys . j. * 510 * , l41 ( 1999 ) .
dodin , s.a .
lamzin , t.m .
sitnova , astron .
39 * , in press , arxiv:1302.1825 ( 2013 )
|
results of measurements of the longitudinal magnetic field in a hot accretion spot in three classical t tauri stars ( ctts ) are presented . the magnetic field in the formation region of the narrow component of the emission line hei 5876 was found for each star in our sample at a level of more than 2@xmath0 in case of ds tau we have found the field in the narrow components of nai d lines , which was equal to @xmath1 kg , i.e. it was equal to the field measured on the narrow component of hei 5876 .
our results indicate that the magnetic field in the hot spots can be studied for ctts down to 13@xmath2 that allow in the future to double a number of ctts with measured field in the accretion zone .
|
= 15000 = 15000 a major payoff period for the _ hubble space telescope _
( hst ) is under way , as the space telescope imaging spectrograph ( stis ) starts to produce black hole detections routinely and in large numbers .
eight new bh detections are available from nuker team work ( gebhardt et nuk .
2000b ) , and improved observations or new detections are available for 6 more ( verdoes kleijn et al .
2000 ; bower et al .
2000 ; green et al .
2000 ; kaiser et al . 2000 ; nelson et al .
the total number of bhs available for demographic studies is now at least 34 .
this is enough for a quantum improvement in our ability to ask astrophysical questions .
stis is well designed for black hole searches based on stellar or ionized gas dynamics .
in addition , maser disks have provided bh detections in three galaxies , including the spectacular case of ngc 4258 ( miyoshi et al .
what dynamical searches securely detect are central dark objects with masses the arguments that these dark masses are bhs and not dark clusters of ( e.g. ) brown dwarf stars , white dwarf stars , neutron stars , or stellar - mass black holes are indirect . however , in ngc 4258 and in our galaxy , the maximum possible radius that a dark cluster can have is so small that astrophysical constraints ( collision times for brown dwarfs and evaporation times for stellar remnants ) rule out the above bh alternatives ( maoz 1998 ) .
this result increases our confidence that the other central dark objects also are bhs . with stis ,
the bh search has become feasible for most nearby galaxies that have unobscured centers and reasonably old stellar populations .
the search is not easy ; e.g. , searching for a 10@xmath0-@xmath1 bh is still difficult at the distance of the virgo cluster and impossible much beyond .
but well chosen samples provide reliable detections or upper limits on @xmath2 that are suitable to address a variety of astrophysical questions .
many observing programs of this sort are currently under way and are , as of mid-2000 , just beginning to yield results .
my oral paper reviewed four particularly good bh cases , ngc 4258 ( the maser galaxy ) , ngc 4374 ( which contains an ionized gas disk with a very clean bh signature bower et al .
1998 ) , ngc 3115 ( whose nucleus , with an observed dispersion of 600 @xmath3 37 km s@xmath4 , would be unbound without a dominant dark object kormendy et nuk .
1996 ) , and our galaxy .
the latter is now the strongest bh case , based on measurements of stellar proper motions in a star cluster within @xmath5 = 0.02 pc of sgr a * ( eckart & genzel 1997 ; genzel et al .
1997 , 2000 ; ghez et al .
the fastest - moving star has a total proper motion of @xmath6 km s@xmath4 .
genzel and ghez and we , as spectators can look forward to observing the galactic center rotate once in our lifetimes !
already preliminary accelerations have been measured ( ghez et al .
one reason why this is important is that the three available acceleration vectors do in fact intersect near sgr a*. in this written version of my paper , i concentrate on results provided by the new bh detections .
two fundamental correlations have emerged from bh detections .
figure 1 ( left ) shows the correlation between bh mass and the luminosity of the `` bulge '' part of the host galaxy ( kormendy 1993a , kormendy & richstone 1995 ; magorrian et nuk .
1998 ) brought up to date with new detections .
a least - squares fit gives @xmath9 since @xmath10 , equation ( 1 ) implies that bh mass is , within errors , proportional to bulge mass : @xmath11 .
figure 1 ( right ) shows a newly discovered correlation between bh mass and the luminosity - weighted velocity dispersion @xmath8 within the effective radius @xmath12 ( gebhardt et nuk .
2000a ; ferrarese & merritt 2000 ) .
the line is a fit to the data assuming that errors in @xmath8 are zero and that errors in @xmath13@xmath2 are the same for each galaxy .
a least - squares fit to the subset of galaxies with most reliable @xmath2 measurements ( see gebhardt et nuk .
2000a ) gives @xmath14 figure 1.1.5em(left ) correlation of bh mass with the absolute magnitude of the bulge component of the host galaxy .
( right ) correlation of bh mass with the luminosity - weighted mean velocity dispersion inside the effective radius of the bulge . in both panels ,
filled circles indicate @xmath2 measurements based on stellar dynamics , squares are based on ionized gas dynamics , and triangles are based on maser disk dynamics .
all three techniques are consistent with the same correlations .
however , note that bh mass error bars are derived within the assumptions of the models used , so the true errors are underestimated for @xmath2 values based on two - integral models .
this is particularly relevant for the most discrepant galaxy at left , ngc 4486b .
the scatter in the @xmath2
@xmath15 correlation is large : the rms dispersion is a factor of 2.8 and the total range of bh masses is about two orders of magnitude at a given @xmath15 .
there are also two exceptions with unusually high bh masses .
the most extreme case , ngc 4486b ( kormendy et nuk .
1997 ; green et al .
2000 ) , is still based on two - integral models .
its bh mass may decrease when three - integral models are constructed . despite the scatter ,
the 34 detections imply a robust correlation .
many authors have wondered whether the apparent @xmath2
@xmath15 correlation is real or only the upper envelope of a distribution that extends to smaller @xmath2 .
the latter possibility is now unlikely : ongoing searches find bhs in essentially every bulge observed , and the error bars in @xmath2 are small . put another way , almost all bhs represented in figure 1 would have been detected even they if were significantly farther away .
in contrast , the scatter in the @xmath2
@xmath8 correlation is small , and the galaxies that were discrepant above are not discrepant here .
gebhardt et nuk .
( 2000a ) show that the scatter is essentially consistent with measurement errors for galaxies with the most reliable @xmath2 measurements .
therefore the @xmath2
@xmath8 correlation is more fundamental than the @xmath2
@xmath15 correlation .
what does this mean ?
what does it tell us about galaxy formation and bh growth ?
both correlations imply a close connection between bh growth and galaxy formation .
they suggest that the bh mass is determined in part by the amount of fuel that is available ; this is connected with the total mass of the bulge .
but when was the mass accreted ?
there are three generic possibilities .
( 1 ) bhs could have grown to essentially their present masses before galaxies formed and then regulated the amount of galaxy that grew around them ( e.g. , silk & rees 1998 ) . ( 2 ) seed bhs that formed early or that were already present at the start of galaxy formation could have grown to their present masses as part of the galaxy formation process .
this would mean that the major merger and dissipative collapse events that made a bulge or elliptical galaxy were the same events that made quasars shine ( e.g. , sanders et al .
1988a , b ) .
( 3 ) the majority of bh mass may have been accreted after galaxy formation from ambient gas in the bulge .
figure 1 provides the first of two arguments ( the other is in 4 ) in favor of ( 2 ) , that bhs and galaxies form together .
exceptions to the @xmath2
@xmath15 correlation satisfy the @xmath2
@xmath8 correlation .
this means that , whenever a bh is unusually high in mass for a given luminosity , it is also high in @xmath8 for that luminosity .
that is , it is high in the faber - jackson ( 1976 ) @xmath16 correlation .
one possible reason might be that the mass - to - light ratio of the stars is unusually high ; this proves not to be the main effect .
the main effect is illustrated in figure 2 .
ellipticals that have unusually high dispersions for their luminosities are unusually compact : they have unusually high surface brightnesses and small effective radii for their luminosities .
similarly , unusually cold galaxies are unusually fluffy : they have low effective surface brightnesses and large effective radii .
therefore , when a galaxy is observed to be hotter than average , we conclude that it underwent more dissipation than average and shrunk inside its dark halo to a smaller size and higher density than average .
we will say that it `` collapsed '' more than average .
_ if bhs are unusually massive whenever galaxies are unusually collapsed , then this strongly suggests that bh masses were determined by the collapse process .
_ consider the alternative : could unusually massive bhs _ cause _ a galaxy to be unusually collapsed .
this seems unlikely , because more massive bhs would power more luminous quasars .
their radiation pressure would tend to push on the protogalactic gas and make it collapse less , not more , than average .
the @xmath2 @xmath15 and @xmath2 @xmath8 correlations contain _ almost _ the same information .
after all , the mass of the bulge involves @xmath17 and ( small ) corrections for velocity anisotropy .
but the @xmath2 @xmath8 correlation contains something new : it contains information about how much the bulge collapsed when it formed .
bh mass correlates better with this new information about the galaxy formation process than it does with the amount of raw material involved .
@xmath18 ; it is the presence of @xmath12 that makes a galaxy an exception to the @xmath2
@xmath15 correlation when it is not an exception to the @xmath2 @xmath8 correlation .
all this emphasizes the close connection between bh masses and the bulge formation process .
figure 2.1.5emcorrelations with absolute magnitude of velocity dispersion ( upper panel ) , effective surface brightness ( middle panel ) and effective radius ( lower panel ) for elliptical galaxies from the seven samurai papers .
galaxies with unusually high or low velocity dispersions for their luminosities are identified in the top panel and followed in the other panels .
it is important to note that bh mass does not correlate with disks in the same way that it does with bulges .
figure 3 shows the correlations of bh mass with the absolute magnitude of the bulge ( left ) and with the total absolute magnitude of the galaxy ( right ) .
( right ) shows that disk galaxies with small bulge - to - total luminosity ratios destroy the reasonably good correlation seen in figure 3 ( left ) .
in addition , figure 3 ( right ) shows four galaxies that have strong bh mass limits but no bulges .
they further emphasize the conclusion that disks do not contain bhs with nearly the same mass fraction as do bulges .
in particular , in the bulgeless galaxy m33 , the upper limit on a bh mass from stis spectroscopy is @xmath2 @xmath20 2000 @xmath1 . if m33 contained a bh with the median mass fraction observed for bulges , then we would expect that @xmath21 @xmath1 .
figure 3 tells us that bh masses do not `` know about '' galaxy disks .
rather , they correlate with the high - density bulge - like component in galaxies .
the above results do not preclude bhs in pure disk galaxies as long as they are small enough .
filippenko & ho ( 2000 ) emphasize that some pure disks are seyfert galaxies .
they probably contain bhs .
the extreme example is ngc 4395 , the lowest - luminosity seyfert known .
ngc 4395 is included in figure 3 .
however , if its bh were radiating at the eddington rate , then its mass would be only @xmath22 @xmath1 ( filippenko & ho 2000 ) .
this is consistent with the dynamical mass limit .
so disks can contain bhs , but their masses are _ much _ smaller in relation to their disk luminosities than are bulge bhs in relation to bulge luminosities .
it is possible that the small bhs in disks are similar to the seed bhs that once must have existed in protobulges , too , before they grew monstrous during the agn and galaxy formation era .
figure 3.1.5em(left ) @xmath2 @xmath15 correlation from figure 1 .
( right ) plot of @xmath2 against the total absolute magnitude of the host galaxy .
blue points denote elliptical galaxies , red points denote bulges of disk galaxies , and green crosses denote galaxies that do not contain either a bulge or a pseudobulge ( 4 ) .
spiral galaxy limits : m33 ( scd ) is from gebhardt et al .
( 2000b ) ; ic 342 ( scd ) is from bker et al .
( 1999 ) , and ngc 4395 ( sm ) is from filippenko & ho ( 2000 ) .
so far , i have discussed elliptical galaxies and the bulges of disk galaxies as if they were equivalent . in terms of bh content , they are indistinguishable : they are consistent with the same @xmath2 @xmath15 and @xmath2 @xmath8 correlations . but a variety of observational and theoretical results show that there are two different kinds of high - density central components in disk galaxies .
both have steep surface brightness profiles .
but , while classical bulges in ( mostly ) early - type galaxies are like little ellipticals living in the middle of a disk , the `` pseudobulges '' of ( mostly ) late - type galaxies are physically unrelated to ellipticals .
pseudobulges are reviewed in kormendy ( 1993b ) and in combes ( 2000 ) .
observational evidence for disklike dynamics includes ( i ) velocity dispersions @xmath23 that are smaller than those predicted by the faber - jackson ( 1976 ) @xmath23 @xmath24 correlation , ( ii ) rapid rotation @xmath25 that implies @xmath26 values well above the `` oblate line '' describing rotationally flattened , isotropic spheroids in the @xmath26 ellipticity diagram , and ( iii ) spiral structure that dominates the pseudobulge part of the galaxy .
these observations and @xmath27-body simulations imply that high - density central disks can form out of disk gas that is transported toward the center by bars and oval distortions .
these heat themselves , for example , by the scattering of stars off of bars ( pfenniger & norman 1990 ) .
kormendy ( 1993b ) concludes that most early - type galaxies contain traditional bulges , that later - type galaxies tend to contain pseudobulges , and that only pseudobulges are seen in sc
sm galaxies .
andredakis & sanders ( 1994 ) , andredakis , peletier , & balcells ( 1995 ) , and courteau , de jong , & broeils ( 1996 ) show that the `` bulges '' of many late - type galaxies have nearly exponential surface brightness profiles .
it is likely that exponential profiles are a signature of pseudobulges , especially since blue colors imply that they are younger than classical bulges ( balcells & peletier 1994 ) .
hst observations strengthen the evidence for pseudobulges .
carollo et al.(1997 , 1998a , b ) find that many bulges have disky properties , including young stellar populations , spiral structure , central bars , and exponential brightness profiles .
it seems safe to say that no - one who saw these would suggest that they are mini - ellipticals living in the middle of a disk .
rather , a morphologist who saw these structures would assign late even i m hubble types . to be sure , peletier et al .
( 2000 ) find that bulges of early - type galaxies generally have red colors : they are old .
true bulges that are similar to elliptical galaxies do exist ; m31 and ngc 4594 contain examples .
but the lesson from the carollo papers is that pseudobulges are more important than we expected . like kormendy ( 1993b ) and courteau et al .
( 1996 ) , carollo and collaborators argue that these are not real bulges but instead are formed via gas inflow in disks .
so there is growing evidence that the `` bulges '' whose luminosities i used in the @xmath2
@xmath15 correlation diagrams are two different kinds of objects .
classical bulges are thought to form like small ellipticals , via a dissipative collapse , possibly triggered by a merger .
pseudobulges are thought to form by secular evolution of the disk .
material flows inward in both cases .
potentially , both formation processes may include bh feeding .
one way to explore this is to ask whether bulges and pseudobulges have the same bh content .
figure 4.1.5emthe @xmath2
@xmath15 ( left ) and @xmath2 @xmath8 ( right ) correlations for elliptical galaxies ( blue symbols ) , traditional bulges ( red ) and pseudobulges ( green ) .
this question is addressed in figure 4 .
it is plausible that pseudobulges have low luminosities , because they are made from disks .
but for their low luminosities , they have normal bh masses .
the indentification of pseudobulges is still somewhat uncertain , and there are not many in the sample
. therefore the above result needs to be checked .
however , it is consistent with the hypothesis that ( pseudo)bulge formation and bh feeding are closely connected .
present data do not show any dependence of @xmath2 on the details of whether bh feeding happens rapidly during a collapse or slowly via secular evolution of the disk .
if disks contain only small bhs while the pseudobulges that form from disks contain standard bhs with 0.2% of the pseudobulge mass , then i conclude that the bhs must have grown to their present masses during pseudobulge formation .
this is the second argument that bhs and ( pseudo)bulges grew together .
the smallest bhs provide an argument that most bh growth did not happen after bulge formation .
some pseudobulges are still forming now ; there is little time after bulge formation .
also , these objects do not contain fuel in the form of x - ray gas . and
galaxies like m32 contain little gas of any sort for late accretion . -20pt
galaxy formation is notoriously complicated ; any conclusions that we reach now are less secure than the observational results discussed in however : _ observations increasingly suggest that the major events that form a bulge and the major growth phases of its bh when it shone as an agn were the same events .
_ the likely formation process is a series of dissipative mergers that fuel both starbursts and agn activity ( sanders et al .
1988a , b ) .
it is a pleasure to thank my nuker collaborators and especially karl gebhardt and douglas richstone for helpful discussions and for permission to use our bh detection results before publication .
i am also most grateful to gary bower , richard green , mary beth kaiser , and charles nelson for communicating stis team bh detections before publication .
|
dynamical searches find central dark objects candidate supermassive black holes ( bhs ) in at least 34 galaxies .
the demographics of these objects lead to the following conclusions : \(1 ) bh mass correlates with the luminosity of the bulge component of the host galaxy , albeit with considerable scatter .
the median bh mass fraction is 0.15% of the mass of the bulge .
the quartiles are 0.09% and 0.7% .
\(2 ) bh mass correlates with the mean velocity dispersion of the bulge component inside its effective radius , i.e. , with how strongly the bulge stars are gravitationally bound to each other . for the best bh mass determinations , the scatter is consistent with the measurement errors .
\(3 ) bh mass correlates with the luminosity of the high - density central component in disk galaxies independent of whether this is a real bulge ( a mini - elliptical ) or a `` pseudobulge '' ( believed to form via inward transport of disk material ) .
\(4 ) bh mass does not correlate with the luminosity of galaxy disks .
if pure disks contain bhs ( and agn observations argue that some do ) , then their bh mass fractions are much smaller than the canonical 0.15% for bulges .
these results lead to the following conclusions : \(a ) present observations show no dependence of bh mass on the details of whether bh feeding happens rapidly during a collapse or slowly via secular evolution of the disk .
\(b ) the above results increasingly support the hypothesis that the major events that form a bulge or elliptical galaxy and the main growth phases of its bh when it shone as an agn were the same events .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in
|
neutral atoms trapped in an artificial periodic potential formed by laser light , the so called far detuned optical lattice , have been proposed as the individual qubits for quantum information processing . in an optical lattice
, neutral atoms can be trapped in the intensity maxima ( or minima ) of a standing wave light field owing to the optical dipole force . a configuration with one single atom trapped in each site of the optical lattice is realized in the configuration of a mott - insulator transition associated to the loading of bose - einstein condensates ( becs ) in optical lattices@xcite .
in order to realize quantum gates with neutral atoms within the ideal environment of the mott insulator several schemes have been proposed .
the common idea is to control the quantum atomic states through the preparation and coherent manipulation of atomic wave - packets by means of application of standard laser cooling and spectroscopic techniques . by using spin dependent , or more precisely state dependent , optical lattice potentials ,
the control can be applied independently to multiple atomic qubits based on different internal states .
a state dependent
potential may be created for a one dimensional optical lattice in the so - called lin-@xmath1-lin configuration , where the travelling laser beams creating the optical lattice are linearly polarized with an angle @xmath1 between their polarizations @xcite . in this configuration
the optical potential can be expressed as a superposition of two independent optical lattices , acting on different atomic states . by appropriately choosing the atomic internal states ,
the atoms will be trapped by one of the two potentials depending on their internal state . by changing the angle @xmath1 between the linear polarizations of the two laser beams producing the optical lattice ,
the wavepackets corresponding to orthogonal atomic states can be coherently transported relative to each other @xcite .
once the atomic qubits are brought together they interact through controlled collisions . in the coherent transport experiment of mandel _
et al _ @xcite , by a proper control of the angle @xmath1 the wavepacket of an atom initially localized at a single lattice site was split into a superposition of two separate wave packets , and delocalized in a controlled and coherent way over a defined number of lattice sites of the optical potential .
+ in an optical lattice created by the counterpropagating standing wave configuration , the spacing between neighboring minima of the optical lattice potential is one half the wavelength of the lasers creating the optical lattice . optical lattices with more widely separated wells can be produced using long wavelength lasers , as co@xmath2 lasers @xcite .
alternatively , optical lattices with a larger spacing between the minima / maxima of the optical potential are formed using the angle - tuned geometry .
there the periodic potential is created by two laser beams propagating at an angle @xmath3 and the lattice constant is @xmath4$],with @xmath5 the laser wavenumber @xcite .
+ the aim of the present work is to analyze the coherent transport associated to the lin-@xmath1-lin polarization configuration for the angle - tuned lattice geometry .
we analyze rubidium atoms in a given zeeman level of a hyperfine state loaded within a 1-d optical - lattice .
the 1-d geometry of the bose gas may be generated by a tight confinement along the orthogonal directions .
for instance a two - dimensional array of 1-d bose gases ( tubes ) is produced by confining the atoms through a two - dimensional optical lattice generated by independent lasers , as realized in ref .
+ section ii defines the geometry of the angle - tuned optical lattice . in section iii
we analyze the effective optical potential created by an optical lattice in the angle - tuned configuration .
the potential contains a component with a vectorial symmetry described through an effective magnetic field , as derived in @xcite .
the dependence of the potential depth on the angles defining the lattice geometry is analyzed in section iv .
section v reexamines the coherent transport for the counterpropagating laser geometry .
section vi extends the coherent transport to the angle - tuned geometry and determines the condition for a constant optical lattice depth during transport .
the application to the hyperfine - zeeman states of rubidium is presented in sec vii , and in the following section the minimum time required to realize the coherent transport in the adiabatic limit is briefly discussed .
the lasers generating the 1-d lin-@xmath1-lin configuration are composed by two phase - correlated propagating electric fields with frequency @xmath6 and amplitude @xmath7 .
their wavevectors @xmath8 lying on the @xmath9 plane create the angle - tuned geometry with angle @xmath3 , as shown in fig .
[ setup ] .
the spatio - temporal dependencies of the electric fields are @xmath10 for @xmath11 , with polarizations @xmath12 defined in the following .
this geometry creates a 1d optical lattice along the @xmath13-axis .
the laser fields confining along the transverse directions are not required for the following analysis and not listed here .
+ plane with wavevectors @xmath14 at an angle @xmath3 between them , and polarizations @xmath15 , with @xmath16 .
these wavevectors and polarizations are generated applying the rotations described in the text to two laser fields counterpropagating along the @xmath13 axis and with polarizations @xmath17 , with @xmath16 within the @xmath18 plane .
] the above laser geometry is obtained by applying proper spatial rotations to a 1d optical lattice initially created by two counterpropagating laser fields along the @xmath13 axis .
let s introduce the rotation @xmath19 of an angle @xmath20 around the @xmath21 axis @xmath22 and the rotation @xmath23 of an angle @xmath20 around the @xmath13 axis @xmath24 in the lin-@xmath1-lin counterpropagating configuration the laser wavevectors are @xmath25 and their polarizations , @xmath26 for @xmath27 , are @xmath28 notice that * , * in addition to the angle @xmath1 between the two polarization directions , we have introduced the angle @xmath29 between the polarization vector @xmath30 and the @xmath21 axis .
the wavevectors @xmath31 , given by eq . ( 1 ) , are obtained applying the following rotations : @xmath32 such rotations are applied as well to the polarization vectors @xcite @xmath33 the electric fields @xmath34 , @xmath35 are obtained by substituting eqs .
( [ vectors ] ) and ( [ polarizzazioni ] ) into eq .
( [ electricfield ] ) , and the total electric field is given by @xmath36 where @xmath37 defines the local polarization , not necessarily unit norm .
the optical potential experienced by the atoms is obtained from the analysis of ref .
for alkali atoms , in the limit that the laser detuning is much larger than the hyperfine splittings in both the p@xmath38 and p@xmath39 excited states , the optical potential has the following form @xcite : [ effpotd ] @xmath40 , \label{field}\end{aligned}\ ] ] where , in order to simplify the notation , we introduced the following quantities : @xmath41 representing the scalar and vector polarizabilities , derived for instance in @xcite .
the operators @xmath42 are the identity and pauli operators in the electron ground - state manifold .
the polarizabilities @xmath43 and @xmath44 corresponding to the excitations to the p@xmath45 and p@xmath46 excited states respectively , depend on the dipole operator reduced matrix element @xmath47 with @xmath48 * : * @xmath49 here @xmath50 is the detuning of the laser frequency @xmath6 from the resonance between the states @xmath51 and @xmath52 or @xmath53 , for the d@xmath54 or d@xmath2 lines of @xmath55rb respectively . + the substitution of eqs .
( [ polarizzazioni ] ) for the local polarization in the right sides of eqs .
( [ uj ] ) and ( [ field ] ) leads to @xmath56 \
, , \\ { \bf e}_{\rm l}^{*}\times{\bf e}_{\rm l}&=i{\bf b}(\theta_1,\theta,\phi ) \sin(2\pi\frac { z}{d } ) \label{angledependence } \end{split}\ ] ] where @xmath57 is the period of the optical lattice .
the @xmath58 parameter describes the spatial dependence of the scalar part of the optical lattice @xmath59 the spatial components of the effective magnetic field are given by @xmath60 determining also the module @xmath61 .
@xmath62 and @xmath63 satisfy the following useful relations : @xmath64 and @xmath65 the effective magnetic field @xmath66 @xcite varies spatially with a @xmath67 period .
its components along the three axes have amplitudes depending on the angles defining the lattice geometry .
if the light field is everywhere linearly polarized , @xmath68 , the effective magnetic field vanishes and the light shift is independent of the magnetic atomic sublevel * : * @xmath69 . for the counterpropagating geometry , i.e. @xmath70 ,
investigated by @xcite and implemented in @xcite , the effective magnetic field is oriented along @xmath13-axis .
making the assumption of neglecting the kinetic energy of the atoms , the effective potential corresponds to the full hamiltonian acting on the atomic states , and the position @xmath13 can be treated as an external parameter .
if we consider the two - dimensional subspace characterized by the electron spin component , _
i.e. _ , @xmath71 and @xmath72 , the eigenvalues of the hamiltonian are @xmath73 with a constant term @xmath74 left out .
these quantities represent the optical potential experienced by the atoms . in an equivalent description ,
@xmath75 define the energies of the @xmath76 atomic states when the electron spin is aligned along the local direction of the magnetic field , _ i.e. _ , @xmath77 .
in fact we write @xmath78 the eigenvalues @xmath75 of eqs ( [ eigenvalue ] ) can be expressed as @xmath79 the potential depth @xmath80 and the relative phase @xmath81 are given by [ generic ] @xmath82 where we have introduced @xmath83 eq .
( [ eigen2 ] ) demonstrates the spatial periodicity of the potentials experienced by the atomic eigenstates .
ultracold atoms are trapped at the spatial positions corresponding to the minima of the optical lattice potentials . for positive @xmath84 * , * the @xmath85 and @xmath86 minima positions for the @xmath87 and @xmath72 states respectively are given by @xmath88 with @xmath89 an integer .
for instance at @xmath90 , the @xmath91 atoms are localized at @xmath92 and the @xmath93 atoms at @xmath94 . at @xmath95
both species are localized at the @xmath96 positions .
we consider the case @xmath97 of an optical lattice created by two counterpropagating laser fields .
then the functions @xmath62 and @xmath98 determining the optical lattice potential reduce to @xmath99 for this geometry the functions @xmath62 and @xmath98 do not depend on the angle @xmath29 but only on the relative angle @xmath1 .
this is a consequence of the symmetry of the system . since the two beams forming the optical lattice propagate along the direction @xmath13 , the system is invariant under rotations around that axis .
+ at @xmath100 the potential depth @xmath101 and the phase shift @xmath102 become @xmath103 ( continuous line ) and @xmath104 ( dashed line ) , in units of @xmath74 , plotted for different values of @xmath1 ( @xmath105 , @xmath106 and @xmath107 from left to right ) in the case of the counterpropagating geometry corresponding to @xmath97 and for the value @xmath108 . ] fig .
( [ fig_2 ] ) reports the two eigenvalues of eqs .
( [ eigen2 ] ) as a function of @xmath13 for different values of the relative angle @xmath1 between the two linear polarizations .
the laser frequency is chosen such that @xmath109 , that is @xmath108 . at @xmath110 with the laser polarizations parallel , @xmath111 and @xmath104 coincide . by
increasing @xmath112 , the minima of the potential curves @xmath111 and @xmath104 move in opposite directions along the @xmath13 axis , and the potential depth decreases . at @xmath113 the minima of @xmath111
coincide with the maxima of @xmath104 , and their amplitudes are at the minimum .
let us suppose to start at @xmath113 preparing a @xmath91 atom at the site @xmath114 and a @xmath93 atom at the site @xmath115 .
varying adiabatically @xmath1 from @xmath107 to @xmath105 ( or to @xmath0 ) the two particle will occupy the same site .
this protocol was used to transport the atoms from one site to the other in order to produce controlled collisions @xcite .
we recall that in this counterpropagating geometry the effective magnetic field is oriented along the @xmath13 axis and is equal to zero for @xmath110 , when the atoms collide .
[ fig_3 ] shows that the potential depth does not remain constant by varying @xmath1 .
as pointed in refs @xcite , this difficulty is avoided for a particular choice of the parameter @xmath116 .
in fact for @xmath117 , the term @xmath118 disappear in eq .
( [ counter ] ) and @xmath84 becomes a constant independent of @xmath1 and equal to @xmath74 . as a function of @xmath1 for @xmath119 ranging between @xmath105 and @xmath120 in steps of @xmath121 , from bottom to top . for @xmath122
the potential is a constant independent of @xmath1 . ] by assuming equal dipole moments for the d@xmath123 and d@xmath124 lines @xmath125@xcite , the parameter @xmath126 becomes * : * @xmath127 fig .
[ fig4 ] reports the parameter @xmath116 versus the laser wavelength @xmath128 .
the constraint @xmath129 implies @xmath130 or @xmath131 .
the first condition is satisfied for @xmath132 which is not an acceptable value , since the whole treatment for the optical lattice potential is valid only for detunings @xmath133 large with respect to the typical hyperfine splitting .
the @xmath131 relation is satisfied if the laser wavelength is equal to the _ magic _ value @xmath134 @xmath135 where @xmath136 and @xmath137 denote the resonant wavelengths for the d@xmath123 and d@xmath124 lines . for @xmath131
the relative phase @xmath138 becomes @xmath139 therefore for the magic wavelength * @xmath134 * the potential depth is independent on @xmath1 and the phase @xmath138 varies linearly with the relative angle @xmath1 . as a function of the laser wavelength @xmath128 .
the two continuous vertical lines denote the position of the resonant wavelengths @xmath137 and @xmath136 , respectively .
the dashed vertical line indicates the magic wavelength @xmath134 . ] and @xmath140 versus @xmath1 for the laser at the magic wavelength , corresponding to @xmath141 .
for @xmath138 the continuous line corresponds to @xmath142 and @xmath143 .
the dashed line plots @xmath138 at @xmath144 whichever value of @xmath3 .
the same @xmath138 values are obtained for @xmath97 whichever value of @xmath29 . for @xmath140
the dotted line corresponds to @xmath144 whichever value of @xmath3 .
the coherent transport requires for the phase shift @xmath107 as initial value and @xmath105 as final value .
a complete transport can not be realized for the conditions corresponding to the continuous line , but instead can be realized for the dashed and dotted lines . ]
for @xmath3 different from @xmath0 the potential depth @xmath84 and the phase @xmath138 depend on the angle @xmath29 as well .
this reflects the fact that for @xmath145 the system is not invariant under rotations around the @xmath13 axis .
+ for the @xmath84 dependence on @xmath29 the magic wavelength @xmath134 plays again a key role .
in fact by choosing @xmath146 eqs .
( [ generic ] ) reduce to @xmath147 therefore the potential depth @xmath84 depends only on the laser wavelength .
+ the remaining dependence of @xmath138 on the angles @xmath148 defining the optical lattice represents a difficulty for the coherent transport operation .
for instance for the values of @xmath142 and @xmath143 the phase shift @xmath138 versus @xmath1 is plotted as a dashed line in fig .
[ fig_5 ] and is compared to the value corresponding to the counterpropagating geometry , plotted as a continuous line . for those values of @xmath3 and @xmath29 the range of variation of @xmath149 is smaller than @xmath0 .
this means that by varying @xmath1 the two potential curves corresponding to the two atomic eigenstates are shifted by a quantity smaller than the spatial period @xmath67 .
thus the minima of the potentials for the two atomic eigenstates do not coincide and a complete transport is not realized .
however under the special condition of @xmath150 eq .
( [ useful2 ] ) holds , and the phase @xmath138 becomes @xmath151 therefore by choosing @xmath152 and @xmath144 , the phase shift becomes fully equivalent to that of the counterpropagating case .
thus , it is possible to move the two potential curves of eqs .
( [ eigen2 ] ) without changing the potential depth , and a spatial coherent transport with amplitude @xmath67 can be produced by varying the angle @xmath1 of the laser polarizations .
for this angle - tuned configuration , even if the effective magnetic field is null at @xmath110 where the atoms collide , it includes a component along the @xmath153 axis at other values of @xmath1 .
the vector component and the total potential of eqs .
( [ effpotd ] ) experienced by the atoms depend on the projection of the electron magnetic moment or , equivalently , of the total angular momentum @xmath154 along the local magnetic field .
different effective potentials are experienced by different zeeman levels of the ground hyperfine state , and the laser parameters required for the coherent control depend on the atomic computational basis . here
we consider two different hyperfine - zeeman states @xmath155 of the @xmath55rb atom as in the analysis of @xcite .
as described in @xcite the potential experienced by the atoms in these internal states is derived from the potential of eq .
( [ u ] ) by considering the components of the electron spin for these states .
however in the angle tuned geometry with @xmath145 , the effective magnetic field also includes components oriented along the @xmath21 and @xmath153 axes . as a consequence the potential experienced by the atomic states depend on the optical lattice loading process and the occupation amplitudes for the zeeman states .
in fact raman coherences of the type @xmath156 are created by the effective magnetic field . as pointed out in ref .
@xcite , in the tight binding regime where each lattice site can be considered as an independent potential well , the wannier states constituting an orthonormal basis within each well in general become spinors .
+ for simplicity we analyze the case of an adiabatic loading of the @xmath155 states in the lattice so that their @xmath157 component is oriented along the local magnetic field @xmath158 .
thus we impose the atomic states under consideration to be the @xmath155 eigenstates along the local magnetic field .
let s consider the following states : @xmath159 the coefficients @xmath160 defining the normalized superposition .
for the @xmath161 state of the explorations in refs .
@xcite the coefficients are @xmath162 and @xmath163 .
the energies @xmath164 of the states @xmath165 at a fixed position @xmath13 within the optical lattice are given by @xmath166 while @xmath84 and @xmath138 are given by eq .
( [ generic ] ) , the potential depth @xmath167 and the phase @xmath168 are given by [ fpotentialphase ] @xmath169 where @xmath170 we obtain two different effective lattice potentials trapping the atoms in the@xmath171 and @xmath172 hyperfine - zeeman states .
+ because eqs .
( [ fpotentialphase ] ) have the same structure as eqs .
( [ eqangletuned ] ) , the coherent transport is determined by the @xmath173 dependence at fixed values @xmath29 and @xmath3 . by using this analogy we conclude that , in order to perform a controlled - collisions experiment , the potentials seen by the two hyperfine states @xmath174 must move in opposite direction when @xmath1 is varied .
the comparison eq .
( [ gamma1 ] ) to eq .
( [ genericb ] ) indicates that this condition is satisfied when the @xmath172 state is chosen such that @xmath175 , that is , when @xmath176 . for the states @xmath177 , @xmath178 ,
@xmath179 this inequality is satisfied .
the optimal coherent transport is obtained when the potential depth is constant by varying the @xmath1 control parameter . in section v the constance of the optical depth was realized by fixing the laser wavelength at the magic value . for the present case of
two hyperfine states a unique magic wavelength where the @xmath84 and @xmath180 potential depths are both independent of @xmath1 does not exist . while the @xmath84 constance imposes @xmath117 and produces the magic wavelength eq .
( [ magic ] ) , the @xmath180 constance imposes @xmath181 leading to a different laser wavelength .
for instance , at @xmath144 and fixing the laser wavelength to the @xmath182 value of eq .
( [ magic ] ) such that @xmath141 , the potential depths for the @xmath174 states become @xmath183 while their phases @xmath184 are @xmath185 when @xmath1 is varied from @xmath186 to @xmath105 or @xmath0 , while the @xmath84 potential depth is independent of @xmath1 , the @xmath180 potential depth depends on @xmath1 and its range is determined by @xmath187 , whence by the @xmath188 values .
therefore , when @xmath1 is varied the potentials experienced by the @xmath165 states move with different velocities .
the phase @xmath138 is linearly dependent of @xmath1 , while the dependence of @xmath140 on @xmath1 has a more complicated behavior . for the case of the @xmath179 state
the @xmath140 dependence is shown by the dotted line in fig .
[ fig_5 ] .
the coherent transport condition of phase shifts @xmath184 varying from @xmath107 to @xmath105 is realized for both the hyperfine - zeeman states .
different results for the change in the potential depth and for the phase dependence on @xmath1 , and therefore for the displacements of the two potentials , are obtained for a laser wavelength different from the magic one .
in order to realize an efficient quantum gate , the time dependence of polarization angle @xmath1 should be chosen so that the transport of the atomic states is realized in the adiabatic limit , _
i.e. _ the atoms remain in the ground state of local optical potential . for an analysis of the adiabaticity constraint necessary for the coherent transport we approximate the atomic potential of eq .
( [ eigen2 ] ) with a harmonic one . for non interacting atoms experiencing a harmonic potential moving with respect to the laboratory frame
, the hamiltonian may be written as @xcite @xmath189 where @xmath13 denotes the coordinate of the atom in the harmonic potential frame , @xmath190 represents the acceleration of the harmonic potential in the laboratory frame , and the oscillation frequency @xmath191 of the harmonic oscillator is @xmath192 if we consider the last term of the above hamiltonian as a time - dependent perturbation , the probability of transferring an atom to the first excited level of the harmonic potential is @xmath193 where @xmath194 is the ground state radius for the harmonic potential , and @xmath195 is the time required for the coherent transport process . + the amplitude of the transfer probability of eq .
( [ probability ] ) depends greatly on the time dependence of the @xmath190 acceleration . at first
we will suppose that , as in the theoretical analysis of ref .
@xcite and in the experimental investigation of ref.@xcite , the potential moves at a constant speed , by imposing an infinite acceleration at @xmath196 and @xmath197 . for this transformation
the adiabatic condition for the atomic transformation requires @xmath198 owing to this dependence on @xmath3 , for a constant speed of the potential the time @xmath195 required to realize an adiabatic coherent transport in the angle - tuned configuration is much longer than in the counterpropagating case , for a given depth of the optical potential .
such result could impose a strong constraint for performing quantum computation with angle - tuned lattices .
+ however , the condition on @xmath195 for realizing the adiabatic limit becomes less restrictive if we assume a different motion for the lattice harmonic potential confining the atoms .
for instance , let us assume that the lattice is constantly accelerated from @xmath196 to @xmath199 and constantly decelerated from @xmath199 to @xmath197 .
thus the adiabatic condition becomes @xmath200 therefore using this motion of the potential , the dependence of the minimum time for the coherent transport on the angle @xmath3 is modified causing a decrease in the time by a factor which for @xmath201 is larger than 2 . moreover the additional dependence on @xmath128 , @xmath84 , and the physical constants of eq .
( [ adiabaticlimit2 ] ) contributes to the decrease of the time scale , so that for the experimental conditions of ref .
@xcite transport times around 10 @xmath202s can be achieved .
in quantum computation experiments with neutral atoms loaded in optical lattices , a crucial aspect is the single site addressability . in the angle - tuned configuration where the lattice constant
could be large , the question of the single site addressability is shifted to a frame of more accessible dimensions . for this geometry
the coherent transport protocol requires specific conditions of the laser beam polarizations , linked to the breaking of the rotational symmetry associated to the counterpropagating geometry .
an additional request is the constance of the optical potential depth during the coherent transport .
this constance is realized by choosing a _ magic _
wavelength for the laser fields producing the lattice .
the value of the magic wavelength is independent of the lattice geometry .
however an unique magic wavelength for the transport of all hyperfine - zeeman atomic states does not exist . coherent transport within an optical lattice represents a component of the process based on ultracold collisions and leading to entanglement of neutral atoms and implementation of quantum logic . by storing the ultracold atoms in the microscopic potentials provided by optical lattices the collisional interactions can be controlled via laser parameters . at the low temperature associated to the mott insulator
, the collisional process is described through s - wave scattering . in the @xmath144 laser configuration of the angle - tuned geometry , at @xmath110
the effective magnetic field is null and the scattering potentials associated to the different atomic states have an identical spatial dependence .
however , as new feature brought by the angle - tuned geometry , at @xmath203 the effective magnetic field is different from zero and oriented along different directions for different hyperfine - zeeman states . therefore during the whole collisional process the colliding atoms may be oriented along different spatial directions . in order to treat this collisional configuration , the atomic interaction may be described through the pseudopotential models introduced in refs .
@xcite for asymmetric trap geometries .
the atomic control is based on a the realization of a mott insulator phase , in which the number of atoms occupying each lattice site is fixed .
the physics of such a system is described in terms of a bose - hubbard model whose hamiltonian contains the on - site repulsion resulting from the collisional interactions between the atoms , and the hopping matrix elements that take into account the tunneling rate of the atoms between neighboring sites .
both the repulsive interaction and the hopping energy can be tuned by adjusting the lasers setup , as reviewed in @xcite .
the mott insulator phase is realized under precise conditions between the on - site repulsion and the hopping matrix elements .
the dependence of these parameters defining the angle - tuned lattice should be investigate in order to realize a mott insulator in an angle tuned geometry .
this researach was financially supported by the eu through the strep project olaqui and by the italian miur through a prin project .
the authors are gratefully to dieter jaksch and carl j. williams for useful discussions .
|
optical lattices with a large spacing between the minima of the optical potential can be created using the angle - tuned geometry where the 1-d periodic potential is generated by two propagating laser beams intersecting at an angle different from @xmath0 .
the present work analyzes the coherent transport for the case of this geometry .
we show that the potential depth can be kept constant during the transport by choosing a _ magic _
value for the laser wavelength .
this value agrees with that of the counterpropagating laser case , and the magic wavelength does not depend of the optical lattice geometry .
moreover , we find that this scheme can be used to implement controlled collision experiments under special geometric conditions
. finally we study the transport of hyperfine - zeeman states of rubidium 87 .
|
a prerequisite for a theoretical description of the properties of @xmath1-mesons in hadronic matter and the possible consequences for @xmath1-production in heavy - ion collisions is a reliable model for elementary processes like @xmath1-production in hadronic collisions and in photo - induced reactions .
new accurate data on the photoproduction of @xmath1-mesons off protons @xcite as well as off deuterons @xcite provide strong constraints on models for the elementary @xmath1-meson hadron and photon hadron interactions .
we present such a model which satisfies the requirements of unitarity and gauge invariance .
an effective field theory is constructed , where the hadronic coupling constants and resonance masses are determined by fitting data on elastic @xmath0n scattering .
the consistency of our model is checked by comparing the cross section for the inelastic channel @xmath0n@xmath3n with experiment ( see ref .
having fixed the hadronic parameters of the theory , we now study electromagnetic processes , namely photoproduction of pions and etas off nucleons as well as off deuterons .
in the @xmath4-matrix approach the @xmath5-matrix for the process @xmath6p@xmath7p is , to lowest order in the electromagnetic interaction , given by @xmath8 where the sum is over all open hadronic channels .
the final - state interaction , included in @xmath9 , is entirely due to strong interactions . under the assumption that the total cross section for photoproduction of @xmath1-mesons is dominated by the s@xmath10 channel @xcite
, we use our model @xcite for @xmath0n scattering and pion - induced @xmath1-meson production in this channel to describe the final state interaction .
a crucial point is the inclusion of both resonances in the s@xmath10 channel , the s@xmath10(1535 ) and the s@xmath10(1650 ) .
since both resonances couple to @xmath0n , @xmath1n and @xmath11n we include three channels , where the two - pion - continuum is parametrized by an effective scalar field @xmath12 ( for details see ref .
@xcite ) . in order to satisfy unitarity
we have to include all three channels @xmath13 in eq.(1 ) .
we identify the resonance contributions to the @xmath4-matrix with the first diagram shown in fig . 1 .
the lagrangian , which describes the coupling of the photon to the resonance , is given by @xmath14 where @xmath15 and @xmath16 are the masses of resonance and nucleon , respectively .
since the interaction has an isovector and an isoscalar part , there are two new coupling constants per resonance . in order to be consistent with the hadronic part of the model , which includes interaction terms of the linear sigma model , and to fulfill the low - energy theorems of pion photoproduction
we also include non - resonant born terms in @xmath17 .
the corresponding interaction terms are obtained by coupling the photon to the lagrangian of the linear sigma model in a minimal way , which yields a coupling of the photon to the electromagnetic currents of nucleons and pions , respectively .
furthermore we add a coupling to the anomalous magnetic moment of the nucleon and vector - meson exchange contributions . in order to satisfy the low - energy theorems within the framework of the linear sigma model in the absence of loop
diagrams an additional contact interaction of the form @xcite @xmath18 is added , which contains an anticommutator of nucleon isospin matrices .
here @xmath19 and @xmath20 are the isoscalar and isovector part of the anomalous magnetic moment of the nucleon .
the diagrams shown in fig . 1 are included in @xmath17 .
we stress that the contact term in fig . 1 is not the kroll - ruderman term obtained by coupling the photon in a minimal way to a pseudovector @xmath0nn - interaction , but corresponds to the interaction ( 3 ) .
consequently it is proportional to the anomalous magnetic moment of the nucleon . as will be discussed below the vector - meson exchange processes shown by the last two diagrams of fig
. 1 also play an important role in @xmath21 .
we introduce form factors for the contact term and the vector - meson exchange contributions .
gauge invariance implies large cancellations between the form factor at the @xmath6nn - vertex , the one at the @xmath0nn - vertex and nucleon self - energy insertions @xcite .
the net effect is that , for the current coupling in the born terms , the form factors in diagrams 2 , 3 and 5 of fig .
1 are cancelled by other contributions .
( 150,50 ) ( 5,5 ) next we have to determine the electromagnetic parameters , namely the isoscalar and isovector couplings of the photon to the two resonances and the cutoffs . the ideal way to proceed
would again be to fix the parameters by fitting the e@xmath2-amplitude of pion photoproduction and then make a prediction for the total cross section for @xmath6p@xmath7p .
however , in view of the fact that there are substantial deviations between different existing partial wave analyses for pion photoproduction this is not a practicable way .
consequently we choose a more pragmatic approach and try to find a consistent description of the two processes .
reasonable agreement with the data on the e@xmath2-amplitude of pion photoproduction in the different isospin channels ( not shown ) is obtained for a parameter set which also reproduces the total cross section for @xmath1-production off proton and deuteron .
we find the following helicity amplitudes for the two resonances ( in units of @xmath22 gev@xmath23 ) : @xmath24=102 , @xmath25=82 for the s@xmath10(1535 ) and @xmath24=83 and @xmath25=24 for the s@xmath10(1650 ) .
the result for the total cross section for the process @xmath6+p@xmath7+p is shown by the solid line in fig . 2 ( left ) .
the dot - dashed line shows the result if one completely neglects the presence of the second resonance and the vector - meson exchange contributions to @xmath21 and scales the result with a factor 0.7 while the long - dashed line is obtained if one includes the second resonance but not the vector - meson exchange terms .
obviously the presence of the second resonance leads to destructive interference and a shift of the peak towards higher energies compared to a model which only includes one resonance .
this shift is compensated by the vector - meson exchange contributions .
we stress that the position of the peak in the photoproduction cross section is not sensitive to the remaining free coupling constants , @xmath26 and @xmath27 .
thus , when the second resonance is included we can not describe the data without vector - meson exchange contributions .
this is an important difference to other models , where the second resonance was ignored @xcite .
we note that the coupling of the s@xmath10(1650 ) to the @xmath1-meson is very weak and consequently neglected in our model . in spite of this
, the second resonance plays an important role in the photoproduction of @xmath1-mesons . in a coupled channel approach
this is possible , in contrast to tree level calculations @xcite , since the s@xmath10(1650 ) couples to the @xmath1n - channel over intermediate @xmath0n , @xmath12n and s@xmath10(1535 ) states .
( 150,70 ) ( -3,2 ) ( 150,70 ) ( -3,2 ) also shown in fig .
2 is our result for photoproduction of @xmath1-mesons off the deuteron , computed in the impulse approximation . except for a deviation near threshold
the agreement with the data is very good .
the photoproduction of @xmath1-mesons off the nucleon is described within a unitary and gauge invariant model .
elastic @xmath0n scattering data are used to determine the hadronic parameters of the model .
the predicted cross section for pion - induced @xmath1-production agrees well with the data .
we obtain a consistent description of photoproduction of @xmath1-mesons and pions satisfying the low - energy theorems for the latter reaction .
we find that the s@xmath10(1650)-resonance plays an important role in all the processes we consider and that the inclusion of @xmath28- and @xmath29-exchange is crucial in the photoproduction of @xmath1-mesons .
estimates of the @xmath1-nucleon pole terms in the framework of a nonlinear su(3)@xmath30su(3 ) sigma model with mesons and baryons indicate that their contributions to photoproduction of @xmath1-mesons is small and can be neglected @xcite .
thus , our model provides a consistent description of total cross sections of both pion- and photoinduced @xmath1-meson production . in order to describe differential cross sections
one has to extend the model by including the p@xmath10 as well as the d@xmath31 channel @xcite .
b. krusche et al .
* 74 * ( 1995 ) 3736 m.wilhelm , phd thesis , bonn , bn - ir-93 - 43 b. krusche et al . ,
submitted to phys .
b c. sauermann , b.l .
friman , w. nrenberg , phys . lett . *
b341 * ( 1995 ) 261 s. fubini , g. furlan , c. rossetti , nuov .
* 40 * ( 1965 ) 1171 f. gross and d.o .
riska , phys . rev . *
c36 * ( 1987 ) 1928 h.w.l .
naus and j.h .
koch , phys .
rev . * c39 * ( 1989 ) 1907 l. tiator , c. bennhold , s.s .
kamalov , nucl .
* a580 * ( 1994 ) 455 m. benmerrouche , n.c .
mukhopadhyay , j.f .
zang , phys .
rev . * d51 * ( 1995 ) 3237 n. kaiser , p.b .
siegel , w. weise , tu munich preprint tum / t39 - 95 - 10 , and private communication
|
an effective field theory with hadrons and photons is constructed , where the hadronic parameters are determined by fitting elastic @xmath0n scattering data .
the predicted pion - induced @xmath1-production on the nucleon agrees well with the data .
the hadronic model is then used to describe the final - state interaction in photoinduced processes .
we present a consistent description of pion photoproduction in the e@xmath2-channel and of the total cross sections for photoinduced @xmath1-production on the proton as well as on the deuteron .
|
here we briefly summarize the most important results of our paper so that the reader interested in applying our formalism to specific models will easily find the relevant material .
- we derive the tn representation of the standard ks lgt hamiltonian for arbitrary compact groups in eq .
[ eq : one_site_op ] , [ eq : u_ks ] , [ eq : h_ks ] .
- we present an extra gauge invariant operator that can be added to the ks hamiltonian to explore generalized ks lgt for arbitrary compact groups in eq .
[ eq : v_op ] .
- we provide the exact tn representation for the projector onto @xmath4 of the ks lgt ( represented as an hyperplane in fig .
[ fig : hilb_emb ] ) in fig .
[ fig : proj_gauge ] .
it is the contraction of several copies of elementary tensors @xmath13 and @xmath14 defined in fig . [
fig : proj_k_s ] and eq .
[ eq : dis ] .
- we describe a truncation of the hilbert space of the ks lgt that is consistent with gauge symmetry in sect .
[ sec : t_ks ] .
this allows to study lgt with continuous groups ( i.e. @xmath12 and @xmath15 ) with constituents leaving in finite dimensional hilbert spaces . - in this way we introduce an exact tn representation with finite bond dimension of the projector onto the @xmath4 of an arbitrary lgt .
it is encoded in the tn of fig .
[ fig : proj_gauge ] with the @xmath13 and @xmath14 tensors defined in fig .
[ fig : proj_k_s_t ] .
- we discuss the relation of this truncation scheme with gauge magnets or quantum link models .
we build the exact tn representation of the projector onto @xmath4 also for gauge magnets .
it is the tn of fig .
[ fig : proj_gauge ] with the @xmath13 and @xmath14 tensors defined in fig .
[ fig : proj_gm ] . - we provide a variational ansatz for generic states of @xmath4 of all the lgt discussed as the tn of fig .
[ fig : gen_ans ] .
the ansatz is the contraction of several copies of sparse tensors that unveils a connection between the lgt gauge invariant hilbert space and spin networks .
- we confirm that the topological entropy detects phase transitions elusive to standard local order parameters .
in particular , we characterize the transition between two different topological phases the 2d 8-vertex and the 6-vertex topological phases .
these results are presented in fig .
[ fig : topo_ent ] . in fig .
[ fig : num_sp ] we show that the same phase transition does not affect the behavior of the lowest part of the entanglement spectrum .
tensors are multi - linear maps @xmath16 acting among different hilbert spaces . in particular the coefficients of a state of a quantum many - body systems
are encoded in the element of a very large tensor @xmath17 , @xmath18 in general , the tensor @xmath19 is too large to be stored on a computer and thus it is useful to express it as the contraction of smaller elementary tensors .
these contracted tensors are called tn .
when dealing with large tn , the formulas become easily large and complex and it is simpler to resort to a graphical notation .
the graphical notation is explained already in the literature @xcite , but we also shortly review it here in order to fix the notation we use in the paper . in the graphical notation
, geometric shapes are associated to tensors and lines or `` legs '' attached to them represent their indexes . as an example , the upper panel of fig .
[ fig : tens ] represents a tensor with three indexes @xmath20 .
a small dot on the shape allows to keep track of the index ordering that is assumed clockwise starting from the dot .
incoming legs _ in - legs _ ( upper indexes ) are drawn with entering arrows while outgoing legs _ out - legs _ are drawn as outgoing arrows .
the hermitian conjugate of a tensor , @xmath21 , involves taking the complex conjugate of the elements and exchanging the in- , with the out - legs , @xmath22 .
this is represented graphically by mirroring the tensor and inverting the arrows on the lines ( left of the upper panel of fig .
[ fig : tens ] ) .
a leg connecting two tensors represents their multiplication through the contraction of the corresponding indexes ( summation over all the values of that indexes ) .
an example is represented on the left of the lower panel of fig .
[ fig : tens ] , where the tensor @xmath23 is contracted with the tensor @xmath24 . in our formulas
we sometime omit explicit summation and use the einstein notation , where the summation is intended over repeated indexes . .
upper indexes are incoming legs while lower indexes are outgoing legs .
complex conjugation is denoted by mirror reflection of the vertex .
different colors denote different tensors .
indexes are ordered clockwise starting from the solid dot on the vertex .
the contraction of two tensors is denoted by an arrow joining them . the dagger operation @xmath21 involves both the complex conjugation of the elements of the tensor , and the inversion of the arrows attached to its legs . as a consequence
the order of the legs also changes , @xmath25 @xcite .
panel b ) an isometric tensor ( or simply isometry ) is always represented by a triangular vertex .
a tensor @xmath24 is isometric if there is a specific choice of legs , which identifies two vector spaces @xmath26 and @xmath27 , such that , when @xmath24 is contracted with @xmath23 through the legs in @xmath26 , the result is the identity tensor in @xmath27 . in the figure , @xmath26 is spanned by the leg @xmath28 , while @xmath27 is spanned by the two legs @xmath29 .
[ fig : tens ] ] a tensor can always be interpreted as a matrix by dividing its legs into two groups , one group identifies a vectors spaces @xmath26 and the other @xmath27 . in this way
the tensor becomes a map from @xmath26 to @xmath27 .
a natural choice is to interpret @xmath26 as the collection of in - legs and @xmath27 as the collection of out legs , but this is not the only possibility .
an important class of tensors are isometric tensors . for a specific choice of @xmath26 and @xmath27 ,
they fulfill @xmath30 in our drawings triangular shapes always represent isometric tensors . to @xmath27 . here
we use an arbitrary orientation for the triangle and explicitly write the isometric property so to identify @xmath26 and @xmath27 ] the lower panel of fig .
[ fig : tens ] illustrates a specific case of a three - leg _ isometric _ tensors @xmath24 , where @xmath26 is spanned by the leg @xmath28 and @xmath27 is spanned by the two legs @xmath29 .
this implies that @xmath23 is defined as @xmath32 and equation reads @xmath33 ( remember that there is a sum over @xmath28 ) .
graphically the contraction is represented by the line connecting the two tensors .
the result of the contraction is a four - leg tensor , explicitly written as the tensor product of two identity tensors represented by straight lines . when tns represent quantum states of many - body systems , the legs related to the constituents ( @xmath34 in eq .
) are called _ physical _ legs and are typically represented by latin letters , while all the others legs ( those that are contracted ) are legs called _ auxiliary _ legs , and represented by greek letters .
here we assume that the reader is familiar with basic concepts of the representation theory of both finite and continuous groups , and we list the relevant results for our paper in order to express them in the tn language . in particular , while this paper deals with gauge theories with continuous groups , we develop the formalism by using discrete groups @xmath3 .
the underlining ideas , are indeed completely independent from the fact whether the group is discrete or continuous , and we feel that discrete groups allow to present these ideas in a simpler way . concretely , all the results we present , which involve the summation over group elements , can be rewritten for continuous compact groups by substituting the sums with integrals over the group , defined through the appropriate invariant measure ( see i.e. chapter 4 of @xcite ) . here , we remind that a collection of elements @xmath35 closed under multiplication forms a group @xmath3 .
we are mostly interested in matrix representations of the group @xmath3 that are obtained by associating to each group element @xmath36 a unitary matrix , @xmath37 , acting on a given vector space . in this way
the group multiplication table is rephrased into specific relations between the representing matrices . in general , given a matrix representation of a group , there is a well defined procedure to reduce it to a block diagonal form , where each of the blocks constitute a `` smaller '' , independent representation of the same group @xmath3 .
if those blocks are not further reducible into smaller blocks they define an irreducible representation of the group @xmath3 that in this paper will be labeled by @xmath38 . the dimension and number of the irreducible representations depend on the group @xmath3 , and their study is the subject of the theory of group representations . in the following
we denote by @xmath39 the matrix representation of @xmath36 in the irreducible representation @xmath38 .
one of the most important results of the theory of group representations is , what is typically called , the `` great orthogonality theorem '' @xcite .
it states that given a group @xmath3 with elements @xmath36 , and given any pair of irreducible representations @xmath40 the following relation holds @xmath41 where @xmath42 are the @xmath43 matrix elements of the irreducible representation ( irrep ) @xmath38 of @xmath36 , @xmath44 is the number of elements of @xmath3 , and @xmath45 and @xmath46 are the dimensions of the irrep s @xmath38 and @xmath47 .
we now want to reinterpret this relation in terms of tn diagrams . in order to do this
we need to identify two vector spaces @xmath26 and @xmath27 .
@xmath26 is spanned by the group elements @xmath36 .
this vector space is called group algebra @xmath48 , and has dimension equal to @xmath44 .
@xmath27 is the vector space spanned by the direct sum @xmath49 .
each of @xmath50 is the defining space of the irrep @xmath38 .
@xmath51 is the defining space of its conjugate representation , obtained by taking the hermitian conjugation of the matrices , @xmath52 , where we have used the property that we are dealing with unitary representations .
the fact that the vector space @xmath27 has a direct sum structure is encoded in the @xmath53 factor in the right hand side of eq . .
we start by focusing on the above formula in the case @xmath54 . in this case
tells that the tensor @xmath55 , @xmath56 is an isometry .
the tensor @xmath55 is represented in the panel a ) of fig .
[ fig : ortho ] .
the fact that it is an isometry means that @xmath57 as represented in the panel b ) of fig .
[ fig : ortho ] .
each of the @xmath55 thus allows to project a vector @xmath58 in the group algebra @xmath59 onto a vector @xmath60 of the tensor product @xmath61 .
this can be used in two ways .
reading eq . from left to right , it tells us that if we know all the @xmath39 for all the elements @xmath62 , we can collect them in a three - leg tensor and obtain an isometry that projects @xmath63 onto @xmath61 . from right to left , we can obtain the irrep s matrices . if we are given the isometry @xmath55 , by acting onto the vector @xmath64 of @xmath48 we obtain @xmath39 , @xmath65 as we represent graphically in the panel c ) of fig .
[ fig : ortho ] . projecting @xmath48 onto @xmath66 .
if one knows the matrices @xmath39 for all @xmath62 , @xmath55 is defined by collecting them inside a three - index tensor .
the index @xmath38 identifying the irrep is written on the top of one of the two legs .
b ) the isometric property of the tensor @xmath55 .
c ) alternatively , if we know the @xmath55 isometry , we can obtain the matrices in the @xmath38 representation by acting with @xmath55 on a vector proportional to @xmath58 ( yellow circle in the figure ) .
the resulting tensor has two indexes @xmath67 , and is the matrix @xmath68 .
[ fig : ortho ] ] a peculiarity of our notation , that we inherit from spin networks , is that each leg of the tensor also carries a representation index @xmath38 that is typically superimposed to the line . in order to avoid confusion , after fig .
[ fig : w_g ] we will drop all the letters labeling the indexes of the tensors ( unless really needed ) , and keep only the letters related to the irrep @xmath38 . a second result of the theory of group representations is that @xmath69 , that is the direct sum of all the @xmath55 is a unitary transformation as encoded in following relations , @xmath70 a graphical representation of the direct sum of @xmath55 leading to @xmath71 is presented in fig . [
fig : w_g ] . is a unitary tensor .
it allows to change basis from @xmath72 to the direct sum of all the irrep s times their conjugate , @xmath73 . on the upper part of the figure
we explicitly draw the collection of all the @xmath55 s .
a simplified picture is shown on the lower part , where the direct sum is implicit in the absence of the label @xmath38 attached to the legs of the tensor .
this is the notation we will use in the following .
when we want to represent @xmath55 we will attach @xmath38 to the leg , while we represent @xmath71 by exactly the same drawing but without the @xmath38 . from now on , for simplicity , we will also omit to label the legs with letters .
b ) the unitarity of @xmath71 is encoded in the fact that the contraction over the two legs acting on @xmath74 gives a delta function in @xmath72 .
[ fig : w_g ] ] in the context of many - body quantum systems , symmetries play a fundamental role in the classification of phases .
recently , in the context of tn states , symmetries have been used to classify gapped phases of 1d systems @xcite . for this reason
a strong effort has been devoted to incorporate the appropriate exact symmetries even when studying many - body systems with approximate variational ansatz . in the context of tn , a sufficient condition in order to have symmetric states is that the constituent tensors are symmetric @xcite . as an example a symmetric tensor with respect to the group @xmath3 with one in - leg and two out - legs obeys the following equation , @xmath75 with @xmath39 the matrix of the appropriate unitary representation of the group @xmath3 .
this relation is sketched graphically in fig .
[ fig : symm ] .
equation can be satisfied by non - vanishing tensors only when the tensor product @xmath76 contains the trivial representation , that is the representation where all group elements are mapped to the identity .
it is thus important to be able to explicitly construct the trivial representation contained in a given tensor product of different representations . for continuous groups
this can be done by diagonalizing the corresponding casimir operators . their zero eigenvalues , if present , identify the trivial factors .
the way this is done in practice is explained in detail in refs .
@xcite . an alternative way , which works both for discrete and continuous groups , is to explicitly build the projector onto the trivial representation as a group sum ( or integral in the case of continuous groups ) .
the projector is given by @xmath77 as can be found i.e. in ref @xcite .
this is the method that we will mostly use .
a third possibility entails disentangling the symmetry constraints following the ideas of refs .
we postpone the discussion about this until sec . [
sec : proj_tn ] , where we will provide an explicit example of this procedure . , and its out - going legs by @xmath37 in the appropriate representation .
here we exemplify the case of a three - leg tensor with an incoming leg that transforms in the irrep @xmath38 , and two out - legs that transform under irrep s @xmath47 and @xmath78 .
[ fig : symm ] ]
gauge theories originated , at a classical level , in the description of the electromagnetic interactions between charged particles and light .
the physical processes , described by the maxwell equations , depend only on the electric and the magnetic fields while maxwell equations can be written in terms of a vector potential and in this form they show some redundancy . the vector potential can be modified by adding to it the gradient of an arbitrary scalar potential , without affecting the corresponding electric and magnetic field , so giving the same physical results . at a quantum level
the vector potential becomes a full quantum field , and the redundancy appears as a local symmetry in the action that drives its dynamics .
the generalizations of these ideas to vector potentials describing non - abelian `` electric and magnetic '' fields , and their success in describing the hadron spectrum , gave rise to the modern gauge theories , and to our understanding of particle physics .
actual calculations away from the perturbative regime are most of the times carried out numerically in the framework of lgt by discretizing the space - time on a lattice . in this formalism
the vector potential is associated to the links of the lattice , while the charged matter fields live on the sites .
a hamiltonian version of the system has been obtained by identifying one of the lattice direction as `` time '' , fixing the temporal gauge , and constructing the hamiltonian operator whose matrix elements coincide , in the time continuum limit , with those of transfer matrix in the time direction @xcite . in the hamiltonian formulation lgt
become many - body quantum systems , whose constituents are divided in two groups , _ gauge bosons _ attached to the links of an oriented lattice @xmath79 and _ matter constituents _ attached to the sites @xmath80 of @xmath79 . here , we will work on an oriented 2d square lattice but , what follows , can be generalized easily to more complex orientable lattices . the original local symmetry of the classical action is then mapped to a residual local symmetry of the quantum hamiltonian
. symmetric hamiltonians typically have symmetric eigenstates , and thus one can decide to characterize the space of locally symmetric states .
the residual local symmetry is defined in terms of a set of constraints , that the quantum sates and operators should fulfill .
as we show in this section , both the operators used to define the symmetry constraints and the local constraints have a natural expression in terms of tn diagrams .
in particular in this section we describe the kogut - susskind version of the hamiltonian lgt ( ks ) .
the hilbert space for gauge bosons is the group algebra @xmath59 . in this case
, one can associate a state @xmath58 to any group element @xmath81 .
states representing different group elements are orthogonal @xmath82 . as a consequence ,
the dimension of the local hilbert space is equal to @xmath44 . in particular , continuous groups require dealing with infinite dimensional hilbert spaces .
the lattice @xmath79 is oriented , and it is made by @xmath83 links so that the total hilbert space is @xmath84 .
changing the orientation of one link sends @xmath85 .
if we act with the operator @xmath86 on the state @xmath58 , this also implies that we need to act with the operator @xmath87 on the state @xmath88 , obtained by reversing the link orientation .
a prerequisite for defining the action of the symmetry operators is to be able to define the action of left and right rotations of a state by arbitrary group elements @xmath89 and @xmath90 .
this is achieved by defining the operators @xmath91 and @xmath92 , which act on @xmath58 and produce @xmath93 this is done first by using @xmath71 of eq . , to change basis from @xmath59 to @xmath94 . at this stage
the rotation is performed through the direct sums of the rotation matrices in each representation , and then one rotates back to @xmath59 with @xmath95 .
this is expressed graphically in the panel b ) of fig .
[ fig : hilb_space ] , and reads @xmath96 , w^{\dagger}_g \label{eq : left_ks}\ ] ] @xmath97 w^{\dagger}_g .
\label{eq : right_ks}\ ] ] of the group algebra @xmath59 .
b ) left and right rotations @xmath91 and @xmath92 are introduced in eq . and
they transform the state @xmath58 into @xmath98 .
both operators require an initial change of basis from @xmath59 to @xmath99 , obtained through @xmath71 of eq . and represented in fig .
[ fig : w_g ] ( the first horizontal triangle ) .
one then applies to each of the legs the corresponding rotation matrix given by @xmath100 $ ] , with the individual @xmath101 defined in panel c ) of fig .
[ fig : ortho ] and represented here by the two vertical triangles . at last @xmath95 , represented by the inverted triangle , allows to go back to to @xmath59 .
[ fig : hilb_space ] ] having the operators that perform the left and right rotations , we are now in the position to define the operators @xmath102 , the building blocks of local gauge transformations .
in particular , the local transformation rotates all the states of the links touching a site @xmath80 by an element @xmath103 .
since the lattice is oriented , the transformation induced by @xmath102 is different for entering @xmath80 ( in - links ) and links leaving @xmath80 ( out - links ) .
all in - links are rotated through @xmath104 while all the out - links are rotated on the left by @xmath91 .
concretely , @xmath102 at site @xmath80 , acting on links @xmath105 to @xmath106 ordered counter - clockwise starting from the left , is defined as @xmath107 with @xmath83 and @xmath108 defined respectively in eq . and . notice that @xmath109= 0 $ ] if @xmath110 , as a consequence of the commutation between @xmath83 and @xmath108 operators defined on the same links .
they are represented graphically in fig .
[ fig : a_ks ] .
a generic gauge transformation is then a product of local gauge transformations , where for each site @xmath80 one chooses a different element @xmath111 to perform the desired rotation .
given a lattice of @xmath112 sites and a choice of @xmath112 elements @xmath113 , we obtain the transformation @xmath114 , @xmath115 , the building block of gauge transformations in the kogut - susskind lgt . on every link
entering a site @xmath80 , it either rotates the state of the link through @xmath104 or @xmath91 , depending on whether the link enters or leaves the site .
the form of @xmath83 and @xmath108 are given in panel b ) of fig .
[ fig : hilb_space ] .
[ fig : a_ks ] ] the set of states in the hilbert space that are invariant under all @xmath116 , the building blocks of the gauge transformations , defined by eq . , constitutes the _ physical hilbert space _ ( or gauge invariant hilbert space ) @xmath117 , @xmath118 where @xmath80 are the sites of the lattice @xmath79 , @xmath83 is the number of links , and @xmath36 is an arbitrary group element .
an example of state in @xmath117 is given by @xmath119 since @xmath120 .
@xmath117 is a subspace of the original tensor product hilbert space @xmath84 , as sketched in fig .
[ fig : hilb_emb ] . by construction
all states in @xmath117 are also invariant under any global transformation @xmath114 , defined in eq . .
we are now interested in introducing the dynamics of a lgt , and this is first done by characterizing gauge invariant operators .
gauge invariant operators are operators @xmath86 that commute with all the @xmath116 , that is @xmath121 = 0 , \forall g \in g , s \in \lambda .
\label{eq : g_inv_op}\ ] ] we would like to find the `` minimal '' gauge invariant operators , that is those that constitute the building blocks for gauge invariant hamiltonians , and with as small support as possible . in order to find them we take inspiration from the qed hamiltonian , whose gauge part can be written as @xmath122 , with @xmath123 and @xmath124 being the electric and magnetic fields .
on the lattice the @xmath125 term is mapped to a link operator ( remember that @xmath126 , with @xmath27 being a scalar field , so that in a mathematical sense @xmath123 is naturally a one form , and thus geometrically attached to links ) . the term @xmath127 is mapped to a plaquette operator , where plaquettes , the smallest possible closed loops made by links , are the elementary pieces of the lattice surface ( again remember that @xmath128 , with @xmath129 being a vector field , and thus @xmath124 is naturally identified with a two form , discretized on plaquettes ) . for this reason we look for a generalization to an arbitrary group @xmath3 of the @xmath123 operator of electromagnetism , as a single - link operator , and of the @xmath124 operator , as an operator acting on plaquettes of the lattice .
we start by the single link operator . in an abelian lgt
any matrix representing the rotation by a group element commutes with all the others , and thus fulfills eq . .
for this reason we can choose any @xmath130 as a link operator @xcite .
when dealing with non - abelian lgt , the only link operator that commutes with an arbitrary rotation , as requested by eq . , is an operator proportional to the identity in each of the irrep s , as a consequence of the schur lemma @xcite .
a gauge invariant link operator acting on a link @xmath131 in the ks lgt thus can be expressed as @xmath132w_g ^{\dagger}\right]_{s_n } , \label{eq : one_site_op}\ ] ] with @xmath133 arbitrary numbers and @xmath71 defined in eq . .
plaquette operators can be defined as matrix product operators whose elementary tensors @xmath26 act on the hilbert space of a single link and an auxiliary hilbert space .
specifically @xmath26 acts on the tensor product of the @xmath134 .
@xmath135 is the defining space of the irrep @xmath136 of @xmath3 that defines how the matter transforms under gauge transformation @xcite .
@xmath26 is defined as @xmath137 it acts diagonally on the hilbert space of the link @xmath131 , and perform a controlled rotation in the auxiliary space , in other words it `` reads '' the state of the link and rotates accordingly the state on the auxiliary space @xmath136 .
the operator @xmath138 is represented graphically in fig . [ fig : u_ks ] .
operator of in the ks lgt .
it acts on the tensor product of the physical space @xmath72 of the link @xmath131 and of an auxiliary space that defines the irrep @xmath136 , @xmath139 .
it is built from the contraction of a copy tensor @xmath140 ( red circle ) , and the corresponding rotation matrices @xmath141 ( defined in eq . ) .
b ) the above definition implies that the @xmath138 transmits the rotation onto the physical index , as induced by conjugation by @xmath83 and @xmath108 , to the auxiliary indexes .
this property allows to use @xmath138 as building block of the gauge invariant plaquette operator in eq . , where , thanks to the trace , the rotations attached to the auxiliary indexes cancel .
[ fig : u_ks ] ] from the definition we can derive its covariance properties under left and right rotations . indeed after the gauge transformation that sends @xmath58 to @xmath142 , a @xmath138 becomes @xmath143\ u\ \left [ l(h^{-1 } ) r(k)\right ] ^{\dagger } = \nonumber \\
= \sum_g { \ket { h^{-1 } g k}\bra { h^{-1 } g k}}_{s_n } \otimes \gamma_{r_m}(g)^{i}_{j } , \label{eq : phys_rot}\end{aligned}\ ] ] as illustrated in the lower panel of fig .
[ fig : u_ks ] . by re - defining @xmath144
we obtain @xmath145 eq . together with show that the @xmath138 allows to transfer the rotation on the physical hilbert space of the link to corresponding rotations on the auxiliary space .
this relation is often considered as the defining relation of a lgt @xcite , since @xmath26 can be thought of the equivalent of the position operator in the group manifold , while @xmath146 are equivalent to translation operators on the group manifold . by appropriately building a close path out of @xmath26 we can get rid of the rotations on the auxiliary legs , and
thus obtain a gauge invariant operator .
in particular we can now construct the simplest closed path that leads to the plaquette operator we are after .
@xmath147 where we denote by @xmath148 a plaquette of @xmath79 , and by @xmath149 the links around it , ordered counter - clockwise .
the dagger is related to the fact that the orientation of the plaquette is in some cases opposite to the natural orientation of the lattice , and the trace is intended over the auxiliary indexes only .
the equivalence between eq . and guarantees that the plaquette operator , represented in fig .
[ fig : plaquette_op_ks ] , commutes with the gauge transformations that have non - trivial overlap with it . of eq .
constructed as a matrix product operator from four @xmath150 s , acting on the links around a plaquette in the ks lgt . as a consequence of the covariance of the @xmath150 under conjugation by @xmath83 and @xmath108 ( see panel b ) of fig .
[ fig : u_ks ] ) .
it transfers the rotation onto the physical legs to rotations onto the auxiliary legs .
the operator commutes with the gauge transformations , and , thus , it be used as building block of a gauge invariant hamiltonian ( see eq . ) .
[ fig : plaquette_op_ks ] ] having both gauge invariant link - operators and plaquette operators , we can now write the generalization of the @xmath151 hamiltonian of qed for a lgt with gauge group @xmath3 . this reads @xmath152 where @xmath153 is the coupling constant , and the first sum runs over links @xmath154 , while the second over plaquettes @xmath155 .
the gauge invariant hilbert space @xmath117 defined in eq . , is made of those states that fulfill all constraints @xmath156 .
this can be obtained through a projector @xmath157 @xmath158 here we show that @xmath157 has an exact tn representation .
the idea is very general and requires the contraction of several copies of two types of elementary tensors , @xmath159 tensors and @xmath160 tensors .
the two have different roles as shown in fig .
[ fig : proj_gauge ] .
the @xmath159 tensors copy the states from the _ physical _ legs to the _ auxiliary _ legs so that gauge constraints are decoupled and are imposed individually by @xmath160 tensors .
there is a @xmath161 , for every link .
they have all elements zero but those corresponding to @xmath162 @xcite .
for every site , there is a @xmath163 that only possesses auxiliary indexes ( all of its indexes are contracted ) .
has as many legs as the coordination number of the lattice , i.e. six for a triangular lattice , three for an hexagonal lattice ... ] it selects , among all states of the tensor product of the four auxiliary constituents around a site , only those that fulfill the gauge - symmetry requirements in eq . .
concretely , consider an horizontal link ( @xmath105 using the notation of fig .
[ fig : a_ks ] ) connecting sites @xmath164 and @xmath80 .
its state is copied , through the corresponding @xmath159 , to two auxiliary states , one located close to @xmath164 , @xmath153 and the other close to @xmath80 , @xmath165 ( see panel b ) of fig .
[ fig : proj_gauge ] ) . in this way we can treat the gauge constraint defined at the site @xmath164 as acting on @xmath153 rather than on the link @xmath105 , analogously the gauge constraint at the site @xmath80 can be imposed on @xmath165 rather than on @xmath105 . in this way we have been able to completely decouple the gauge - symmetry constraints acting on site @xmath164 and @xmath80 . before the copy tensors they were acting on the same link @xmath105 , while after it they act on two different auxiliary sites @xmath153 and @xmath165 .
this allows to address each gauge constraint individually . by contracting as many copies of @xmath159 as there are links on the lattice with as many copies of @xmath160 tensors as there are sites on the lattice , following the pattern in fig .
[ fig : proj_gauge ] , we obtain the desired projector @xmath157 . , which copies the physical hilbert space onto the auxiliary hilbert space , and @xmath160 , which selects only configurations fulfilling the gauge invariance condition .
b ) an example of gauge - constraint decoupling at site @xmath164 and @xmath80 obtained through the insertion of the copy tensor @xmath159 .
c ) the projector on the gauge invariant states for a @xmath166 square lattice with pbc.[fig : proj_gauge ] ] let us write explicitly the tensors @xmath159 and @xmath160 .
the only non zero elements of the copy tensor are @xmath167 .
regarding @xmath168 , there are several ways to obtain them ( all providing equivalent tensors ) . here
we follow the one inspired by the known tn expressions for the ground states of quantum doubles @xcite .
the idea is to disentangle the gauge - symmetry constraint .
it originally acts on four auxiliary sites , and we want to design an appropriate unitary transformation that transforms it to a single auxiliary site operator . in practice , we obtain this by constructing @xmath160 itself as the result of an elementary tn contraction , whose building blocks are unitary tensors that act on two constituents , and perform controlled rotations .
in particular we define the tensors , @xmath169 and @xmath170 , @xmath171 where @xmath153 and @xmath165 specify the location of the constituents in the virtual lattice .
the generic properties of these tensors are independent on the position of the constituents .
they act on the tensor product @xmath172 and transform the state @xmath173 as @xmath174 and @xmath175 .
the two operators have the following properties : @xmath176 these properties can be used in order to simplify the gauge - symmetry building blocks .
concretely , we now specify one possible arrangement of @xmath177 operators that allows to simplify the gauge condition . if we define the following unitary operator @xmath178 by using the properties we obtain @xmath179 where @xmath102 are the building blocks of local gauge transformation defined in eq . .
the physical interpretation of this transformation is that we have concentrated a four - body constraint onto a single - body constraint acting on @xmath180 that it is now easy to fulfill .
on @xmath180 , gauge invariance requires indeed to pick the only state that is invariant under the rotation for an arbitrary group element .
this is the state @xmath181 .
every state of the other three auxiliary sites , forming @xmath182 is by construction gauge invariant , since after the unitary transformation the gauge constraint does not act on those sites .
the projector operator is obtained by the equal superposition of all the gauge invariant states .
this can be done by projecting each of the three @xmath59 onto @xmath183 .
this construction is sketched in fig .
[ fig : proj_k_s ] . and @xmath160 for the kogut - susskind lgt . the tensor @xmath160 is obtained by composition of several copies of unitaries @xmath184 and @xmath185 described in the main text .
they are followed by the projection onto the states @xmath186 and @xmath187 defined in the main text.[fig : proj_k_s ] ] before proceeding , let us summarize what we have obtained so far . by re - expressing the ks lgt in the language of tns we have been able to provide a tn prescription for the projector onto the gauge invariant hilbert space @xmath117 .
this tn has bond dimension @xmath1 equal to the number of elements in the group , @xmath188 .
this means that for discrete groups this tn can be used in actual computations since it has finite bond dimension .
furthermore , the construction can be improved , as discussed in section [ sec : proj_tn_tlgt ] , where we explain an alternative choice of @xmath159 and @xmath160 that allows to express @xmath157 with a tn with bond dimension @xmath189 where @xmath190 is the dimension of the irrep @xmath38 .
this is of the order of the square root of @xmath44 since @xmath191 . in any case , when we study lgt with continuous groups , the bond dimension of the tn is infinite and thus is not useful for numerical simulations .
it provides , nevertheless , an interesting analytical results since it encodes , in a tn , the exact projector onto the gauge invariant hilbert space .
we now generalize the ks lgt to models that are described by finite dimensional local hilbert spaces while invariant under continuous groups .
this will allow to generalize our tn construction and obtain a tn expression for @xmath157 with finite bond dimension that can be used in actual numerical calculations .
the constructive approach that we have followed so far allows to generalize the original ks lgt .
here we are departing from the hamiltonian lgt whose lagrangian formulation provides the standard yang - mills action in the continuum limit .
however , we would still like to construct models that are related to the original ks lgt through a truncation of the local hilbert space that commutes with local gauge transformations . in this way , what we have discussed
so far applies as a whole to the truncated models @xcite . within our formalism
the truncation of the hilbert space we are looking for is very natural .
we use @xmath71 of eq . to pass from @xmath59 to @xmath192 and to truncate the direct sum to an arbitrary finite set of irrep s .
the minimal choice requires keeping at least two irrep s .
@xmath193 ( the reason why we need at least two irrep s will become clearer in the following ) , so that the projector can be written as @xmath194 where the @xmath55 are defined in eq . . after the projection
, if the group @xmath3 is compact , the hilbert space on a link becomes finite dimensional .
it still preserves the property of the group algebra that we have used extensively ; namely in each block it is the tensor product of @xmath195 .
this immediately allows to write left and right rotations @xmath83 and @xmath108 as @xmath196 as represented graphically in panel b ) of fig .
[ fig : hilb_space_t_lgt ] . with the isometry @xmath197 defined in eq . .
it is isomorphic to @xmath198 .
each term of the direct sum is a tensor product of two constituents so that we represent it by two solid circles on each link .
b ) the operators implementing the left and right rotations defined in eq . and are obtained from those of fig .
[ fig : hilb_space ] after conjugation with @xmath197 .
[ fig : hilb_space_t_lgt ] ] as already mentioned , there is a lot of freedom on the choice of @xmath38 and @xmath47 .
a legitimate choice is the one that minimizes the dimension of the local hilbert space , since this dimension takes part in the computational cost of tn algorithms . in this case
one should choose @xmath47 as the trivial representation and @xmath38 as the smallest faithful irrep of the group .
for example , for @xmath15 groups , the dimension of the local hilbert space with such choice is @xmath199 , where @xmath200 is indeed the dimension of the fundamental representation .
the building blocks of gauge transformations @xmath102 are still defined by eq . with @xmath83 and @xmath108
given by eq . and as represented graphically in fig .
[ fig : a_t_ks ] .
they allow to define the gauge invariant hilbert space @xmath117 by using eq . . that generates the gauge transformations in the truncated ks lgt
it is the result of the truncation of the operator @xmath102 of fig .
[ fig : a_ks ] with @xmath197 defined in eq . . on each link , the operator acts on the constituent closer to @xmath80 through a rotation , @xmath201 or @xmath202 , which is controlled by the irrep of the neighboring constituent on the same link .
it thus performs controlled operations in between the two constituents of a link , as suggested by a dot on the controller leg and by an arrow that joins it with the controlled leg .
[ fig : a_t_ks ] ] gauge invariant operators have to commute with all the @xmath116 defined above .
both single - link operators ( electric - like ) and plaquette operators ( magnetic - like ) are just the truncation with @xmath197 of the corresponding operators in the ks lgt .
the single - link operator is the truncation of eq .
, @xmath203w_t\right]_{s_n } \nonumber \\
= \left [ c^r \left ( \mathbb{i}_{r } \otimes \mathbb{i}_{\bar{r}}\right)\oplus c^{r ' } \left ( \mathbb{i}_{r ' } \otimes \mathbb{i}_{\bar{r'}}\right ) \right]_{s_n } \label{eq : one_site_op_t}\end{aligned}\ ] ] and depends only on the two free parameters @xmath204 and @xmath205 .
here is where it becomes clear that we need to keep in @xmath197 at least two irrep s .
the truncation of eq . to a single irrep is indeed proportional to the identity .
the truncated plaquette operator is built from @xmath138 in eq . ,
@xmath206 and is represented graphically in fig . [ fig : u_t_ks ] . depending on the choice of @xmath38 and @xmath47 inside @xmath197
, @xmath207 could vanish . in order to see this
we have to remind that in the ks lgt @xmath135 , is chosen as the smallest faithful irrep ( for @xmath15 this is the fundamental irrep of dimension @xmath200 ) . from the above definition of @xmath208 in eq . , and the definition of @xmath138 in eq .
, we see that @xmath208 entails terms of the kind @xmath209 .
these terms are proportional to the projector onto the trivial representation defined in eq .
, where now the @xmath108 in eq .
is given by @xmath210 .
this implies that the above terms will be non - vanishing only if the decomposition of @xmath108 in direct sum of irrep s contains the trivial representation .
this consideration implies that in order to have a non - trivial model , a certain care should be taken when choosing @xmath40 and that their choice depends on the choice of @xmath136 .
in particular , the minimal prescription provides a valid truncation scheme with non - trivial dynamics since in that case @xmath211 and @xmath47 is by itself the trivial representation .
once the @xmath208 is non trivial , it automatically fulfills the desired commutation relations with the @xmath83 and @xmath108 operators defined in eq . and , as illustrated once more in the panel b ) of fig .
[ fig : u_t_ks ] . in this way
it can be used to construct the desired plaquette operators , using the same formula of eq . where the @xmath26 are substituted with the @xmath208 just defined .
operator in the truncated ks lgt as defined in eq .
[ eq : u_t ] .
b ) it has by construction the desired covariance properties , meaning that rotations on the physical legs ( upper part of the panel ) are transmitted to rotations on the auxiliary legs ( lower part of the panel ) , as described by eq . and .
[ fig : u_t_ks ] ] the hamiltonian of the truncated lgt has thus the same form as the one of the ks lgt defined in eq .
, with the appropriate substitution of @xmath212 by @xmath213 and @xmath26 inside the plaquettes by the corresponding @xmath208 operators . in the truncated lgt we can also consider the projector on to the hilbert space of gauge invariant states , @xmath214 ^{\otimes l } \to { \cal h}_p,\ ] ] with @xmath117 defined as in eq . with the @xmath116 of containing the operators @xmath83 and @xmath108 of eq . and
even in this case this projector can be written as an exact tn .
the construction is very similar to the one used for the ks lgt in sect .
[ sec : proj_tn ] .
there , however , the specific form of @xmath160 has been derived by exploiting that , in @xmath59 , we are able to disentangle the symmetry requirements .
in general we are unable to do this explicitly , so that here we introduce a generic approach ( which can also be used for the full ks lgt and provides , indeed a tn with lower bond dimension ) .
the tn structure of @xmath215 is the same as the one of @xmath157 in the ks .
it consists of the contraction of various copies of @xmath159 tensors ( one per link ) and @xmath160 tensors ( one per site ) following the pattern in fig .
[ fig : proj_gauge ] . in particular , the four - leg tensor @xmath159 copies the physical states to the auxiliary states , while the gauge fixing tensor @xmath160 selects among the auxiliary states only those that fulfill the gauge invariance condition , @xmath216 .
chosen a basis @xmath217 , with @xmath218 and @xmath219 of the hilbert space of a link @xmath220 $ ] , the @xmath159 tensor copies the left constituent to the left and the right constituent to the right : @xmath221 @xmath160 acts then non trivially only on @xmath222 . on this space ,
the requirement of gauge invariance is equivalent to asking that @xmath160 is a symmetric tensor with respect to rotations of elements in the group @xmath3 , that is it fulfills the requirements in eq . @xmath223
as explicitly shown in panel a ) of fig .
[ fig : proj_k_s_t ] .
furthermore , since we are interested in building the projector @xmath215 , we need to build @xmath160 out of the equal superposition of all symmetric tensors acting on the above space . as explained in sec .
[ sec : symm ] , there are two possible ways of building a symmetric tensor .
the first one consists of applying the projector onto the trivial irrep defined in to the space @xmath224 .
the projector acts separately in each block of irrep s and involves terms of the type @xmath225 with @xmath226 .
it is important to notice that not all of the blocks contain a trivial irrep , and the projector will then give zero when acting on those blocks without it .
after the action of the projector , we take the equal superposition of all symmetric tensors with equal weight .
it is important at this point to find the rank of the previous projector , and take an equal superposition of all possible normalized states on its support .
this is obtained by using the tensor @xmath227 , where @xmath228 is the number of copies of the trivial irrep ( the rank of the projector ) as is represented in panel b ) of fig .
[ fig : proj_k_s_t ] .
alternatively , one can use the technology developed in ref .
@xcite for constructing invariant tensors .
the idea is to decompose the tensor @xmath160 into a @xmath229 part ( the part that is fully dictated by the symmetry constraints ) and a @xmath230 part , the part that contains the variational parameters .
the @xmath229 part , in our case , is built out from the clebsch - gordan coefficients that decompose the tensor product of the two irrep s @xmath231 into the direct sum of irrep s @xmath232 .
then every @xmath232 is again decomposed into the tensor product of two irrep s @xmath233 . in general
, this implies that the tensor @xmath230 has an extra index corresponding to @xmath232 .
the projector is obtained by taking the equal superposition of all the @xmath229 tensors with different @xmath232 , that is @xmath234 .
this is represented in panel c ) of fig .
[ fig : proj_k_s_t ] . for concreteness
we discuss how to construct @xmath160 for the ks lgt with @xmath235 truncated to the sum of the trivial irrep , plus the product of two @xmath236 irrep s .
@xmath159 copies the trivial irrep on both sides , while it copies one of the two @xmath236 to the left and the other to the right .
@xmath160 is then used to project onto the invariant states .
each auxiliary site then lives on @xmath237 that has dimension @xmath238 .
the possible blocks of the four fold tensor product @xmath239 that contain the trivial irrep are those with an even number of @xmath236 factors .
this implies that there is one block @xmath240 six blocks with @xmath241 and one block @xmath242 .
this last block furthermore leads to two values of @xmath232 , @xmath243 and @xmath244 that need to be equally taken into account .
the equal superposition of all these blocks gives the @xmath160 necessary to build the projector @xmath215 .
the explicit form of @xmath160 is given in the appendix [ app : tensors ] . that is used to build the projector onto the physical hilbert space is an invariant tensor under the action of @xmath3 as defined in eq . .
the legs of @xmath160 carry an index of the irrep , since the hilbert space on which they act has a direct sum structure labeled by the irrep @xmath38 .
b ) @xmath160 can be obtained by using the explicit form of the projector onto the trivial irrep .
the red circle is the copy tensor in the group algebra . after acting with the projector
, we have to take an equal superposition of all vectors in the trivial - irrep space , so to obtain @xmath215 .
this implies finding the image of the projector , taking an equal superposition of vectors spanning such image , state that we represent by the cyan @xmath186 , and contracting on it all open legs .
c ) alternatively one can use the construction introduced in ref .
@xcite for building symmetric tensors .
the tensor is then divided into a piece @xmath229 that takes care of correctly matching the various irrep s ( made by the clebsch - gordan coefficients ) , and a degeneracy tensor @xmath230 , made of free parameters , that assigns a four - leg tensor to each of the irrep s which enter in composition of representations . in the example
we draw , @xmath230 depends on the irrep s @xmath232 that are contained into the tensor product of @xmath245 and @xmath246 .
once more , the projector @xmath215 is obtained by taking a uniform superposition of the possible values of @xmath232 , that is setting @xmath247.[fig : proj_k_s_t ] ] we conclude this section by discussing the bond dimension of the tn @xmath215 .
it clearly depends on the choice of irrep s that one decides to consider in @xmath197 of eq . .
in the minimal case for a @xmath15 lgt , it is @xmath248 .
the truncated lgt we have just presented is not the only lgt with continuous groups , defined on a finite dimensional hilbert space . in particular
a set of models have been proposed in the literature that have the same features and are known as gauge magnets ( gm ) @xcite . here
we briefly recall their construction using the tools that we have described in the previous sections . in gauge magnets
the local hilbert space is the direct sum @xmath249 . by rewriting the direct sum into a tensor product of @xmath250
we obtain a natural basis spanned by @xmath251 in terms of a position qubit times a spin vector in @xmath252 ( more details are given in sect .
[ sect : p_gm ] ) . in this case
the constituents live in a space of dimension @xmath253 , where @xmath45 the dimension of the irrep @xmath38 .
this is , in general , smaller than @xmath254 the hilbert space of the ks lgt truncated to the same irrep @xmath38 .
for this reason gm , in their simplest version , can be considered the `` minimal lgts '' , that is the lgt with the smallest local hilbert space .
however , the two terms in the direct sum still allow to define left and right rotations of the state of a link for an arbitrary element of the group that , as we have seen , is the prerequisite for being able to define gauge transformations .
the left and right rotation for elements @xmath255 in @xmath3 are defined through @xmath256 the above equation is represented graphically in panels c)-d ) of fig .
[ fig : hilb_space_gm ] and can be rewritten in the tensor product basis as @xmath257 and @xmath258 . in this notation
it is clear that @xmath259 represents the left end of the link while @xmath260 represents the right end of the link .
the above equation allows to identify the gauge boson with a boson that can occupy one of the two extremes of the link ( for a physical implementation of this ideas with cold - atoms please see @xcite ) . , which can be identified as the left and the right constituents at the ends of each link .
graphically , we represent this local hilbert space as a single circle embracing the two copies of @xmath50 , each of them being a smaller circle identified by a different color and position inside the bigger circle .
b ) the legs of the tensors acting on this hilbert space are represented by bands rather than lines so that we can specify operators that only act on one sector as acting on half of the band .
each sub - sector is colored differently .
c)-d ) the left and right rotations in the gauge magnets only act on half of the direct sum as depicted here and discussed in eq . .
[ fig : hilb_space_gm ] ] the definition of local gauge transformations at a site slightly differs from the one in ks lgt of eq .
, @xmath261 since it rotates all the links by the element @xmath89 , independently if they are entering or leaving the site @xmath80 . in terms of left and right
constituents @xmath102 only acts non trivially on those constituents of the links that are located close to @xmath80 as illustrated in fig .
[ fig : a_gm ] .
once the building blocks of gauge transformations are defined , the discussion parallels the one for the others lgt , and in particular we can use the @xmath262 s in order to define the physical hilbert space of gauge invariant states @xmath117 .
it rotates both incoming and outgoing links by the same group elements . on every link it only acts on the constituent that is closer to the site . ]
furthermore we use the new @xmath102 in order to define gauge invariant operators as from eq . .
as a result all gauge invariant link operators are defined as @xmath263 with @xmath264 and @xmath265 arbitrary numbers . the equivalent of the @xmath138 operator is used to build gauge invariant plaquette operators . in the literature it is possible to find the specific form for the @xmath138 operators for @xmath266 and @xmath267 groups @xcite . here
we provide a recipe to generalize it to an arbitrary group @xmath3 , either discrete or continuous .
we use a similar construction than the one used for the ks lgt and define an operator that acts on @xmath268 , @xmath269 .
\label{eq : u_gm}\ ] ] if we now study how @xmath270 changes under a left rotation , we immediately see that @xmath271 , \label{eq : u_left_gm } \end{aligned}\ ] ] and that under a right rotation @xmath272.\label{eq : u_right_gm } \end{aligned}\ ] ] operator in the gauge magnet lgt defined in eq . .
it can be defined , once more , as the contraction of three @xmath39 s , and a copy tensor in @xmath59 ( red circle ) .
b)-c ) from the definition , one can check that it allows to pass the @xmath83 and @xmath108 rotations on the physical legs , induced by the gauge transformation of fig .
[ fig : a_gm ] , to analogous rotations onto the auxiliary legs .
this is exactly the covariance property that allows to define a gauge invariant plaquette operator , defined as in eq . , where the rotation on the auxiliary legs cancel by taking the trace onto the auxiliary legs .
[ fig : u_gm ] ] once more , @xmath270 allows to transmit the rotations on the link to rotations onto the auxiliary space .
this allows to use it as the building block for a plaquette operator analogous to the one of the ks lgt in eq . .
the hamiltonian of a gauge magnet has thus the same structure as the hamiltonian for the ks lgt defined in eq .
, where this time @xmath273 is substituted by @xmath274 and plaquette operators are built from the @xmath270 operators just described . as a final comment , it is important to notice that in general for non - abelian groups the gauge magnets and the truncated lgt are different models . indeed , there is no consistent truncation at the level of a single link that allows to map the ks lgt onto the gm .
this would require to decompose @xmath48 into a direct sum of irrep s and keep only two of them .
unfortunately such decomposition can not be performed in an invariant way .
this is a consequence of the fact that @xmath59 can be decomposed in an invariant way only as a direct sum of tensor products of irrep s ( as we have done through @xmath71 of in the previous sections ) .
any further decomposition of each of the terms @xmath195 into a direct sum of irrep s requires a choice of basis in one of the two factors , and thus can not be invariant under rotations by elements of @xmath3 @xcite .
there is an exception to this rule in the case of abelian lgt where @xmath275 , and consequently @xmath276 . in ref .
@xcite we have shown how to construct abelian gm as a specific truncation of the abelian ks lgt .
gauge magnets have been independently re - formulated in ref .
@xcite as _ quantum link models_. the same authors generalized them to arbitrary groups and representations @xcite introducing the concept of rishons .
both the truncated lgt and the gauge magnets , can be understood as specific quantum link models constructions , where the gm corresponds to a quantum link model with a single rishon per link while the truncated lgt corresponds to a quantum link model with two rishons per link .
it is however interesting to point out that some of the link models can be obtained as a consistent truncation of the ks lgt and others not .
we now discuss how to obtain the projector @xmath277 onto the physical hilbert space @xmath4 for gauge magnets as a tn .
the construction is similar to the one used for the other lgt . in particular @xmath277
is obtained by contracting as many copies of tensors @xmath159 as there are links on the lattice and as many copies of tensors @xmath160 as there are sites , following the patterns of fig .
[ fig : proj_gauge ] .
the tensor @xmath159 is a four - leg tensor that copies the state of physical hilbert space half into the left auxiliary space and half into the right auxiliary space .
notice that both auxiliary spaces need to be extended so that @xmath50 is embedded in a larger vector space that has at least one extra orthogonal direction that we call @xmath278 . concretely chosen a basis @xmath279 with @xmath280 and @xmath281
, @xmath159 has elements @xmath282 where @xmath283 ( remember that @xmath284 for all @xmath283 ) .
this means that the bond dimension of the @xmath159 tensor is @xmath285 with @xmath45 the dimension of @xmath50 .
we can now define the operator @xmath160 that acts on such hilbert space .
once more @xmath14 needs to be an invariant tensor , that is it needs to fulfill eq . .
interestingly the auxiliary hilbert space is isomorphic to auxiliary hilbert space of the truncated lgt tn .
nevertheless since @xmath102 is defined differently in the two models , @xmath160 is different .
the symmetry requirements induced by the gauge transformation on @xmath286 are shown in panel a ) of fig .
[ fig : proj_gm ] .
the explicit form of @xmath160 can be obtained once more using different techniques , either by using the projector onto the trivial irrep of eq . as done in the panel b ) of fig .
[ fig : proj_gm ] , or by using the standard recipes for constructing symmetric tensors of ref .
@xcite , as shown in panel c ) of fig .
[ fig : proj_gm ] . in the latter case , one first fuses the irrep s @xmath245 and @xmath246 to @xmath232 , then @xmath287 and @xmath288 to @xmath289 and then fuse @xmath232 and @xmath289 to the trivial irrep . in this way the @xmath229 part of the tensor is well defined , while the @xmath230 part depends explicitly on all @xmath289 and @xmath232 compatible with the incoming irrep .
furthermore it should be chosen such to provide a uniform superposition of all symmetric tensors .
is an invariant @xmath3-tensor , i.e. , it is left invariant under multiplying all its legs by @xmath39 .
b ) @xmath160 is obtained , for a finite group , by summing all the representation matrices through the projector on the trivial representation .
after projecting , the uniform superposition of all the states of the trivial representation is obtained by closing the free legs with the product of @xmath186 .
c ) for continuous gauge groups , the projector can either be written as an integral over the group elements with the appropriate invariant measure , or it can be obtained by using the clebsch - gordan coefficients , which transform the product of the four representations into a direct sum of irrep s , and , then , by projecting onto to the trivial irrep.[fig : proj_gm ] ] in the appendix [ app : tensors ] we provide the explicit tensors for both a @xmath12 and @xmath290 gauge magnets .
so far we have discussed how to construct the projector onto the gauge invariant hilbert space @xmath5 defined in eq . as a tn .
this means that any state of the gauge invariant hilbert space can be constructed by acting with @xmath157 on a generic state @xmath291 .
the challenging problem is still how to express @xmath278 , since in general it is a state of an exponentially large hilbert space .
one possibility would be to express @xmath278 itself as a tn , and then project it with @xmath157 .
however , even in the best scenario , in which the two states share the same structure as a tn , this rapidly becomes computationally intractable since the bond dimension of the combined tn would be the sum of the bond dimensions for @xmath278 and @xmath292 .
the alternative is to construct all states in @xmath4 directly from a symmetric variational tn ansatz .
this has the big advantage of being a sparse tn so that computations are much cheaper than in full tns having the same bond dimension .
the idea is represented schematically in fig .
[ fig : hilb_emb ] , where we show that @xmath5 is embedded in @xmath6 by drawing it as a membrane inside a 3d box . states described by the variational ansatz belong to @xmath4 and are represented as orange ovals on it .
their size increases by increasing the bond dimension of the elementary tensors , as represented by the label @xmath293 .
the simplest gauge invariant states are the ground states of the hamiltonian in the strong coupling limit , which typically are product states .
as we increase slightly the complexity of the ground - state wave function we find another simple class of gauge invariant states .
they consist of uniform superpositions of all gauge invariant states and are obtained by acting with @xmath157 onto a product state @xmath294 , @xmath295 where @xmath296 is a normalization constant such that @xmath297 and @xmath298 .
these states are an example of generalized rokshar - kivelson ( rk ) states for arbitrary gauge groups @xcite , and are described by tns with the same bond dimension as @xmath157 .
slight generalizations allow to describe a larger family or rk states by still a tn with the same bond - dimension .
this can be done for example by projecting different sates than @xmath186 with @xmath157 with the net - effect of moving away from the equal superposition of all gauge invariant states by changing the matrix elements inside @xmath159 .
alternatively , one could change the weights in the linear superposition of gauge invariant states by changing the tensors @xmath160 ( we will discuss the operators that allow to deform rk state by acting onto the physical hilbert space in sec .
[ sec : vertex_op ] ) . at this point
we have exhausted all possible of gauge invariant states that can be obtained without extending the tn .
for this reason , in order to describe all other states in @xmath4 we have to introduce a more complex variational ansatz .
this is done through a tn ansatz that consists in a superposition of spin networks @xcite .
it is again formed by two types of tensors , @xmath159 and @xmath160 , and as before , every bond of the tn is decorated by an irrep index .
there is one @xmath159 tensor for each link @xmath131 of the lattice , and one @xmath160 tensor for each site of the lattice .
the elementary tensors are contracted following the usual pattern of fig .
[ fig : proj_gauge ] .
the structure imposed by gauge invariance is exactly the same as the one discussed for the projector @xmath157 ; we will refer to it as the `` symmetry part '' of the tn .
now , however , we add to each tensor a `` degeneracy part '' .
we promote every element of the elementary tensors in @xmath157 to a full tensor acting on the appropriate degeneracy space .
this implies that to every irrep @xmath38 , we attach a degeneracy space , unconstrained by the symmetry , with dimension @xmath299 ( that can be chosen independently for every irrep r ) that contains the variational parameters of the tensor .
the various blocks @xmath38 are patched correctly thanks to the symmetry part of the tn .
we illustrate the construction for the ks lgt , but this is applicable to all the other cases we have discussed so far . on each link ,
the hilbert space is isomorphic to @xmath300 .
the tensor @xmath159 is a three - leg tensor composed by a symmetry part and a degeneracy part .
this is illustrated in panel a ) of fig .
[ fig : gen_ans ] , where the symmetry part is just made of lines representing identity matrices , while the degeneracy part is made by a tensor of free parameters .
its symmetry part is indeed the same as the @xmath159 tensor of an rk state discussed above . inside each block defined by the irrep @xmath38
, it copies the states in @xmath50 to the left and those in @xmath51 to the right .
its degeneracy part is novel with respect to the previous examples of rk states .
it is made of matrices of size @xmath301 .
these matrices ( one per block ) are shown in cyan in panel a ) of fig .
[ fig : gen_ans ] and contain the variational parameters . turning to the tensor @xmath160 in panel b ) of fig . [
fig : gen_ans ] , it is also made of two pieces , a symmetry part and a degeneracy part . the symmetry part ( once more the lower piece in the graphical representation of the tensor in panel b ) of fig .
[ fig : gen_ans ] ) does not contain any free variational parameter but simply takes cares of correctly matching the irrep s in such a way that the obtained tensor is symmetric , that is it fulfills eq . .
in the graphical representation , this is done by the clebsch - gordan tensors represented by the small black circles . to every irrep label , one associates now a degeneracy tensor ( shown in orange ) in panel b ) of fig .
[ fig : gen_ans ] , which is populated by the variational parameters .
importantly , the degeneracy tensor is obtained as a sum over the irrep s corresponding to the internal lines of the symmetric part ( @xmath232 in the figure ) . the variational state is thus expressed by the contraction of the various @xmath159 and @xmath160 given by panel c ) of fig .
[ fig : gen_ans ] , where it is represented for the specific case of a lattice made by @xmath302 sites and periodic boundary conditions are assumed in both directions .
tensor used to build a variational ansatz for a gauge invariant state .
the tensor acts on a link @xmath131 and embeds its state onto the auxiliary space .
it is composed by two parts , a `` symmetry part '' that does not contain any variational parameter ( the two lower lines ) , and a degeneracy part that contains the variational parameters ( the cyan tensors ) .
the tensor has several blocks labeled by the irrep @xmath38 .
b ) the @xmath160 tensor only acts on the auxiliary space , and it has again a block structure .
it is divided in two pieces .
the first is responsible of the correct symmetry properties of @xmath160 ( see panel a ) of fig . [
fig : proj_k_s_t ] ) .
this part does not contain any free parameters and is given in terms of the clebsch - gordan coefficients of the group ( small black dots ) .
the second part is a degeneracy part that is formed by a sum of several tensors ( one for each allowed value of @xmath232 ) acting on @xmath303 .
these tensors store the variational parameters of the @xmath160 tensor in the appropriate symmetry blocks .
c ) variational ansatz for gauge invariant states on a lattice of @xmath302 sites and periodic boundary conditions .
the network contains one @xmath159 per link of the lattice , and one @xmath160 per site .
the double lines connecting the tensors are used to remind that each of the elementary tensors has a double structure , one part dictated by the symmetry and the other one containing the actual variational parameters . in the figure
we are assuming the sum over all the irrep s @xmath38 on every bond of the tn , so that the specific irrep label is omitted.[fig : gen_ans ] ] as a second example , we can consider the @xmath12 gm , whose @xmath157 is defined in appendix [ app : tensors ] .
we start with the @xmath159 tensor for a generic gauge invariant state . in this case
the physical hilbert space of a link is two dimensional and involves only two blocks labeled by the irreducible representations @xmath304 .
on the other hand the auxiliary hilbert space has two blocks with arbitrary dimension @xmath305 and @xmath306 so that the auxiliary space is @xmath307 .
the non - zero elements of @xmath159 are @xmath308 , a full matrix that acts on the sub - block @xmath309 of the hilbert space @xmath310 @xmath311 similarly @xmath312 is an independent matrix that acts on a different block of @xmath310 , @xmath313 the same idea applies to @xmath314 that now can be thought as a collection of six tensors each of them acting on one of the only six blocks of the four fold tensor product @xmath315 ( this has 16 blocks ) allowed by the symmetry constraints .
just to give an example , one of the allowed blocks is @xmath316 , where @xmath314 has entries @xmath317 . before continuing we
summarize the results of this section .
we have proposed a variational ansatz for states of the physical hilbert space @xmath4 .
the ansatz involves the contraction of several copies of two families of tensors , @xmath159 tensors ( one per link ) and @xmath160 tensors ( one per site ) .
each of those tensors has two components , one completely determined by the requirements of gauge symmetry and a second one that contains the free variational parameters .
the resulting tn has bond dimension @xmath318 where @xmath45 is the dimension of the irrep @xmath38 , @xmath299 is the dimension of the degeneracy space associated to the irrep @xmath38 ( which is a free parameter ) , and the sum over @xmath38 extends to all the irrep s one needs to consider .
there are several advantages in dealing with such symmetric ansatz .
on one side , the ansatz can be manipulated with a cost smaller than the cost involved in manipulating a non - symmetric ansatz with the same @xmath1 , since one can work in each block separately @xcite .
also , the ansatz can be used to obtain approximations of interesting gauge invariant states ( such as eigenstates of gauge invariant hamiltonians ) , while exactly preserving the gauge symmetry .
furthermore , the ansatz allows to target not only the invariant states , but also co - variant states belonging to separate symmetry sectors .
a typical application of this scenario is the characterization of the effects of background charges on the physics of the gauge bosons .
we have just discussed how one can change , given the projector @xmath157 on @xmath4 as a tn , the parameters defining @xmath319 and @xmath160 and thus one can define a family of rk states . here
we want to address the question about what are the operators that allow to modify the rk wave function by acting on the physical hilbert space .
those operators , in principle can be added to the standard lgt hamiltonian of eq .
so to extend it and allow to explore extended phase diagrams .
the operators that allow to change the entries of @xmath159 are just the @xmath273 operators already discussed . on the other hand ,
the operators that allow to modify the entries of @xmath160 act on crosses , and thus are are co - diagonal with @xmath160 .
in particular , @xmath160 consists in an isometric tensor followed by a uniform projection onto the state @xmath320 , that is @xmath321 as explicit in fig .
[ fig : proj_k_s ] .
we can now use @xmath322 to get to the correct basis of the gauge invariant configurations and weight each of them differently through a diagonal tensor @xmath323 , @xmath324 in the ks construction of @xmath160 given in sect .
[ sec : proj_tn ] , @xmath323 has @xmath325 elements that can be chosen arbitrarily .
both the isometry @xmath326 and the vertex operator @xmath327 are represented graphically in fig .
[ fig : vertex_op ] . can be used to go to the gauge invariant states . now a diagonal operator in this space commutes with the gauge transformations , and is thus gauge invariant .
right ) the vertex operator described in eq .
is built by concatenating the tensor @xmath326 , a diagonal tensor @xmath323 ( a red circle ) ( acting on @xmath328 ) , and then @xmath329 . it acts as a potential for different gauge invariant configurations and allows to favor one with respect to another .
[ fig : vertex_op ] ] in particular we will use these operators , in oder to characterize the transition from the eight - vertex to the six - vertex model in the next section .
in this section we benchmark our proposal against known analytic results .
we start by describing the rk wave function of the @xmath11 lgt .
this is the exact ground state of the hamiltonian for @xmath330 . in this case
the @xmath26 operator is just the standard @xmath331 pauli matrix , while @xmath273 is the @xmath332 pauli matrix .
the hamiltonian reads @xmath333 where @xmath148 are the plaquettes and @xmath334 the links of the lattice @xmath79 . in the limit
@xmath335 the ground state becomes the rk state compatible with the symmetry constraints @xmath336 where as usual @xmath337 are the links around a site @xmath80 .
the above rk state is obtained by contracting a tn with @xmath338 of the form of the one in fig .
[ fig : proj_gauge ] , where @xmath339 , @xmath340 , and @xmath341 , @xmath342 , @xmath343 , @xmath344 , where as always we denote the indexes by @xmath345 , following the pattern of fig .
[ fig : a_ks ] .
as written explicitly in the appendix [ app : tensors_u1 ] , the @xmath157 on @xmath4 for the @xmath12 gauge magnet has a very similar tensor structure . in that case
however , the tensor @xmath160 misses the last two entries since @xmath346 .
the @xmath12 rk state is the ground state of the following hamiltonian @xmath347,\end{aligned}\ ] ] with @xmath348 , that is it is the ground state of the gauge magnet hamiltonian at its rk point @xcite . by applying the vertex operators defined in eq . to the @xmath11 rk state
, we can switch off the two extra elements in @xmath160 , thus effectively interpolating between the @xmath11 rk state and the @xmath12 rk state .
in particular , in order to study the transition between the two models we parameterize the elements @xmath349 , with @xmath350 . at @xmath351
we have the @xmath11 rk state while at @xmath352 we have the @xmath12 rk state .
we characterize the corresponding rk wave functions for 2d infinite cylinders with circumference @xmath83 as sketched in the panel a ) of fig .
[ fig : num ] .
( cyan ) and @xmath160 ( orange ) has bond dimension @xmath338 and its matrix elements are those described in the main text .
the elementary tensors are contracted so to provide the quantum states @xmath353 of a 2d infinitely large cylinder of circumference @xmath83 . in the drawing one
has to assume pbc along the vertical direction .
here we represent the norm of the state that defines the partition function @xmath296 of a 2d classical model .
b ) we compute the spectrum of the transfer matrix @xmath354 across the cylinder ( sketched on the left ) , which characterizes the decay of the correlations addressed in sect .
[ sec : corr ] , and the spectrum of the reduced density matrix of half of the cylinder @xmath355 ( shown on the right ) , which gives access to the entanglement characterization of the states addressed in sect .
[ sec : ent ] .
both calculations are performed via sparse exact diagonalization and their cost increase exponentially with @xmath83 so that we can address at most systems with @xmath356 .
[ fig : num ] ] for each value of @xmath357 the norm of the state obtained from the above tensors gives the partition function of the eight - vertex model , whose phase diagram was uncovered by baxter @xcite . in particular in the language of the eight - vertex model
we follow the line at @xmath358 and vary @xmath10 in the range @xmath359 . such line has also been studied in fig .
@xcite .
the interest in this specific line stands in the fact that along it , the model approaches a transition between two topological phases . at @xmath360
the eight - vertex is in a @xmath11 deconfined phase .
this is the paradigm of a @xmath11 gapped spin liquid . at @xmath361
, there is the transition between the eight- and six - vertex model so that the system enters in an algebraic spin - liquid phase .
here we analyze how to characterize the two phases and the transition between them by using our numerical ansatz .
the correlations across the cylinder are mediated by the transverse transfer matrix ( tm ) , made by the contraction of all @xmath159s and @xmath314s along a transverse slice of the cylinder as sketched in panel b ) of fig .
[ fig : num ] .
in particular at the @xmath11 lgt point @xmath351 the tm has only two degenerate non - vanishing eigenvalues @xmath362 and @xmath363 .
the first gap @xmath364 , while all others are infinite .
the model thus has zero correlation length .
as we start departing from @xmath365 , the two degenerate eigenvalues start to split so that @xmath366 starts to diverge as @xmath367 .
a family of new eigenvalues starts to appear and the model acquire a non - zero correlation length .
the decay of correlation functions is thus exponential in all this region .
the new eigenvalues tend to approach @xmath362 .
the bigger of them @xmath368 is separated from @xmath362 by a gap @xmath369 .
this is the gap that closes to zero when approaching @xmath361 ( that is at @xmath370 ) as @xmath371 .
the model thus develops algebraic correlations at @xmath370 .
our benchmark numerical results agree with this exact picture .
all the eigenvalues are computed by exact sparse diagonalization of the tm with a cost that increases exponentially with @xmath83 .
these results give a first check that our numerical technique works and we can now apply it to characterize the phase transition in terms of two of the proposed order parameters based on the scaling of the entanglement entropy : the topological entropy and the schmidt gap .
both quantities could be relevant in understanding the phase diagram of gauge theories .
gauge theories indeed present phases that can not be distinguished by using a local order parameter .
a legitimate question is whether the scaling of entanglement allows to discern those eluding phases . in the @xmath11
spin - liquid phase closes as @xmath357 approaches @xmath372 , where there is the transition from the @xmath11 spin - liquid phase to the @xmath12 algebraic spin - liquid phase .
the collapse of the data for the @xmath373 close to @xmath372 , for the values of @xmath83 in the range @xmath374 , confirms that the gap at the transition closes as @xmath375 as expected .
thus , what we are studying is a transition from a gapped phase with exponential decay of correlations to a gapless phase governed by algebraic decay of correlation .
lower panel ) on the other hand , the first gap of the tm @xmath366 , representing the gap between the two different topological sectors appearing on the cylinder , opens exponentially with @xmath83 as @xmath357 tends to @xmath372 .
this is again confirmed by the collapse of our numerical data for @xmath376 , with @xmath83 in the range @xmath377 .
[ fig : num_corr ] ] we compute the entanglement entropy of the reduced density matrix of half of the infinite cylinder @xmath355 , sketched in panel b ) of fig . [
fig : num ] .
the spectrum @xmath378 , @xmath379 of @xmath355 is computed using sparse exact diagonalization of the eigenvectors of the tm as explained in detail in ref .
the cost of the computation increases exponentially with @xmath83 .
we are interested in characterizing the scaling of the entanglement entropy as a function of @xmath83 since we want to extract the topological entropy @xmath380 . in a gapped spin liquid
the entanglement entropy @xmath381 of a region @xmath129 with boundaries of length @xmath83 scales as @xmath382 where the dots stand for the omission of higher order corrections starting with @xmath383 .
the constant @xmath380 , in the topological phases , is negative and universal , and encodes the topological entropy @xcite . in the specific case of a @xmath11 spin liquid
it is known to be @xmath384 equation holds also for the gapless spin - liquid phase described by the six - vertex model . in a series of seminal works st ' ephan and collaborators @xcite
have shown that it is indeed possible to get an exact expression for @xmath380 for the whole phase diagram of the six - vertex model .
in particular they have expressed it in the language of the @xmath385 spin chain defined on a space - like section of the cylinder .
the eigenvectors of the hamiltonian of the xxz model are indeed equal to those of the transfer matrix of the six - vertex model @xcite .
the entanglement entropy of half of the infinite cylinder corresponds , in the @xmath385 model , to the shannon entropy of the ground - state wave function of a chain with periodic boundary conditions and length @xmath83 .
the @xmath385 hamiltonian is given by @xmath386 in the range @xmath387 the hamiltonian is critical and the low - energy physics is described by a cft with @xmath388 describing a free boson compactified on a circle with radius @xmath389 . in the whole phase the topological entropy is given by @xcite , @xmath390 the specific point we are studying , called the ice - point of the six - vertex model corresponds to @xmath391 .
the numerical results we have obtained are presented in fig .
[ fig : topo_ent ] .
we have only access to modest sizes in the transverse direction @xmath392 .
@xmath380 is easy to extract close to @xmath351 and @xmath393 where we are able to recover , from our numerics , its exact analytical value . as a cross - check
we have further reduced the size of the system and considered only the smaller cylinders from @xmath394 .
the results , obtained with those sets of data ( represented by different symbols in fig . [ fig : topo_ent ] ) , still provide estimates of @xmath380 in agreement with the theory close to @xmath351 , where finite size corrections are completely negligible , and close to @xmath395 where they are reasonably small ( see the rightmost inset of fig .
[ fig : topo_ent ] ) .
the situation is very different for intermediate values of @xmath357 . especially in the region between @xmath396 , we observe strong cross - over effects .
the value of @xmath380 extracted from different series of @xmath83 do not agree as shown by the fact that curves made by different symbols are distinct .
this effect could be related to the sub - leading corrections that become more important when we approach the transition .
we also observe that for larger systems , the cross - over region shrinks and move towards the phase transition at @xmath395 .
it looks like that , if we were able to reach the thermodynamic limit , @xmath380 would present a very sharp jump between the two asymptotic values .
these results give further evidence that , provided one is able to address large enough systems , @xmath380 can be used as an order parameter , away from a relatively small cross - over region ( which further - more shrinks with increasing system size ) . in our specific case
indeed it allows to discern the gapped @xmath11 spin - liquid phase from the algebraic @xmath12 spin - liquid phase at @xmath370 .
defined in eq . , and extracted from the scaling of the entropy of half infinite cylinders with respect to their circumference @xmath83 .
the red dots are obtained by considering the scaling of the entropy in the interval @xmath397 , blue up - facing triangles @xmath377 , green squares @xmath398 , and the yellow triangles pointing to the right @xmath399 .
the red solid orange line represents the exact value of @xmath380 for the @xmath11 spin liquid in eq . , while the cyan dashed line represents its value for the @xmath12 spin liquid from eq . with @xmath108 associated to @xmath400 .
we see that , for small @xmath357 , @xmath380 coincides , independently of the size of the cylinder considered , with the expected exact value @xmath11 ( left inset ) .
as we move towards the transition @xmath380 shows a transient oscillation that tends to become sharper , deeper and move towards @xmath372 for larger @xmath83 . at @xmath372
it attains again the expected analytical value with very small corrections induced by considering the two different set of data ( right inset ) . in the main text
we provide a discussion of these results.[fig : topo_ent ] ] we now analyze another possible order parameter based on the scaling of the entanglement .
li and haldane in ref .
@xcite suggested that phases could be easier to identify by considering the scaling of the full entanglement spectrum rather than focusing on a single number as the topological entropy .
the entanglement spectrum is the collection of the logarithm of the eigenvalues of the reduced density matrix @xmath401 . in particular , numerical studies of 1d systems
have provided a precise characterization of the scaling of the lowest part of the entanglement spectrum , the one associated to the largest eigenvalues of the reduced density matrix @xcite . in many cases , the lowest gap of the entanglement spectrum , called the schmidt gap ,
vanishes when approaching a quantum phase transition following universal scaling laws .
the authors thus have proposed to use the schmidt gap as an order parameter .
this idea is further supported by the recent results that show that for conformal invariant critical points , several gaps in the entanglement spectrum close in a way that allows to identify the critical exponents of the underlying cft @xcite .
it is still unclear how general these results are .
for this reason we have decided to analyze the behavior of the lowest part of the entanglement spectrum of @xmath355 of half cylinders , when approaching the transition at @xmath402 .
in particular we address the cross - over region between @xmath403 , where the analysis of the topological entropy is unreliable .
the results are presented in fig .
[ fig : num_sp ] .
the main panel shows the first 100 values of the entanglement spectrum @xmath401 as a function of @xmath357 in the cross - over region .
surprisingly nothing strange seems to happen .
the spectrum presents the plateaux structure characteristic of the rk wave functions .
the structure of the first plateaux seems quite stable , and the only effect of increasing @xmath357 towards @xmath372 is to shift the relative height of the plateaux so to accommodate the appearance of new ones in the tails .
in particular the first two eigenvalues are degenerate for all the interval considered .
the schmidt gap is indeed constantly zero and does not detect the transition . even from the plot of the second schmidt gap , which increases monotonically with @xmath404 , we are unable to appreciate that we are approaching a phase transition as shown in the inset of fig .
[ fig : num_sp ] for several values of @xmath83 .
we thus conclude that the low - energy part of the entanglement spectrum seems to fail to detect the phase transition between the the gapped @xmath11 spin - liquid phase and the algebraic @xmath12 spin liquid occurring at @xmath370 .
( @xmath405 is on the y - axis ) of the reduced density matrix of half infinite cylinders @xmath355 defined in the panel b ) of fig .
[ fig : num ] with @xmath406 .
we plot them as as a function of @xmath357 ( x - axis ) .
@xmath357 varies in the cross - over range @xmath407 identified during the analysis of the topological entropy in fig .
[ fig : topo_ent ] .
the entanglement spectrum presents clear plateaux , footprint of rk wavefunctions .
nevertheless , its lower part does not seem to detect the phases transition .
the first schmidt gap is identically zero everywhere ( there are two degenerate eigenvalues ) , and the second schmidt gap increases while approaching the transition ( inset in the figure ) for all the @xmath377 .
[ fig : num_sp ] ]
here we try to list the works that have contributed to our understanding of various topics and to the final formulation of our proposal as we have outlined .
there are several good references for group theory in a diagrammatic notation .
in particular the book @xcite deals with continuous groups .
continuous groups and the specific case of @xmath290 have been extensively studied in the literature of tns @xcite .
the reader not familiar with the elementary concepts in the theory of group representations would benefit from studying the first few chapters of some of the standard text - books @xcite .
a nice summary of relevant material can also be found in the appendices of @xcite . a nice introduction to lgt in the lagrangian formulation can be found in @xcite and in any of the standard textbook on the subject , i.e. @xcite . the standard hamiltonian formulation was obtained by kogut and susskind , and by creutz @xcite .
truncated lgt where discussed independently by many authors @xcite that have called them gauge magnet or quantum link models .
horn originally introduced a model similar to the one described here in sect .
[ sec : t_ks ] , which we generalize here to arbitrary continuous and discrete groups . later in ref .
@xcite , the quantum link models where generalized to several continuous gauge groups with a different strategy from the one presented here . in the same work ,
they where also given an interpretation in terms of rishons . in that language , the truncated lgt we have discussed in sec .
[ sec : t_ks ] when dealing with continuous groups , is a specific quantum link model with two rishons per link , while the one in sect .
[ sec : gm ] is a quantum link model with one rishon per link .
still the results we present here can be applied also to discrete groups , and thus provide a further generalization of quantum link models .
also they allow to easily distinguish between those quantum link models that can be obtained as a consistent truncation of the ks lgt and those that can not .
the study of lattice gauge theories with matrix product states , the simplest 1d tn structure , has by now a quite long tradition , @xcite 2d lgt with tns on the other hand have been less studied .
lgt with discrete groups have been addressed with entanglement renormalization in @xcite , some aspects of them have been studied in @xcite with categorical tns , a construction that has strong connections to the present proposal . there however the emphasis is on abelian discrete groups and the lgt are addressed at the exactly solvable point obtained at @xmath408 while here we discuss a construction that allows to tackle generic groups both abelian and non - abelian , discrete and continuous at any point in the phase diagram .
lgt have also recently been addressed by using tensor renormalization schemes @xcite .
there have been several proposal on how to obtain the gauge invariant hilbert space .
the original proposal by ks was to act on a reference state with all possible gauge invariant operators .
however , there the states generated are not necessarily orthogonal and thus special care need to be taken .
t. osborne has pointed out refs .
@xcite to us that is related ( implicitly ) to our work , and we also found the discussion in ref .
@xcite very illuminating .
t. osborne himself is working along similar ideas @xcite and in particular he has independently worked out operators similar to the ones we have presented in eq . and . in the context of characterizing a family of states ( such as the rk states we have discussed ) ,
t. osborne has pointed out refs .
@xcite . in condensed matter ,
in particular , the recent results presented in ref .
@xcite have been obtained by applying similar ideas to the characterization of singlet states .
at last we have also used the available literature about tn and topological order in order , whenever possible to make connections between our ideas and the one presented in that context .
in particular we have found particularly useful the ref .
@xcite . to our knowledge
the only previous mention to vertex operators in the context of lgt is the one of ref .
@xcite however there the analysis is limited to @xmath0 and @xmath12 lgt while here we give a prescription for arbitrary groups .
it is also worth mentioning that there is an alternative connection between lgt and tn through a map of the low - energy physics of qcd in the chiral limit to the physics of specific spin chains as pursued in @xcite .
finally , several groups have recently addressed the experimental implementation of truncated lgt @xcite . with respect to the absence of the closure of the schmidt gap across a phase transition ,
a similar observation was made in ref .
@xcite and while we were preparing the final version of our manuscript in ref .
@xcite the authors have provided a plausible argument to understand that this is a quite general phenomenon for phase transitions between different rk states .
in this paper we have defined a tn framework for studying lgt .
it allows to use tns as both a lgt model - building tool ( and as such we have used to construct the minimal consistent - truncation scheme for the ks lgt ) and as a practical tool to numerically explore lgt , their phase diagrams , and their emerging properties .
the ansatz we have proposed follows the same spirit as the one proposed in ref .
the tn indeed has a symmetric part that allows to exactly encode the constraints imposed by the gauge - symmetry conditions , and a variational part , which can be optimized numerically in order to characterize interesting physical states such as the low - energy states of gauge invariant hamiltonians .
the new framework is also very powerful from the theoretical point of view . in this paper
we have been indeed able to derive through it a consistent truncation of local hilbert space of the ks lgt with continuous group to finite dimensional hilbert spaces .
we have obtained also an explicit alternative construction of gauge magnets and of their @xmath26-operators for arbitrary gauge groups [ eq . ]
that is also applicable to discrete groups , and the construction of gauge invariant vertex operators for arbitrary gauge groups [ eq . ] .
we have also been able to show that , differently from the abelian case , the non - abelian gauge magnets can not be obtained as a consistent local truncation of the ks lgt . for this reason they stand as an alternative microscopic formulation of lgt .
this result does not exclude the possibility that both gauge magnets and the ks lgt can encode the same emergent physics .
it excludes however that they are locally ( where by locally we mean at the level of a single link ) unitarily equivalent .
furthermore , the distinct form of their projectors onto @xmath4 in terms of tn ( given explicitly in app .
[ app : tensors_su2 ] and [ app : tensors_su2gm ] ) also points to the fact that their rk states are probably different ( as we will analyze in a subsequent paper ) .
this is not particularly surprising since the relation between the low - energy physics of quantum link models and of standard lgt was already discussed in ref .
@xcite and required the use of dimensional reduction arguments ( so that @xmath409 dimensional quantum link models are expected in some limit to be equivalent to @xmath1 dimensional standard lgt ) .
the tools that we have developed here can be used to analyze the entanglement content of interesting lgt states .
the entanglement , studied in a basis of states belonging to the original hilbert space , tensor product of the constituents , does not have a direct physical meaning ( since the only measurable operators in a real gt are gauge invariant operators , c.f .
the recent discussion in @xcite ) .
however , it still provides an estimate of the computational cost of simulating such states using a tn ( c.f . the related discussion in @xcite ) .
as a benchmark , we have considered the transition between the eight - vertex and the six - vertex model in terms of the rk wave function of the corresponding lgt .
we have shown that , while the transition is correctly detected by the behavior of the entanglement entropy , it is hard to detect it by observing the behavior of the schmidt gap and the lower part of the entanglement spectrum .
we envisage that the tools that we have developed will play an important role in the characterization of the real time dynamics in lgt , and in the quest for finding model displaying stable topological phases even at finite temperature .
all the recent developments about the characterization of topological phases in terms of 2d tn such as the ones of ref .
@xcite can be easily applied to our construction as we plan to in the near future .
recently , gauge magnets have received a lot of attention from the amo community , due to the possibility of implementing them in experiments based on the emerging new quantum technologies such as cold - atoms , trapped ions etc @xcite .
there is still a large room for improvement on these first proposals and the tools we have designed will help in this task .
we have indeed just become aware that a new proposal for simulating @xmath290 lgt along the lines of our discussion has already been independently designed @xcite .
furthermore , our analysis is just the starting point in the development of a tn approach to lgt . in particular ,
questions relevant for high - energy lgt like , e.g. , taking the continuous limit , have not been addressed here and constitute a logic next step to be done .
some of such questions are subject of an ambitious collaborative project , coordinated by tobias osborne , that is open to contributions and available on - line @xcite .
we would like to acknowledge i. cirac , a. ferris , f. gliozzi , g. misguich , f. verstraete , g. vidal for stimulating discussions on the topics presented .
we also acknowledge the correspondence with t. obsorne and u. j. wiese who pointed to us to many relevant references .
the results of the paper have been already presented at several conferences and the feedback from the audience allowed us to improve them .
in particular the slides of the talk given at the cecam conference networking tns hosted by the eth zurich in may 2013 are available on - line since then @xcite .
a large part of this work was developed while lt was financed by the project fp7-people-2010-iif engages 273524 .
we also acknowledge financial support by the erc adg osyris , eu ip siqs , and eustrep equam .
while we were completing this manuscript we become aware of the preprint arxiv:1312.3127 ( ref .
@xcite ) that discusses the formulation of quantum link models ( with emphasis on 1d systems ) in the context of tn . during the review process
another related work appeared on the arxiv @xcite .
130ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevb.83.035107 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.075102 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.165139 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.68.3487 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.69.2863 [ * * , ( ) ] \doibase doi : 10.1103/physrevb.80.235127 ; eprintid : arxiv:0903.5017 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-020911 - 125018 [ * * , ( ) ] link:\doibase 10.1103/physrevx.4.011025 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.010303 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.052338 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.125118 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.115126 [ * * , ( ) ] link:\doibase 10.1021/ct300211j [ * * , ( ) ] link:\doibase 10.1063/1.4798639 [ * * , ( ) ] link:\doibase 10.1126/science.1201080 [ * * , ( ) ] link:\doibase 10.1038/nphys2465 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.067201 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.040502 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.076401 [ * * , ( ) ] link:\doibase 10.1088/1742 - 5468/2004/04/p04005 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/16/3/033014 [ * * , ( ) ] http://arxiv.org/abs/1405.3259 [ ( ) ] \doibase doi : 10.1103/physrevb.83.115127 ; eprintid : arxiv:1007.4145 [ * * , ( ) ] link:\doibase 10.1007/jhep11(2013)158 [ * * , ( ) ] http://arxiv.org/abs/1310.4118 [ ( ) ] , http://arxiv.org/abs/1312.6654 [ ( ) ] link:\doibase 10.1103/physrevd.88.056005 [ * * ( ) ] link:\doibase 10.1103/revmodphys.72.1 [ * * , ( ) ] @noop _ _
( , ) link:\doibase 10.1038/nature08917 [ * * , ( ) ] link:\doibase 10.1126/science.235.4793.1196 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.78.17 [ * * , ( ) ] link:\doibase 10.1038/475280a [ * * , ( ) ] link:\doibase 10.1016/s0003 - 4916(02)00018 - 0 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.070602 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.085114 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.107203 [ * * , ( ) ] link:\doibase 10.1103/physrevd.10.2445 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.1349 [ * * , ( ) ] link:\doibase 10.1016/0370 - 2693(75)90162 - 8 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.161804 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2004.08.009 [ * * , ( ) ] link:\doibase 10.1126/science.1204621 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.072301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.202301 [ * * , ( ) ] link:\doibase 10.1016/0370 - 2693(81)90763 - 2 [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(90)90646-u [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(92)90369-m [ * * , ( ) ] http://dx.doi.org/10.1016/s0550-3213(97)80041-7 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.96.220601 [ * * , ( ) ] link:\doibase 10.1088/1742 - 5468/2007/08/p08024 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.070502 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.052104 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2376 [ * * , ( ) ] http://www.iop.org/ej/abstract/0295-5075/57/6/852/ [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.115125 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.195114 [ * * , ( ) ] link:\doibase 10.1103/physrevd.52.5743 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2004.01.004 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.110404 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.110405 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.010504 [ * * , ( ) ] link:\doibase 10.1007/s10955 - 011 - 0237 - 4 [ * * , ( ) ] http://arxiv.org/abs/1308.3318 [ * * , ( ) ] http://arxiv.org/abs/1306.2164 [ ] @noop _ _ ( , ) @noop _ _ ( , , ) @noop `` , '' ( ) , link:\doibase 10.1103/physreva.79.042308 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.050301 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/12/2/025010 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2012.07.009 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.070404 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2011.07.007 [ * * , ( ) ] link:\doibase 10.1103/physrevd.11.395 [ * * , ( ) ] link:\doibase 10.1103/physrevd.15.1128 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2012.11.009 [ * * , ( ) ] @noop _ _ ( , ) @noop _ _ ( , ) http://arxiv.org/abs/1001.4517 [ ] link:\doibase 10.1063/1.3672009 [ * * , ( ) ] link:\doibase 10.1088/1751 - 8113/45/1/015309 [ * * , ( ) ] @noop `` , '' ( ) , link:\doibase 10.1103/physrevlett.109.175302 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms3615 [ * * ( ) ] link:\doibase 10.1103/physrevd.63.085007 [ * * , ( ) ] http://arxiv.org/abs/hep-lat/0110148 [ ( ) ] link:\doibase 10.1016/j.aop.2005.01.006 [ * * , ( ) ] link:\doibase 10.1103/physrevb.65.024504 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.220403 [ * * , ( ) ] link:\doibase 10.1088/1742 - 5468/2013/12/p12010 [ * * , ( ) ] @noop _ _ ( ) link:\doibase 10.1103/physrevb.83.245134 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.184421 [ * * , ( ) ] link:\doibase 10.1088/1742 - 5468/2012/02/p02003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.237208 [ * * ( ) ] http://arxiv.org/abs/1303.0741 [ ] link:\doibase 10.1088/1742 - 5468/2007/10/p10014 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.115421 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.51.659 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physrevd.66.013002 [ * * , ( ) ] link:\doibase 10.1088/1126 - 6708/2005/07/022 [ * * , ( ) ] http://dx.doi.org/10.1103/physrevlett.112.201601 [ * * , ( ) ] p. silvi , e. rico , t. calarco , tommaso and s. montangero lattice gauge tensor networks , arxiv : 1404.7439
. http://stacks.iop.org/1367-2630/14/i=3/a=035008 [ * * , ( ) ] http://stacks.iop.org/1367-2630/14/i=10/a=103004 [ * * , ( ) ]
huan he , heidar moradi , xiao - gang wen , modular matrices as topological order parameter by gauge symmetry preserved tensor renormalization approach , arxiv : 1401.5557 .
link:\doibase 10.1006/aima.1996.0012 [ * * , ( ) ] link:\doibase 10.1006/aphy.2000.6070 [ * * , ( ) ] link:\doibase 10.1016/s0550 - 3213(99)00533 - 7 [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(79)90178 - 0 [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(81)90005 - 5 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.037202 [ * * , ( ) ] http://arxiv.org/abs/1404.5268 [ ] link:\doibase 10.1016/j.aop.2010.05.008 [ * * , ( ) ] http://arxiv.org/abs/1305.3877 [ ] http://arxiv.org/abs/1403.6780 [ ] link:\doibase 10.1038/nphys1614 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.125302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.125304 [ * * , ( ) ] link:\doibase 10.1103/physrevx.3.041018 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.120406 [ * * , ( ) ] http://arxiv.org/abs/1403.1221 [ ] link:\doibase 10.1103/physrevd.89.085012 [ * * , ( ) ] http://arxiv.org/abs/1404.3717 [ ] http://www.cecam.org/upload/files/file_1608.pdf[__ ] e. zohar , i. cirac and b. reznik _ `` local non - abelian gauge invariance out of atomic conservation laws : quantum simulation of su(2 ) lattice gauge theory using ultracold atoms '' _ , in preparation .
j. haegeman , k. van acoleyen , n. schuch , norbert , j. i. cirac , f. verstraete , _ `` gauging quantum states '' _ , arxiv : 1407.1025 .
in the case of @xmath12 both the local hilbert space and the auxiliary space are two dimensional and the tensor @xmath159 is given in the computational basis as @xmath410 while the tensor @xmath160 is given in the computational basis , @xmath411 the truncated @xmath290 lgt has hilbert space of dimension five . in it
we call the vectors as @xmath412 , in order to remember that we have two blocks the irrep @xmath413 and the irrep @xmath414 , which are direct sum , one of dimension 1 and the other of dimension @xmath415 that is tensor product of two @xmath416 dimensional spaces .
the tn that encodes the rk states can be highly simplified by noting that only a part of the hilbert space needs to be copied on the left and another part on the right .
in particular we can write the @xmath159 tensor as @xmath417 from the second line we immediately recognize that @xmath159 in the four dimensional block that is tensor product copies the left factor to the left and the right factor to the right . in this way
the auxiliary dimension is only @xmath418 .
we now need to select gauge invariant configurations on the auxiliary links by using @xmath160 .
we give the experssion of the blocks individually @xmath419
|
we discuss how to formulate lattice gauge theories in the tensor network language . in this way
we obtain both a consistent - truncation scheme of the kogut - susskind lattice gauge theories and a tensor network variational ansatz for gauge invariant states that can be used in actual numerical computations .
our construction is also applied to the simplest realization of the quantum link models / gauge magnets , and provides a clear way to understand their microscopic relation with the kogut - susskind lattice gauge theories .
we also introduce a new set of gauge invariant operators that modify continuously rokshar - kivelson wave functions , and can be used to extend the phase diagrams of known models . as an example we characterize the transition between the deconfined phase of the @xmath0 lattice gauge theory and the rokshar - kivelson point of the u(1 ) gauge magnet in 2d in terms of entanglement entropy .
the topological entropy serves as an order parameter for the transition , but not the schmidt gap .
tensor network ( tn ) techniques are starting to play an important role in our understanding of many - body quantum systems , both on the lattice and in the continuum .
they can be used as a framework to classify the phases of quantum matter @xcite , or as powerful numerical ansatz in actual computations of 1d @xcite and 2d strongly correlated quantum magnets @xcite , fermionic systems @xcite , or anyonic systems @xcite .
they have also recently made their way into quantum chemistry as computational tool to study the structure of molecules from the first principles @xcite .
while numerical simulations based on monte carlo ( mc ) are still the most successful techniques in some of these fields , tns start to provide viable alternatives to them , particularly in those contexts where mc has troubles , such as the physics of frustrated anti - ferromagnets @xcite , and the real time evolution of out of equilibrium systems @xcite . at present
, the main limitation of numerical tn techniques is that the cost of the simulations increases rapidly with the amount of correlations in the system ( which is encoded in the bond dimension @xmath1 of the elementary tensors ) , and thus tns tend to be biased towards weakly correlated phases .
however , the steady improvement of the tn algorithms @xcite makes us confident that these limitations will soon be overcome , and as a consequence tn will become more and more useful in the physics of quantum many - body systems . among interesting quantum many - body systems , we focus here on gauge theories , a context in which tn
have recently made a spectacular debut @xcite .
gauge theories ( gt ) @xcite describe three of the four fundamental interactions ( electromagnetic , weak , and strong interactions ) .
in particular , strong interactions , are described by an @xmath2 gauge theory , called quantum chromo - dynamics ( qcd ) @xcite ) .
gt also allow to understand emergent phenomena at low energies in condensed matter systems , e.g. anti - ferromagnets @xcite and high - temperature superconductors @xcite .
the phase diagrams of gt , similarly to those of most strongly correlated many - body quantum systems , are still debated .
still , there are exactly solvable gt that display topological phases .
recently , topological states have been proposed as possible hardware for quantum computers , and thus there is an urgent need for clarification of generic gt phase diagrams where these states could appear @xcite .
wilson s formulation of lattice gauge theories ( lgt ) @xcite was obtained by substituting the continuous space - time with a discrete set of points ( the lattice ) .
it provided the breakthrough that has allowed to develop numerical tools based on mc , able to address the strong coupling regime of gt .
these tools are , as today , the main resource to compare various aspects of qcd at strong - coupling with experiments @xcite . those aspects of qcd that are hard or impossible to address with mc are indeed in most of the cases still unclear .
for example , the mechanism of charge confinement @xcite , invented to explain the absence of isolated quarks @xcite , still stands as a conjecture in full qcd , four decades since it was first understood in abelian models .
furthermore , mc simulations struggle to address hot and dense nuclear matter @xcite , probed by heavy nuclei collisions at cern and rhic @xcite . in their current formulation mc simulations of lgt
can not be used to characterize the real time out of equilibrium dynamics of gt .
the aim of this paper is two - fold .
on one side we develop the theory of tns for lgt with arbitrary groups . on the other
, we provide a constructive approach to lgt using the tn formalism ( reviewed in sec .
[ sec : tn ] ) . in this framework , tn are used as a model - building tool that , given a group @xmath3 , allows to design the most general gauge invariant theory out of the simple knowledge of the group representation matrices .
the approach is based on reformulating very simple results about the theory of group representations ( reviewed in sec .
[ sec : gt ] ) in the tn formalism .
in particular , we use the approach in which lgt differ from standard many - body quantum systems due to the presence of a large amount of local symmetry constraints ( sec . [
sec : lgt ] ) , which arise as a consequence of generalizations of the gauss law .
one of the guiding principles in designing lgt will thus be the possibility of defining such local symmetry constraints ( sec .
[ sec : lgth ] and sec .
[ sec : lgt_gt ] ) .
in particular , we identify the `` physical hilbert space '' , @xmath4 , as the space of states that fulfill those constraints ( sec .
[ sec : lgt_ph ] ) .
we show how these constraints can be naturally embedded in a tn .
we thus construct an exact projector onto @xmath4 as a tn ( sec .
[ sec : proj_tn ] ) .
we re - derive , with our formalism , the elementary gauge invariant operators ( sec . [
sec : g_inv_op ] ) necessary to describe the dynamics inside @xmath4 .
in the course of our discussion we explain that continuous groups are associated to infinite dimensional local hilbert spaces , and thus the tn network construction for them has infinite bond dimension @xmath1 , and thus is computationally intractable . in order to cure this
, we introduce a scheme that allows to truncate , in a gauge invariant way , the infinite dimensional local hilbert spaces . in this way
we obtain a version of the ks lgt for ( compact ) continuous groups defined on finite dimensional hilbert spaces ( sec . [
sec : t_ks ] ) .
the projector onto @xmath4 for these models can be expressed as a tn with finite bond dimension , and can thus be used in practical computations .
we review the alternative constructions of lgt with continuous gauge symmetry and discrete local hilbert spaces , called gauge magnets ( or link models ) @xcite ( sec . [
sec : gm ] ) .
we generalize it to arbitrary groups and we show that in the non - abelian case , the gauge magnets are not equivalent to a local gauge invariant truncation of the ks lgt . at this stage
we are able to introduce a general tn variational ansatz for lgt , with both discrete and continuous groups , that again automatically embeds all local constraints dictated by the gauge symmetry .
the states described by this ansatz are indeed gauge symmetric by construction .
gauge - symmetry constraints , indeed , allow to restrict the attention to @xmath4 , which is still , however , exponentially large as shown in fig .
[ fig : hilb_emb ] .
low - energy states of local gauge invariant hamiltonians are expected to live only on a small region of @xmath4 , in the same way as low - energy states of generic local hamiltonians live in a small region of the unconstrained hilbert space , since they fulfill the `` area - law '' for the entanglement @xcite . of a quantum many - body system ( represented here by a 3d box )
is exponentially large , since it is the tensor product of the hilbert spaces of the constituents .
gauge symmetry allows to identify a smaller space that we call the _ physical hilbert space _ @xmath5 .
this is the subspace spanned by those states that fulfill all the local constraints imposed by the gauge symmetry and is represented by a membrane inside @xmath6 .
@xmath5 is smaller than the full @xmath6 but it is still exponentially large .
low - energy states of local gauge invariant hamiltonians , however , are expected to live in a small corner of @xmath5 , in the same way as low - energy states of generic local hamiltonians live in a small corner of @xmath6 @xcite . for this reason we design a variational ansatz based on tn that allows to explore this small corner of @xmath5 ( orange oval ) . by increasing the bond dimension of the elementary tensors in the tn ( @xmath7 )
we can explore increasingly large regions of @xmath5 , and eventually for @xmath8 we can cover the whole @xmath5 .
the projector on @xmath4 , and a family of interesting rokshar - kivelson ( rk ) states , are obtained exactly with tn with minimal bond dimension that scales as @xmath9 where @xmath10 is the dimension of the local hilbert space .
[ fig : hilb_emb ] ] we explicitly construct a tn ansatz that allows to explore this small corner of @xmath5 ( sec .
[ sec : var_anst ] ) . in its simplest form
, the tn ansatz requires the same bond dimension as the projector onto @xmath4 . in this case @xmath9 , with @xmath10 the dimension of the local hilbert space , and it allows to characterize the physics of generalized rokshar - kivelson states ( rk ) @xcite . by increasing the bond dimension
, one can gradually explore all the space of gauge invariant states @xmath4 , as represented in fig .
[ fig : hilb_emb ] , by increasingly large orange circles .
the tn ansatz depends on several elementary tensors , each made by two distinguished parts .
one part is completely determined by the gauge - symmetry constraints , while the other contains the free parameters to be used in variational calculations . as in the case of globally invariant tn @xcite
, our formalism allows to unveil interesting connections between gauge symmetric quantum states and spin networks @xcite . as a further application
we define gauge invariant vertex operators @xcite for arbitrary gauge theories ( sec . [
sec : vertex_op ] ) . in this way we open new possibilities to use them as extensions of the standard hamiltonians in order to explore extended phase diagrams of the known models .
we benchmark these new tools in the context of rk states .
in particular we focus on the recent proposals about the characterization of quantum phases based on the analysis of the entanglement scaling of the ground state wave function @xcite .
we analyze the well known transition between the eight - vertex and the six - vertex models . in the gauge theory language
this transition is induced by applying the vertex operators @xcite onto the rk @xmath11 wavefunctions . in this way we provide an example of phase transition between a @xmath11 gapped spin liquid and a @xmath12 algebraic spin liquid that is detected by the topological entropy , but elusive for the lowest part of the entanglement spectrum ( sec .
[ sec : num_res ] ) .
all the discussion about connections of our results with other works in the literature is postponed to sec .
[ sec : about ] , and we conclude with a summary of our results and an outlook on future developments in sec . [
sec : conc ] .
|
the discovery by cosmologists that the universe is currently dominated by two distinct types of unknown forms of matter is a development with truly copernican implications for our picture of the universe as a whole .
we have known for some time that visible matter likes to cluster on large scales into galaxies and galaxy clusters , with 90% or more of the mass of these objects consisting of an unknown nonbaryonic ` dark matter ' @xcite .
the more recent surprise was the discovery that this clustered dark matter itself makes up no more than 30% of the overall energy density of the universe , with the remaining 70% apparently consisting of a different kind of unknown substance , sometimes called ` dark energy ' @xcite .
it is absolutely breathtaking that so little is known about these two most abundant forms of matter .
what is known is usually expressed in terms of their equations of state , through the ratio of pressure to energy density , @xmath0 .
the formation of galaxies and galaxy clusters by gravitational attraction appears to require the dark matter to be ` cold ' , with @xmath1 close to zero .
similarly , the current acceleration which the universal expansion is undergoing indicates @xmath2 for the dark energy .
perhaps the simplest explanation for the dark energy is that it is simply the energy density of the quantum vacuum , since this satisfies @xmath3 and is generically nonzero in realistic quantum field theories .
the problem with this explanation is that the predicted vacuum energy is typically at least @xmath4 times larger than what is observed .
the cosmological constant problem @xcite is the recognition that at present no way is known to naturally obtain a vacuum energy anywhere near the required size within a realistic microscopic theory .
quintessence models @xcite take a different tack to explain the dark energy . in these models
the dark energy is attributed to the dynamics of a scalar field , @xmath5 , which is currently evolving in a cosmologically interesting way . since @xmath6 , where @xmath7 and @xmath8 are the scalar s kinetic and potential energies , the condition that @xmath1 be negative requires the scalar s evolution at present must be slow in the sense that @xmath9 .
( the vacuum energy is a special case of this kind of evolution , where @xmath10 . )
these models do not solve the cosmological constant problem , since they do not provide natural reasons for @xmath8 to be currently so small , but they can ( potentially ) explain why an evolving scalar field could naturally have an energy density which is now so similar to that of other forms of matter , like the dark matter .
they can do so , firstly because their equations of motion often admit ` tracking ' solutions , within which the scalar energy density closely follows ( or tracks ) the dominant energy density of the universe as it evolves .
furthermore , the late - time evolution of the scalar field is often drawn to these ` tracking ' solutions for wide choices of initial conditions , because these solutions are also ` attractors ' for the scalar equations of motion . unfortunately , since these tracker solutions typically require the scalar itself not to be the dominant energy density , in order to become the dark energy the scalar must eventually leave the tracker solution . although one might hope that this could also be naturally achieved such as being perhaps due to a transient behaviour due to the crossover from radiation to matter domination
so far it has proven difficult to make a completely convincing cosmology along these lines .
the purpose of this article is to describe a new category of quintessence model , which could be called ` walking quintessence ' @xcite , that may offer new ways to accomplish this crossover .
they may do so because within these models the couplings of the scalar field slowly run ( or walk ) as the universe evolves , and this walking may help facilitate the crossover between tracking solutions .
what is remarkable is that these models were developed in an attempt to address a completely different set of ( very serious ) problems which arise once one tries to obtain quintessence from a realistic model of microscopic physics .
since they play such an important part in the motivation of the models , the bulk of the article is devoted to these serious microphysical problems .
the problems themselves are first summarized in the next section , followed by a description how they are addressed within the attractive brane - world picture which has emerged as a potential low - energy consequence of string theory .
the results of a sample cosmology built on this model are then briefly presented , intended as first step towards a more systematic exploration of the cosmologies which are suggested by this class of models .
from a microscopic perspective any viable quintessence model must have two very remarkable properties , which turn out to be quite difficult to arrange @xcite .
as is explained below , they must have : * * extremely light scalars * , whose mass must be at most @xmath11 ev , and ; * * no new long - range forces * , to the extent that these would ruin the agreement between general relativity and observations on earth , within the solar system and beyond @xcite . the first requirement ( very light scalars ) is a very generic property of explanations of the dark energy in terms of a rolling scalar field . its necessity may be seen from the scalar field equation of motion within a cosmological context : [ kgeq ] + 3 h + v = 0 , h^2 = 3 m_p^2 , where @xmath12 gev is the rationalized planck mass , and @xmath13 is the total energy density ( which is at present dominated by the dark energy ) . to quantify how small this requires the scalar mass to be , it is instructive to consider the very broad class of models within which [ vform ] v ( ) = ^4 u(/f ) . here
@xmath14 and @xmath15 are arbitrary mass scales , whose size may be determined if the dimensionless function @xmath16 and its derivatives are at present @xmath17 . in this case
the present value of the scalar potential and its derivatives are @xmath18 , @xmath19 , _ etc .
_ , and the square of the scalar mass is of order @xmath20 .
this turns out to be of order @xmath21 given the values which are required for @xmath14 and @xmath15 .
the value @xmath22 ev is inferred by recognizing that these estimates also apply to the total scalar field energy , @xmath23 , since present observations require the scalar field to be rolling with @xmath24 less than but of the same order of magnitude as @xmath8 . in this way
we see that @xmath14 controls the present dark energy density , @xmath25 , and so also @xmath26 and @xmath27 .
the value @xmath28 is determined from the scalar field equation , eq .
, together with the above slow - roll conditions , since this implies the @xmath29 term should be much smaller than the other two . except for the case of a pure cosmological constant ( for which only the last term is important )
, we therefore have @xmath30 and so @xmath31 , from which we learn @xmath32 and hence @xmath33 ev .
such an extraordinarily small scalar mass is extremely difficult to achieve in a realistic microscopic theory .
there are two separate aspects to this difficulty . 1 .
* heirarchy problem 1 : * how does such a small nonzero mass arise as a combination of microscopic parameters ? 2 .
* heirarchy problem 2 : * given that such a small mass is predicted by the theory of microscopic physics , how does it remain small as one integrates out all the physics between these microscopic scales and the cosmological scales at which it is measured ?
this is a problem because , for instance , a particle of mass @xmath34 and coupling @xmath35 shifts the scalar mass by an amount m ~m^2 4 f when it is integrated out .
both of these problems are the direct analogs of two aspects of the famous heirarchy problem as applied to the higgs field which breaks electroweak gauge symmetry within the standard model of particle physics . there the weak scale , @xmath36 gev is controlled by the scalar higgs mass , and one asks how this can be so much smaller than , say , the planck scale , @xmath37 gev .
the problem for quintessence models , however , is arguably much worse for two reasons .
first , the quintessence scalar is many more orders of magnitude lighter than the basic microscopic physics scale @xmath38 than @xmath38 itself is from @xmath39 .
second , the range of scales between @xmath38 and @xmath40 is well studied by experiment , which presumably makes it harder to hide the other degrees of freedom which are typically invoked to alleviate heirarchy problem 2 ( such as the superpartners which do the job if supersymmetry is the solution ) .
almost none of the extant quintessence models address these issues , with the exception being those based on pseudo - goldstone bosons @xcite , which address heirarchy problem 2 .
( modifications of these models based on the brane world can also address heirarchy problem 1 @xcite . )
the models presented in the later sections are unique among those yet proposed in that _
_ b__oth of these problems are related to the same microscopic quantities which explain why @xmath41 .
the incredibly small quintessence scalar mass also raises a related observational problem , since it implies that the exchange of this scalar must mediate a very long - range force .
furthermore , the strength with which this force couples to particles of energy @xmath42 is typically of order @xmath43 , which makes it comparable to gravity .
this is problematic , since many observations now strongly constrain the existence of gravitational - strength forces between macroscopic objects having a range longer than about 0.1 mm . in the models which are described in subsequent sections this problem
is evaded because the scalar couplings turn out to evolve over cosmological time scales . in the examples
given it happens that these couplings evolve to become extremely small during the present epoch , which is when all of the very constraining observations have been made .
another way to ensure acceptably small couplings is to arrange the scalar to couple to quantities , such as spin , which ordinary matter in bulk does not carry .
this kind of coupling can occur in pseudo - goldstone - boson models @xcite .
new insights into naturalness problems , like the heirarchy between @xmath38 and @xmath39 , have been made based on the recently much - discussed brane world picture , in which all observed elementary particles are confined to a domain - wall - like surface ( or ` brane ' ) which sits within a higher - dimensional ` bulk ' spacetime .
the simplest choice puts us on a 4-dimensional surface ( or 3-brane ) within a bulk space which has anywhere from 5 to 11 dimensions .
by contrast , gravitational interactions in this picture are not confined to these branes .
this kind of picture is very well motivated within string theory .
the fact that gravity and other interactions do not see the same number of dimensions lies at the root of the surprising realization that the fundamental string scale , @xmath44 , can be much smaller than @xmath39 @xcite , and of the new perspective this has allowed for understanding low - energy naturalness problems . in particular , it can imply the relationship [ wpexp ] ( m_s m_p ) ^2 ~^2 ( m_s r)^n , where @xmath45 is the ( open ) string coupling and @xmath46 is the radius of the @xmath47 extra dimensions . for supersymmetric systems
one generically has @xmath48 and so this remarkable formula allows the observed ratio @xmath49 ( hierarchy problem 1 , above ) to be understood using parameters themselves no smaller than @xmath50 if @xmath51 gev and there are @xmath52 extra dimensions @xcite .
( in this picture it is supersymmetry which accounts for heirarchy problem 2 . )
the brane - world variant which is of most interest for the present purposes , puts the string scale as low as is consistent with experiment , @xmath53 @xcite . from eq .
( [ wpexp ] ) , one sees this is permitted ( even if @xmath45 is not small ) provided @xmath46 is large enough .
for instance , if @xmath54 then @xmath55 mm ( or @xmath56 ev ) . in this picture @xmath57 , so the heirarchy problem is not solved so much as translated into the problem of understanding the origin of the large heirarchy @xmath58 .
a remarkable feature of this scenario is that it provides a framework within which scalars can be naturally as light as @xmath59 ev , and it is instructive to see how this works for specific examples .
the success of the models of ref .
@xcite relies on choosing the extra dimensions to be torii and the quintessence field to be a component of the extra - dimensional metric
the radion , @xmath46 .
this construction directly solves heirarchy problems 1 as follows .
first , because the radion field starts life as a component of the six - dimensional metric its kinetic term has the same origin as does the 4d graviton .
dimensionally reducing the 6d einstein - hilbert action to four dimensions gives [ 4dkin ] l_kin = - m_p^2 2 g^ , from which we see the canonically - normalized field is @xmath5 where @xmath60 . from this
it follows that @xmath61 . for toroidal compactifications , direct dimensional reduction of the higher - dimensional einstein - hilbert action gives no radion potential at all ( provided that the cosmological constant is chosen to vanish , as usual )
this is an artifact of the classical approximation , however , and a potential for @xmath46 is generated once quantum effects are included , such as through the casimir effect which predicts ( for large @xmath46 ) a potential of the form @xmath62 @xcite .
if this potential can be stabilized to have a minimum for some @xmath63 ( more about this later ) , then it predicts @xmath64 .
more remarkably , this model also addresses heirarchy problem 2 , since these predictions are protected from being ruined as physics between the weak scale and the quintessence mass is integrated out .
although the stability of the prediction for @xmath15 against quantum corrections is fairly trivial , it is the stability of the potential which bears closer examination .
there turn out to be two reasons for this success .
the casimir effect predicts @xmath65 largely because the potential is generated when modes having energies of order @xmath66 are integrated out .
now consider the effect of integrating out modes with energies lower than @xmath66 . for scales
@xmath67 the effective theory is four - dimensional , and the naive corrections to @xmath68 are correct , since no symmetries preclude generating a radion potential .
integrating out a mode of mass @xmath34 then contributes to @xmath68 terms of order @xmath69 , which is not dangerous since it predicts @xmath70 .
the key is at energies above the scale @xmath66 , where the effective theory is six - dimensional and so is constrained by additional symmetries like 6d general covariance . for these scales
the result of integrating out modes with @xmath71 may be expanded in powers of the 6d curvature @xmath72 .
if the extra dimensions were to have spherical geometry then @xmath73 , and these curvature terms constitute dangerously large contributions to the radion potential . for flat spaces like torii ,
however , @xmath72 is independent of @xmath46 and these terms are not dangerous .
the question for torii becomes whether it is possible to keep the internal dimensions flat despite the existence of large vacuum energies on the various branes on which our observed particles live .
it is a special feature of two dimensions that this is so , since einstein s equations predict flat geometries around point sources @xcite .
we see that in this picture the quintessence scales @xmath14 and @xmath15 are predicted to be respectively given by @xmath66 and @xmath39 because the quintessence field has its microscopic origin as part of the higher - dimensional geometry .
the success of the predicted dark energy density , @xmath74 , and quintessence mass , @xmath75 , then follow from the choice @xmath56 ev which is required in any case to solve the electroweak heirarchy , since @xmath76 .
this success does not crucially hinge on identifying the quintessence field as a mode of the metric , since variant brane - world models with similar properties may also be built wherein the quintessence field is a pseudo - goldstone mode in the bulk @xcite .
in order to more precisely pin down the cosmology the explicit form for the quintessence potential is required , and within the brane world picture being presented this amounts to providing a mechanism for stabilizing the radion at large values . this section relates a concrete proposal for doing so , as made in ref .
@xcite .
besides providing an interesting cosmology in its own right , there is another reason for exploring this specific proposal in more detail .
this other reason is to illustrate why worrying about naturalness issues is useful when exploring phenomenological applications ( such as quintessence cosmology ) .
naturalness issues are useful precisely because the cosmology of quintessence models depends in such a detailed way on the precise form of the low - energy potential . on one hand ,
the naturalness problem states that only very specific kinds of potentials are likely to arise as the low - energy limit of realistic microphysics . on the other hand ,
it has proven difficult to build a completely convincing quintessence cosmology purely by guessing different kinds of potentials .
it may be true that it is only the very few potentials which can arise from real microphysics that can also provide a realistic description of cosmology as well .
if so , it should be invaluable to be able to explore in detail any potential which _ _ d__oes arise as the low - energy limit of real microphysics .
in the present instance the stabilization mechanism proposed suggests a qualitative new feature of the low - energy theory : it predicts that the effective low - energy couplings and masses depend logarithmically on the quintessence field , and so these all slowly run ( _ i.e. _ walk ) over cosmological time scales .
this leads in a natural way to extended quintessence models @xcite , but with a specific and cosmologically - interesting type of field - dependence . indeed
the proposal of ref .
@xcite was initially motivated by the earlier discovery of the attractive cosmologies which can result from quintessence models having exponential potentials with logarithmic corrections @xcite .
the models now described are based on the observation that @xmath68 would naturally be minimized at large values for @xmath46 if it were to depend logarithmically on @xmath46 : [ logform ] v(r ) = v_0 ( r)^p + , where @xmath77 and @xmath78 are constants and @xmath79 is a microscopic length scale , such as @xmath80 .
the ellipses indicate other terms which fall off with a higher power of @xmath81 relative to those shown .
a potential of this form has a minimum at @xmath82 , which is exponentially larger than @xmath79 if @xmath78 is moderately smaller than one .
( values @xmath83 are sufficient to generate the desired heirarchy in the models explored in @xcite . )
the key point is that a potential of precisely this form is expected if the effective six - dimensional theory at scales @xmath84 contains a renormalizable ( _ i.e. _ marginal ) coupling , @xmath85 .
if so , then this coupling runs logarithmically with @xmath46 and loop corrections to the radion potential involving only this coupling have the logarithmic form of eq . , with a coefficient @xmath86 which is naturally small . in six dimensions most interactions
are not renormalizable , but such couplings can exist , such as a cubic coupling amongst six - dimensional scalar fields . given this generic mechanism it is clear that logarithms in @xmath46 are not restricted to appear in the low - energy theory only within the radion potential .
rather , they arise generically as radiative corrections to _ _
a__ll couplings , and it is this logarithmic dependence which is responsible for the walking of these couplings as @xmath46 evolves over cosmological time scales .
logarithmically evolving couplings are very attractive for quintessence cosmology for two reasons .
first , they introduce non - minimal couplings between the quintessence field and ordinary matter , and these turn out to provide new kinds of scaling attractor solutions .
which attractor is the endpoint for a given choice of initial conditions is determined by the parameters of the quintessence potential , and this observation is at the root of the second attractive cosmological feature of having walking couplings .
a given cosmological solution can cross over from the basin of attraction of one attractor to another since the walking of the couplings moves the boundaries of these domains of attraction within field space .
@xcite provides a preliminary exploration of the cosmology of models built along these lines .
an example of a viable cosmology which is obtained in this way is illustrated in figure 1 , which shows the cosmological evolution of the energy density in radiation , matter and the radion kinetic and potential energies .
figure 2 shows , for the same model , the coupling , @xmath87 , of the radion to ordinary matter as a function of the universal scale factor , from which it is clear that the model evades current constraints on long - range forces ( which require @xmath88 ) . a constraint which is specific to models where it is the radion which is the quintessence field is the requirement that the radion not roll too far since the epoch of nucleosynthesis , since such a roll would predict an unacceptably large change in the ratio of weak and gravitational couplings , @xmath89 , between then and now .
figure 3 shows how @xmath90 evolves for the cosmology shown in figure 1 , and so demonstrates that its evolution can be acceptably small .
although a promising start , this cosmology still leaves much to be desired . in particular , the constancy of @xmath46 is not generic and must be partially arranged by adjusting the parameters of the potential and the initial energy in the scalar roll .
however the broader framework of walking quintessence models seems to provide a fruitful place to search for a quintessence model which is truly natural in all senses of the word .
i would like to thank the organizers of this workshop for providing such stimulating environs , and for their kind invitation to speak .
my part of the work summarized in this article was done in collaboration with andy albrecht , finn ravndal and constantinos skordis .
it was started at the aspen center for physics , and was largely carried out at the institute for advanced study in princeton with financial help from the ambrose monell foundation .
my own research during this period was supported by grants from n.s.e.r.c .
( canada ) and f.c.a.r .
( qubec ) .
peebles , _ publ .
pacific _ * 111 * ( 1999 ) 274 ; m.s .
turner , _ publ .
pacific _ * 111 * 264 ; a. dekel , d. burstein and s.d.m .
white , _ critical dialogues in cosmology _ ( 1997 ) p. 175 .
k. coble , s. dodelson and j. a. frieman , phys .
d * 55 * , 1851 ( 1997 ) ( astro - ph/9608122 ) ; r. r. caldwell and p. j. steinhardt , phys .
d * 57 * , 6057 ( 1998 ) ( astro - ph/9710062 ) ; i. zlatev , l. m. wang and p. j. steinhardt , phys .
* 82 * , 896 ( 1999 ) ( astro - ph/9807002 ) ; p. j. steinhardt , l. m. wang and i. zlatev , phys .
d * 59 * , 123504 ( 1999 ) ( astro - ph/9812313 ) .
p. horava and e. witten , nucl .
* b475 * ( 1996 ) 94114 ( hep - th/9603142 ) ; nucl .
* b460 * ( 1996 ) 506524 ( hep - th/9510209 ) ; e. witten , nucl .
* b471 * ( 1996 ) 135 ( hep - th/9602070 ) ; j. lykken , _ phys .
rev . _ * d54 * ( 1996 ) 3693 3697 ( hep - th/9603133 ) ; i. antoniadis , phys.lett . *
b246 * ( 1990 ) 377 384 . n. arkani - hamed , s. dimopoulos and g. dvali , _ phys .
lett . _ * b429 * ( 1998 ) 263272 ( hep - ph/9803315 ) ; phys . rev .
( 1999 ) 086004 ( hep - ph/9807344 ) ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. dvali , phys .
* b436 * ( 1998 ) 257263 ( hep - ph/9804398 ) .
|
although quintessence models have many attractive cosmological features , they face two major difficulties . first , it has not yet been possible to find one which convincingly realizes the goal of explaining present - day cosmic acceleration generically using only attractor solutions .
second , quintessence has proven difficult to obtain within realistic microscopic theories , largely due to two major obstructions . both of these difficulties are summarized in this article , together with a recent proposal for circumventing the second of them within a brane - world context .
it is shown that this proposal leads to a broader class of dynamics for the quintessence field , in which its couplings slowly run ( or : ` walk ' ) over cosmological time scales .
the walking of the quintessence couplings opens up new possibilities for solving the first problem : that of obtaining acceptable transitions between attractor solutions .
|
in this note we solve the additive eigenvalue problem ( or the time - independent system ) associated to the dynamics of a basic 2d - traffic model considered in @xcite .
it is a system of two circular roads crossing on one junction managed with the _ priority - to - the - right _ rule .
the model is based on petri nets and on minplus algebra @xcite .
it is an extension to an existing 1d - traffic model @xcite , which gives the average speed of vehicles on one circular road as an eigenvalue of a minplus matrix , and thus allows the derivation of the fundamental diagram of 1d - traffic ( the relation between the density and the flow of vehicles on the road ) .
we give a solution of the eigenvalue problem .
we show that the eigenvalue is not necessarily unique , but is given in terms of two main quantities which are interpreted in terms of traffic as the density @xmath0 of vehicles in the system , and a parameter @xmath1 giving the ratio of the non priority road size with respect to the size of the whole system .
we give a condition on the parameter @xmath1 such that the eigenvalue is unique and positive for non - high densities . in this case
the eigenvalue problem can be written as a dynamic programming equation of a stochastic optimal control problem .
we use the minplus algebra notations , mainly for reason of compactness but also to use some classical results of this algebra @xcite .
in addition , the following notations are also used : @xmath2 denotes @xmath3 , @xmath4 denotes @xmath5 , and @xmath6 denotes @xmath7 .
the traffic dynamics is the following ( see @xcite ) : @xmath8 with the ( traffic ) constraints : @xmath9 for example , in the usual algebra , the equation ( [ nimplicit ] ) is written : @xmath10 when the equation ( [ ar1 ] ) is written : @xmath11 this system of equations is implicit but it is triangular , so its trajectory is unique .
we denote by @xmath0 the following quantity ( which is interpreted in terms of traffic as the density of vehicles in the system ) : @xmath12
the additive eigenvalue problem corresponding to the dynamics ( ds ) is : @xmath13 with the constraints ( [ cont ] ) and the notation ( [ nota ] ) . our aim in this note is to solve the system @xmath14 .
solving the system @xmath14 is equivalent to solving the following simplified system ( ss ) : @xmath15 where @xmath16 , @xmath17 , @xmath18 and @xmath19 .
we proceed in two steps : * first we show that if @xmath20 is a solution of the system @xmath14 then @xmath21 . indeed , from the equations ( [ ev - nimplicit ] ) , ( [ ev - ar1 ] ) and ( [ ev - ar2 ] )
, we have : @xmath22 then by multiplying ( standard adding ) the terms of these inequalities , we obtain @xmath23 , since @xmath24 . * we see that if @xmath25 the systems @xmath14 and @xmath26 coincide . for @xmath27 and @xmath28 fixed ,
we denote by @xmath29 and @xmath30 the corresponding systems . by induction on @xmath27 and @xmath28
, we suppose that @xmath31 , and we show that @xmath32 and @xmath33 . * * to show that @xmath32 , we eliminate the variable @xmath34 in @xmath35 which gives a system @xmath29 , then we use the induction assumption . + indeed the problem @xmath35 is written as follows : @xmath36 + using the expression of @xmath34 in ( [ i - eq ] ) , we replace it in the expression of @xmath37 in ( [ i - nimplicit ] ) .
we obtain : @xmath38,\ ] ] which gives : @xmath39 because @xmath40 since @xmath41 and @xmath42 .
+ using the expression of @xmath34 in ( [ i - eq ] ) , we replace it in the expression of @xmath43 in ( [ i - eq ] ) also .
we obtain : @xmath44,\ ] ] which gives : @xmath45 because @xmath46 , since @xmath47 and @xmath48 .
+ let us denote by @xmath49 and @xmath50 the following sets of indexes : @xmath51 thus we can conclude the following equivalence : @xmath52.\ ] ] + the equations @xmath53 combined with the equations ( [ eqnm1 ] ) , ( [ eqnp1 ] ) , ( [ i - implicit ] ) , ( [ i - ar1 ] ) and ( [ i - ar2 ] ) form the following @xmath29 system of variables @xmath54 and @xmath55 : @xmath56 by using the induction assumption , this later system is equivalent to the following system : @xmath57 by adding the equation @xmath58 to the later system , we obtain @xmath59 . *
* we show with the same manipulations that @xmath33 @xmath60 [ thmp](baccelli et al .
@xcite ) given @xmath61 a @xmath62 minplus matrix , if the weights of all the circuits of the graph @xmath63 associated to @xmath61 are positive , then the equation @xmath64 admits a unique solution @xmath65 , where @xmath66 solving the system @xmath14 is equivalent to solve the following system of four variables : @xmath67 taking into account the equivalence @xmath68 , suppose that @xmath69 , @xmath70 , @xmath34 , @xmath37 , @xmath71 are known . to determine the other variables i.e. @xmath72 , @xmath73 we have to solve the system of equation ( [ ev - eq ] )
this system is written : @xmath74 where @xmath75 @xmath76 with @xmath77 and @xmath78 and @xmath79.\ ] ] using lemma [ thmp ] , all the circuits of the graph associated to the matrix @xmath61 have the average weight @xmath80 which is positive because @xmath81 .
thus the matrix @xmath82 exists and the solution of the system ( [ eq - aff ] ) is given by : @xmath65 @xmath60
let us use the notations : * @xmath84 * @xmath85 [ eigen ] there exists a solution @xmath20 of @xmath86 such that @xmath69 satisfies : @xmath87 before we give the proof of theorem [ eigen ] , let us explain it . using the notations : * @xmath88=(1+\rho)(1/4)\;,$ ] * @xmath89=(2r+1-\rho)/4\;,$ ] the result can be explained as follows ( see figure [ diag ] ) : * if @xmath90 then @xmath83 admits a solution @xmath20 such that : @xmath91 , * if @xmath92 the @xmath83 admits a solution @xmath20 such that : @xmath93 , * if @xmath94 or @xmath95 which cases correspond respectively to @xmath96 or @xmath97 then @xmath83 admits a solution @xmath20 such that : @xmath98 , * if @xmath99 then @xmath83 admits a solution @xmath20 such that : @xmath100 . *
if @xmath101 , then a solution @xmath102 is given by : @xmath103 @xmath104 which is a solution of : @xmath105 indeed : + + @xmath106{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_n & = & 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^4 , \quad \text { because } \;\;\bar{a}_n=1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(a_n a_{n+m}),\\ \\ & \geq & e , \;\ ; \text { because } \;\;\lambda=\frac{n+m-1}{n+m}\;d\leq\frac{n+m-1}{n+m}\ ; 1/4\leq 1/4 .
\end{array}$ ] + + + @xmath107{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+m } & = & 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^3 , \quad \text { because } \;\;\bar{a}_{n+m}=1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(a_n a_{n+m}),\\ \\ & \geq & e , \;\ ; \text { because } \;\;\lambda\leq 1/4<1/3 .
\end{array}$ ] + + + @xmath108{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_1 & = & ( 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^2)^{n-1 } , \quad \text { because }
\;\;\bar{b}_n b_n=1^{n-1},\\ \\ & \geq & e , \;\ ; \text { because } \;\;\lambda\leq 1/4<1/2 .
\end{array}$ ] + + + @xmath109{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+1 } & = & 1^{m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(b_nb_ma_na_{n+m})\lambda^{n - m+2 } , \quad \text { because } \;\;\bar{b}_m=1^{m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_m,\\ \\ & = & ( 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^2)^{m-1 } , \quad \text { because } b_nb_ma_na_{n+m}=d^{n+m-1}=\lambda^{n+m},\\ \\ & \geq & e , \;\ ; \text { because } \;\;\lambda\leq 1/4<1/2 .
\end{array}$ ] + + + * if @xmath92 then a solution @xmath102 is given by : @xmath110 @xmath111 which is a solution of : @xmath112 indeed : + + @xmath113{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_n & = & ( a_na_{n+m}b_nb_m){\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^{n+m-4 } , \quad \text { because } \;\;\bar{a}_n=1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(a_n a_{n+m}),\\ \\ & = & d^{n+m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(1/4)^{n+m-4 } , \\ \\ & & \text { because } \;\;a_na_{n+m}b_nb_m = d^{n+m-1},\quad\text{and}\quad \lambda=1/4,\\ \\ & \geq & \frac{n+m}{4}-1-\frac{n+m-4}{4}=e , \quad \text { because } \;\;d\geq d_1\;. \end{array}$ ] + + + @xmath107{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+m } & = & \lambda = 1/4 \geq e. \end{array}$ ] + + + @xmath108{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_1 & = & 1^n{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(a_na_{n+m}b_nb_m){\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^{m - n-2},\\ \\ & & \text { because } \;\;\bar{b}_n=1^{n-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_n , \quad\text{and}\quad \bar{a}_n=1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(a_n a_{n+m}),\\ \\ & = & 1^n{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}d^{n+m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^{n - m+2 } , \quad \text { because } \;\;a_na_{n+m}b_nb_m = d^{n+m-1},\\ \\ & \geq & n-\frac{3n+m-2}{4}-\frac{n - m+2}{4 } = e , \quad \text { because } \;\;d\leq d_2 .
\end{array}$ ] + @xmath109{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+1 } & = & ( 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^2)^{m-1 } , \quad \text {
because } \;\;\bar{b}_m b_m=1^{m-1},\\ \\ & = & \frac{m-1}{2}>e , \quad \text { because } \;\ ; \lambda=1/4 .
\end{array}$ ] * if @xmath94 or @xmath95 , then @xmath114 @xmath115 which is a solution of : @xmath116 indeed , @xmath69 given by ( [ lam3 ] ) satisfies @xmath117 because : 1 .
if @xmath118 , which corresponds to @xmath119 , then @xmath120 , and we can check that : @xmath121 .
2 . if @xmath122 , which corresponds to @xmath123 , then @xmath124 , and we can check that : @xmath125 .
+ then : + + @xmath113{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_n & = & \left[1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^2\right]^{n-1 } , \quad \text { because } \;\;b_n \bar{b}_n=1^{n-1},\\ \\ & \geq & e,\quad \text { because } \;\;\lambda\leq 1/4<1/2 . \end{array}$ ] + + @xmath107{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+m } & = & 1^n\lambda^{m - n-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}a_n{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}a_{n+1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_m{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_n,\\ \\ & & \text { because } \;\;\bar{a}_{n+m}=1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}a_n{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}a_{n+m } \quad \text{and } \quad \bar{b}_n=1^{n-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_n,\\ \\ & = & ( 1^n\lambda^{n - m+2}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}d^{n+m-1})\lambda\;,\\ \\ & & \text { because } \;\;a_n a_{n+m}b_mb_n = d^{n+m-1}\;,\\ \\ & = & \lambda\geq e\ ; , \quad \text{because } \quad \lambda^{n - m+2}=1^n{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}d^{n+m-1}. \end{array}$ ] + + @xmath126{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_1 & = & a_na_{n+m}b_nb_m{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}1^{n-1}\lambda^{n - m-2},\quad \text { because } \;\;\bar{b}_n=1^{n-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_n,\\ \\ & = & ( d^{n+m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}1^n \lambda^{n - m+2 } ) ( 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^4)\;,\ ; \text { because } \;\;a_n a_{n+m}b_mb_n = d^{n+m-1}\;,\\ \\ & = & 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^4 \quad \text{because } \quad \lambda^{n - m+2}=1^n{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}d^{n+m-1},\\ \\ & \geq & e,\quad \text { because } \;\;\lambda\leq 1/4 .
\end{array}$ ] + + @xmath109{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+1 } & = & 1^{m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^{2m-2},\quad \text { because } \;\;b_m\bar{b}_m=1^{m-1}\;\;,\\ \\ & = & ( 1{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}\lambda^2)^{m-1},\\ \\ & \geq & e,\quad \text { because } \;\;\lambda\leq 1/4<1/2 .
\end{array}$ ] * if @xmath99 , then @xmath127 + @xmath128 which is a solution of : @xmath129 indeed ; + + @xmath113{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_n & = & 1^{n-1}\;\geq e\;,\quad \text { because } \;\;b_n\bar{b}_n=1^{n-1 } , \end{array}$ ] + + @xmath130{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+m } & = & a_na_{n+m}b_nb_m{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}1^n\;,\\ \\ & = & d^{n+m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}1^n\ ; , \quad \text { because } \;\ ; a_na_{n+m}b_nb_m = d^{n+m-1},\\ \\ & \geq & e,\quad \text { because } \;\;d\geq r. \end{array}$ ] + + @xmath131{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_1 & = & 1\;\geq e\ ; , \end{array}$ ] + + @xmath109{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}x_{n+1 } & = & 1^{n+m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}(a_na_{n+m}b_nb_m)\;,\quad \text { because } \;\;\bar{b}_m=1^{m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}b_m,\\ \\ & = & 1^{n+m-1}{\text{$/$\hspace{-1.42mm}\tiny${}^\circ $ \normalsize}}d^{n+m-1}\;,\quad \text { because } \;\ ; a_na_{n+m}b_nb_m = d^{n+m-1},\\ \\ & \geq & e\ ; , \quad \text { because } d\leq 1\text{{\scriptsize ~$\blacksquare$ \normalsize } } \end{array}$ ] [ fond - cor1 ] in the case @xmath132 , a non negative eigenvalue @xmath69 of @xmath83 is given by : @xmath133 follows directly from theorem [ eigen ] @xmath60 [ fond - cor2 ] for large values of @xmath27 and @xmath28 such that @xmath134 ( which is the case @xmath132 ) , a non negative eigenvalue @xmath69 of ( s ) is given by : @xmath135 follows directly from corollary [ fond - cor2 ] @xmath60 [ rem1 ] we can check that as soon as we assume @xmath136 , we get @xmath137 , so we have : @xmath138 .
the position of @xmath1 with respect to @xmath139 and @xmath140 gives three cases and divides the interval @xmath141 $ ] , in each case in four regions , where the eigenvalues @xmath69 of @xmath83 satisfy ( see figure [ diag ] ) : * @xmath142 , * @xmath143 , * @xmath144 , * @xmath145 , * @xmath146 , * @xmath147 \quad \rightarrow \quad \lambda=0 $ ] .
given in theorem [ eigen ] depending on @xmath0.,title="fig:",width=170,height=132 ] given in theorem [ eigen ] depending on @xmath0.,title="fig:",width=170,height=132 ] given in theorem [ eigen ] depending on @xmath0.,title="fig:",width=170,height=132 ] [ th - pos ] if @xmath148 then the system @xmath83 ( solved on @xmath20 ) is equivalent to the following system ( solved on @xmath149 ) : @xmath150 from the equations ( [ s - nimplicit ] ) and ( [ s - implicit ] ) we obtain : @xmath151 since @xmath152 we have : @xmath153 .
thus : @xmath154 by replacing @xmath71 in ( [ s - nimplicit ] ) , we obtain : @xmath155 x_1.\ ] ] the system @xmath83 is then equivalent to the system \{([eq-02 ] ) , ( [ eq-01 ] ) , ( [ s - ar1 ] ) , ( [ s - ar2])}. on this later system we use the following changing variable : @xmath156 and we obtain the system @xmath157
@xmath60 if @xmath96 ( that is @xmath158 ) , and for the densities @xmath0 satisfying @xmath159 , the system @xmath83 , and thus the eigenvalue problem @xmath14 , admit a unique positive eigenvalue @xmath69 given by : @xmath160 this situation corresponds to the phases a , c and e of the case 3 on figure [ diag ] of remark [ rem1 ] .
* let @xmath69 be positive .
lemma [ th - pos ] gives the equivalence of the systems @xmath83 and @xmath157 .
the later system is the eigenvalue problem associated to the following dynamical system : @xmath161 if @xmath96 , that is @xmath158 , then this dynamical system is implicit but triangular .
indeed , an iteration of the dynamics is to compute @xmath162 and @xmath163 in parallel , then compute @xmath164 , and finally compute @xmath165 .
so the system @xmath157 can be interpreted as a dynamic programming equation of a stochastic optimal control problem , where @xmath69 is the average optimal cost by unit of time .
since @xmath69 is supposed to be positive , and from corollary [ fond - cor1 ] , we obtain ( [ lamb ] ) .
* let @xmath166 and @xmath167 be two positive eigenvalues of @xmath83 .
lemma [ th - pos ] tells us that both of @xmath168 and @xmath169 are eigenvalues of the system @xmath157 . since in the case
when @xmath170 , the system @xmath157 is a dynamic programming equation of a stochastic optimal control problem , which thus admits a unique eigenvalue , we conclude that @xmath171 @xmath60
the results of this note give a solution to the additive eigenvalue problem associated to a dynamics of an elementary 2d - traffic system ( two circular roads crossing at one junction , managed by the priority - to - the - right - rule ) . the eigenvalue @xmath69 , which is not necessarily unique , is given as a function of two main quantities which are interpreted in terms of traffic as the density @xmath0 of vehicles in the system , and the ratio @xmath1 between the non priority road size and the size of the whole system .
moreover , we showed that when @xmath1 satisfies @xmath96 , that is when the size of the non priority road is bigger than the size of the priority road , the uniqueness of @xmath69 , which is positive in this case , is proved for densities satisfying @xmath172 .
n. farhi , m. goursat , j .- p .
quadrat : _ derivation of the fundamental traffic diagram for two circular roads and a crossing using minplus algebra and petri net modeling _ , in proceedings of the 44th ieee - cdc , sevilla , 2005 .
|
this is a technical note where we solve the additive eigenvalue problem associated to a dynamics of a 2d - traffic system .
the traffic modeling is not explained here .
it is available in @xcite .
it consists of a microscopic road traffic model of two circular roads crossing on one junction managed with the priority - to - the - right rule .
it is based on petri nets and minplus algebra .
one of our objectives in @xcite was to derive the fundamental diagram of 2d - traffic , which is the relation between the density and the flow of vehicles .
the dynamics of this system , derived from a petri net design , is non monotone and additively homogeneous of degree 1 . in this note
, we solve the additive eigenvalue problem associated to this dynamics .
|
an impressive series of experiments has been realized with ultracold atomic gases over the last years , reproducing systems , or physical situations , relevant to several areas of physics .
the use of fano - feshbach resonances to control the interaction strength between particles , in particular , has been the cornerstone of many of these recent experimental achievements . besides providing analog quantum models and simulators for systems of interest in different fields of physics ,
ultracold gases offer also the possibility to construct novel many - body systems with no corresponding counterparts in other domains of physics .
resonant bose - fermi mixtures constitute an interesting example in this respect , and have been the object of active theoretical and experimental investigation recently .
nonresonant bose - fermi mixtures were initially studied theoretically and experimentally in @xcite and @xcite , respectively .
bose - fermi mixtures in the presence of a _ narrow _ fano - feshbach resonance were then considered in the theoretical works @xcite . for a narrow resonance one
has to take into account explicitly the molecular state forming in the closed channel , leading to a hamiltonian which includes three different species ( bosons , fermions , and molecules ) from the outset .
both cases of narrow or broad resonance are , however , relevant to current experiments in bose - fermi mixtures @xcite , depending on the mixture and/or the resonance actually chosen in the experiment .
a broad fano - feshbach resonance is characterized by the smallness of the effective range parameter @xmath2 of the boson - fermion scattering amplitude with respect to both the average interparticle distance and the boson - fermion scattering length @xmath3 @xcite . under these conditions ,
the system can be described by a hamiltonian made just by bosons and fermions mutually interacting via an attractive contact potential .
initial works studying bose - fermi mixtures in the presence of a broad resonance focused on lattice models @xcite , or considered separable interactions , as inspired by nuclear - physics models @xcite . the continuum case in the presence of an attractive contact potential was first tackled in ref .
@xcite , which concentrated on the thermodynamic properties of the condensed phase of a bose - fermi mixture at zero temperature .
a first study of the competition between bose - fermi pairing and boson condensation in a broadly resonant bose - fermi mixture has been presented in a previous work by us @xcite . by using many - body diagrammatic theory , with a @xmath4-matrix approximation for the bosonic and fermionic self - energies
, we were able to show that , for increasing bose - fermi attraction , the boson - fermion pairing correlations progressively reduce the boson condensation temperature and make it eventually vanish at a critical coupling above which the condensate is completely depleted ( thus revealing the presence of a quantum phase transition at zero temperature ) .
in the present paper we extend the work of ref .
@xcite by analyzing the effect of a mass imbalance between the bosonic and fermionic species .
we present results both at finite temperature ( for a specific mass ratio ) and at zero temperature ( for several mass ratios ) .
the effect of mass imbalance in a broadly resonant bose - fermi mixture has been studied recently also in ref .
@xcite , within a path - integral approach which is complementary to our approach and that was limited to zero temperature only . a comparison between our results and those of ref . @xcite will be discussed later on in our paper .
the paper is organized as follows . in section [ formulation ]
we present the theoretical formalism and the main equations describing the physical system of interest in our paper .
section [ fintemp ] reports the numerical results at finite temperature for a @xmath0rb @xmath1k mixture .
the phase diagram for different values of the density imbalance is presented , together with the curves for the boson and fermion chemical potentials at the critical temperature .
section [ qpt ] treats the zero temperature limit .
we report the results for the dependence of the critical coupling on the mass ratio and density imbalance , the corresponding chemical potentials , and the momentum distribution functions .
section [ conclusions ] presents finally our conclusions .
we consider a homogeneous mixture , composed by single - component fermions and bosons .
the boson - fermion interaction is assumed to be tuned by a broad fano - feshbach resonance , as in most of current experiments on bose - fermi mixtures @xcite . under this condition
, the boson - fermion interaction can be adequately described by an attractive point - contact potential .
we consider then the the following ( grand - canonical ) hamiltonian : @xmath5 here @xmath6 , creates a particle of mass @xmath7 and chemical potential @xmath8 at spatial position @xmath9 , where @xmath10=b , f indicates the boson and fermion atomic species , respectively , while @xmath11 is the bare strength of the contact interaction ( we set @xmath12 throughout this paper ) . as for two - component fermi gases @xcite ,
the ultraviolet divergences associated with the contact interaction in ( [ hamiltonian ] ) are eliminated by expressing the bare interaction @xmath11 in terms of the boson - fermion scattering length @xmath3 : @xmath13 where @xmath14 is the reduced mass of the boson - fermion system .
we have not considered in the hamiltonian ( [ hamiltonian ] ) an explicit boson - boson interaction term . in the physical systems relevant to experiments , the corresponding nonresonant scattering length is normally small and positive . in the homogeneous and normal system we are going to consider
, it yields then only a mean - field shift of the boson chemical potential .
note however that , according to the variational analysis of ref .
@xcite , the assumption of having a homogeneous system actually requires the boson - boson scattering length to exceed a certain value , in order to guarantee the mechanical stability of the bose - fermi mixture against collapse .
we thus implicitly assume to work in this stable regime .
note finally that a fermi - fermi @xmath15wave scattering length is excluded by pauli principle . a natural length scale for the system
is provided by the average interparticle distance @xmath16 ( where @xmath17 is the total particle - number density , @xmath18 and @xmath19 being the individual boson and fermion particle - number density , respectively ) .
we thus introduce a fictitious fermi momentum of the system @xmath20 ( as for an equivalent two - component fermi gas with a density equal to the total density of the system ) , and use the dimensionless coupling parameter @xmath21 to describe the strength of the interaction . in the weak - coupling limit , where the scattering length @xmath3 is small and negative and @xmath22 , the two components behave essentially as bose and fermi ideal gases . in the opposite , strong - coupling , limit where @xmath3 is small and positive and @xmath23
, the system is expected to be be described in terms of composite fermions , i.e. boson - fermion pairs with a binding energy @xmath24 , and excess fermions .
[ we restrict our analysis to mixtures where the number of bosons never exceeds the number of fermions .
this is because only in this case a quantum phase transition associated with the disappearance of the condensate is possible @xcite .
in addition , mixtures with @xmath25 are expected to be severely affected by three - atom losses . ]
the physical behavior of a bose - fermi mixture across all the resonance is captured by the @xmath4-matrix set of diagrams for the boson and fermion self - energies represented in fig.[diagrams ] . as shown in our previous work @xcite ,
this set of diagrams is able to recover the expected physical behavior in both the weak- and strong - coupling limits , thus providing a meaningful theoretical framework for the whole resonance .
the corresponding equations for the bosonic and fermionic self - energies @xmath26 and @xmath27 and many - body @xmath28matrix @xmath29 read : @xmath30 @xmath31+b[\xi_{\rm b}({\bf p})]}{\xi_{\rm f}({\bf p}-{\bf p})+\xi_{\rm b}\left({\bf p}\right)-i\omega_{m } } -\frac{2m_{r}}{{\bf p}^{2 } } \right]\right\}^{-1}. \label{gamma}\end{aligned}\ ] ] here @xmath32 , @xmath33 , @xmath34 , where @xmath35 and @xmath36 , @xmath37 are bosonic and fermionic matsubara frequencies , respectively , ( @xmath38 being integer numbers ) , while @xmath39 and @xmath40 are the fermi and bose distribution functions at temperature @xmath4 , and @xmath41 .
the self - energies ( [ selfb ] ) and ( [ selff ] ) determine the dressed green s functions @xmath42 via the dyson s equation @xmath43 ( with the bare green s functions given by @xmath44^{-1}$ ] and @xmath45^{-1}$ ] ) .
the dressed green s functions @xmath42 allow to calculate the boson and fermion momentum distribution functions through the equations : @xmath46 which in turn determine the boson and fermion number densities : @xmath47 at fixed densities and temperature , the two coupled equations ( [ nb ] ) and ( [ nf ] ) fully determine the boson and fermion chemical potentials , and therefore the thermodynamic properties of the bose - fermi mixture in the normal phase .
coming from the normal phase , the condensation of bosons starts when the condition @xmath48 is first met .
( [ eqtc ] ) corresponds to the requirement of a divergent occupancy of the boson momentum distribution at zero momentum : @xmath49 .
@xcite we pass now to consider the zero - temperature limit of the previous finite - temperature formalism .
when @xmath50 , the spacing @xmath51 between two consecutive matsubara frequencies tends to zero .
the sums over discrete matsubara frequencies can then be replaced by integrals over continuous frequencies , provided the corresponding integrands are not too singular @xcite .
we have then the equations @xmath52 for the bosonic and fermionic self - energies , where @xmath53 , @xmath54 , @xmath55 while the frequencies @xmath56 , @xmath57 and @xmath58 are now continuous variables .
similarly , the equations ( [ nbq ] ) and ( [ nfk ] ) for the momentum distribution functions are changed to @xmath59 from which the number densities can be calculated as before .
a closed - form expression can be finally derived for the many - body t - matrix @xmath60 at zero temperature , when the fermi and bose distribution functions appearing in eq .
( [ gamma ] ) are replaced by step functions .
we report in particular the expression for @xmath60 when @xmath61 and @xmath62 , which is in practice the only one relevant to our calculations at zero temperature .
we have @xmath63^{-1 } , \label{gammazero}\ ] ] where we have defined @xmath64 and @xmath65 , while @xmath66 , which results from the integration of the term with the fermi function in eq .
( [ gamma ] ) , is given by @xmath67 \nonumber\\ & - & \frac{m_r k_\omega}{4\pi^2}\left\ { \ln\left[\frac{(k_{\mu_{\rm f}}+k_\omega)^2-k_{{\bf p}}^2}{k_{{\bf p}}^2-(k_{\mu_{\rm f}}-k_\omega)^2}\right]-i \pi { \rm sgn}(\omega)\right\}\nonumber\\ & + & \frac{m_r k_{\mu_{\rm f}}}{2\pi^2},\label{if}\end{aligned}\ ] ] where @xmath68 , @xmath69 , while @xmath70 equations ( [ selfbzero])-([if ] ) determine the thermodynamic properties of a bose - fermi mixture at zero temperature in the absence of boson condensation .
they are thus relevant for a sufficiently strong coupling @xmath71 , such that the system remains in the normal phase even at zero temperature .
in particular , upon lowering the coupling constant @xmath71 , the condensation will start at a critical coupling @xmath72 , when the condition ( [ eqtc ] ) is first satisfied .
the theoretical approach developed in section [ formulation ] , can be used to explore the normal phase of a homogeneous bose - fermi mixture at arbitrary values of the boson and fermion masses and densities . in this section , where we present finite - temperature results
, we focus on the specific mass ratio @xmath73 , relevant for the @xmath0rb @xmath1k mixture . more general mass ratios will be considered at zero temperature .
figure [ tc_8740 ] presents the dependence of the condensation critical temperature on the boson - fermion coupling @xmath74 for a @xmath0rb @xmath1k mixture , at different values of the density imbalance @xmath75 .
the critical temperature was obtained by solving numerically eqs .
( [ selfb])-([nf ] ) , supplemented by the condition ( [ eqtc ] ) , while the ending point at @xmath76 was calculated independently by solving the equations ( [ selfbzero])- ( [ if ] ) ( with the condition ( [ eqtc ] ) defining now the critical coupling @xmath72 at zero temperature ) .
the matching between finite - temperature and zero - temperature results confirms the validity of the equations derived in the zero - temperature limit , while providing simultaneously a check of the numerical calculations .
the overall behavior of the critical temperature as a function of coupling is similar to what found for equal masses @xcite .
the critical temperature starts from the noninteracting value ( @xmath77 ) in the weak - coupling limit and eventually vanishes at a critical coupling @xmath72 , when the effect of the boson - fermion coupling is so strong that the bose - einstein condensate is completely depleted , in favor of the formation of molecules .
the weak dependence of the critical coupling on the density imbalance previously found for equal masses is confirmed also for this case with different masses . in this case , the critical coupling varies in the range @xmath78 , to be compared with the range @xmath79 found for @xmath80 .
a minor difference with respect to the case with equal masses is finally the presence of a weak maximum in the critical temperature , which reaches a value slightly above the noninteracting value @xmath81 before its final decrease .
this feature is more pronounced for intermediate values of the density imbalance .
we attribute this feature to the delicate balance between the effects of the boson - fermion interaction on the boson dispersion , and the condensate depletion due to molecular correlations .
the boson dispersion may in fact be hardened by the interaction ( similarly to what one finds for a dilute repulsive bose gas @xcite ) , thus leading to an increase of the critical temperature .
the predominance of one effect over the other one depends on a fine tuning of the mass ratio and density imbalance , and may explain the different behavior found for different values of these parameters .
figure [ mutc_8740 ] reports the coupling dependence of the chemical potentials @xmath82 and @xmath83 at the critical temperature for a @xmath0rb @xmath1k mixture at the same values of the density imbalance considered in fig .
[ tc_8740 ] . in this case , the critical boson and fermion chemical potentials present the same qualitative behavior already found for a mixture with equal masses @xcite .
the boson chemical potential decreases markedly with increasing coupling , and changes from @xmath84 for weak coupling to @xmath85 for strong coupling , with a small dependence on the density imbalance .
the fermion chemical potential ( reported in the inset ) remains instead almost constant across the whole resonance .
the decrease of the chemical potential due to the attractive interaction with the bosons is , in fact , partially compensated by the decreasing of the temperature when moving along the critical line ( which increases @xmath83 ) and by the pauli repulsion between unpaired fermions and bose - fermi pairs .
the behavior of the critical temperature as a function of the boson - fermion coupling discussed in the previous section , evidenced the presence of a quantum phase transition at zero temperature associated with a transition between a superfluid phase with a boson condensate to a normal phase , where the condensate is completely depleted . in this section
we examine in more detail this quantum phase transition by solving numerically the equations formulated at exactly zero temperature .
figure [ gc_massb ] reports the critical coupling @xmath72 as a function of the mass ratio @xmath86 , with two distinct panels for the cases @xmath87 and @xmath88 .
the different curves reported in fig . [ gc_massb ] correspond to different values of the density imbalance , ranging from the density - balanced case to the fully imbalanced one @xmath89 , which represents the system with just one boson immersed in a fermi sea .
this is actually the same as a spin - down fermion sorrounded by a fermi sea of spin - up fermions , since for a single particle the statistics is irrelevant .
the critical coupling reduces thus to that for the polaron - to - molecule transition , recently studied in the context of strongly imbalanced two - component fermi gases @xcite .
the equations governing the one boson limit are reported in the appendix .
the critical coupling is strongly influenced by the mass ratio , especially for @xmath90 . in this case , for all values of the density imbalance , @xmath72 increases very rapidily as @xmath86 is decreased . for the single boson problem , an asymptotic expansion of the equations determining @xmath72 for @xmath91 yields the result @xmath92 which proves quite accurate
even before the asympotic regime is reached ( the deviation from the numerical solution is 15% for @xmath93 and 2% for @xmath94 ) .
[ see the appendix for the details of the derivation of eq .
( [ asysmall ] ) . ] the rapid increase of the critical coupling for the polaron - to - molecule transition when @xmath95 was already noticed in ref .
, even though no asymptotic expression was reported there .
note however that , according to the analysis of the polaron - to - molecule transition of ref .
@xcite , for a mass ratio @xmath96 the molecular state acquires a finite momentum in its ground state .
similar results were obtained in ref .
the equations for the single boson problem here adopted assume that the formation of the molecule occurs at zero center of mass momentum ( as in refs . ) .
the curve corresponding to the single boson problem in fig . [ gc_massb ] may thus change for @xmath97 , after taking into account the possibility of pairing at finite momentum .
we expect however this change to be minor on the basis of our calculations with a finite boson density , which allow for pairing at finite momentum and yield results close to the single - boson curve also for @xmath98 .
note further that the study of the three body system with two equal fermions with mass @xmath99 and one different particle with mass @xmath100 interacting through a zero - range potential , shows that for @xmath101 the system is unstable due to a sequence of three - body bound states with energy @xmath102.@xcite a similar instability is expected to occur also in the many - body system with @xmath103 equal fermions plus one different particle ( a recent work has proven indeed that the above critical value for the three - body system provides a lower bound for the location of the instability in the many - body system @xcite ) .
we note , however , that the unboundness from below of the efimov spectrum ( and the associated global instability ) occurs only for a pure zero - range interaction . in a real system
, physical two - body interactions will provide a natural cut - off at distances of the order of the van - der - waals length @xmath104 , thus limiting the position of the lowest efimov level at an energy @xmath105 .
the global mechanical instability is thus avoided by the real system , even though the presence of efimov states is expected to lead to an enhancement of three - body losses when they lie close to the three - particle or atom - dimer continuum .
we further note in this context that the presence of the ( non - universal ) three - body bound - states recently found by kartavtsev and malykh @xcite for @xmath106 could also lead to enhanced three - body losses in this mass - ratio range .
we observe in any case that since our calculations were taken for @xmath107 ( @xmath108 for equal densities , due to numerical difficulties ) , the above three - body effects ( which are out of the scope of the present theory ) should affect our study only marginally .
consistently with our previous results , we observe a weak dependence of @xmath72 on the density imbalance .
such a weak dependence on the densities remains valid also for @xmath109 . in this case , all curves reach a minimum value of @xmath72 ( = [email protected] ) for mass ratios @xmath111 in the range 3.5@xmath1105 , after which they increase slowly with the mass ratio . for the single boson problem with a large mass we have obtained the asymptotic expression ( see appendix ) : @xmath112 where @xmath113 one can see from eqs .
( [ asylarge1 ] ) and ( [ asylarge2 ] ) that at large mass ratios @xmath72 increases very slowly , with a logarithmic dependence on the mass ratios . due to this log - dependence the leading behavior , @xmath114 , is reached only at extremely large mass ratios .
equation ( [ asylarge1 ] ) , which includes the first two corrections to the leading behavior , provides a better approximation , the deviation from the numerical solution being 15% for @xmath115 and 1.5% for @xmath116 .
the increasing behavior of @xmath72 at both small and large mass ratios implies the existence of a minimum in the curve for @xmath72 at intermediate mass ratios , consistently with the results of fig .
[ gc_massb ] .
note however that in earlier work for the polaron - molecule transition , the critical coupling for the transition was found to move away from the bec limit for increasingly heavier impurities . in particular , ref .
@xcite reported that the critical coupling should approach the unitary limit for an infinitely heavy impurity .
similar results were also found in two - dimensions by m. parish @xcite ( with the critical coupling approaching the weak - coupling limit in this case ) .
a reason for such a difference may be that our theory in the limit of a single boson reduces to the `` first level approximation '' of ref .
@xcite , with no particle - hole dressing of the molecule ( while the polaron is described with the same accuracy obtained with the variational wave - function introduced by chevy @xcite , which is deemed quite accurate for the polaron energy @xcite ) .
the `` second level approximation '' of ref .
@xcite includes instead a particle - hole dressing of the molecule .
this inclusion is sufficient to recover in the strong - coupling limit the correct dimer - fermion scattering length , and shifts the position of the critical coupling for equal masses from the value @xmath117 to the value @xmath118 ( the factor @xmath119 appears here because of a different definition of @xmath120 between us and the above references ) .
the eventual ( slow ) increase of @xmath72 at large mass ratios could then be an artifact of the `` first order approximation '' .
calculations at large mass ratios with alternative methods ( such as fixed - node or diagrammatic quantum monte - carlo methods ) could definitively clarify this issue . at equal densities , our results for @xmath72 as a function of the mass ratio agree well with the results reported in ref .
@xcite for the same case .
this is because the equations used in ref .
@xcite for calculating @xmath72 correspond , in a diagrammatic formalism , to the same choice of the self - energy as ours , but with the dyson s equation expanded : @xmath121 , instead of @xmath122 . even though such an expansion is justified
only when @xmath123 is small , apparently it leads only to minor differences in the values for @xmath72 .
for instance , for equal masses and densities we obtained @xmath124 , to be compared with @xmath125 in ref .
@xcite . according to the analysis of ref .
@xcite , however , only for sufficiently large values of the boson - boson scattering length @xmath126 , the critical coupling @xmath72 lies in a stable region of the phase diagram . for small values of @xmath126 ,
the second - order quantum phase transition between a condensed phase and a normal phase is in fact superseded by a phase separation between the two phases .
according to our calculations , the compressibility matrix @xmath127 is always positive in the normal phase , indicating that the second order phase transition here explored lies , at worst , in a metastable region of the phase diagram . in order to examine its absolute stability within our approach
, one should extend our study to the superfluid phase and make a comparison of the free energies for the normal and superfluid phases .
this nontrivial extension is postponed to future work .
we observe however , that if ratios @xmath128 of the order of 0.2 - 0.3 are sufficient to suppress phase separation in most of the phase diagram ( as the results of ref .
@xcite seem to indicate ) , then the effect of @xmath126 on @xmath72 will be minor . the weak dependence of the critical coupling on the density imbalance is emphasized in fig . [ gc_imbal ] , which presents @xmath72 as a function of the density imbalance for some representative values of the mass ratio @xmath86 .
all curves show a weak dependence on the density imbalance , with a weak maximum at an intermediate imbalance ( = 0.55 for equal masses , and similarly for the other mass ratios considered here ) .
figure [ chem_potb ] reports the chemical potentials at @xmath72 as a function of the mass ratio @xmath86 , for different density imbalances . as it can be seen from the main panel , the ( negative ) bosonic chemical potential increases very rapidly in absolute value for @xmath91 .
this is because at @xmath72 the bosonic chemical potential is already close to its strong - coupling limit , @xmath85 , such that the dimensionless ratio @xmath129 . since @xmath130 for @xmath131
, we have @xmath132 in this limit . in the opposite limit of large @xmath86 ,
the ratio @xmath133 slowly increases and eventually saturates to 1 for large @xmath100 , such that @xmath134 follows the slow logarithmic increase of @xmath72 in this limit .
the fermionic chemical potential ( reported in the inset ) depends weakly on the mass imbalance , reflecting the weak dependence on @xmath72 . as a matter of fact
, the fermion chemical potential is determined essentially by the fermion density @xmath19 , independently from the coupling value or mass ratio .
we pass now to study the momentum distribution functions @xmath135 and @xmath136 for the bosons and fermions , as obtained from eqs .
( [ nbzero ] ) and ( [ nfzero ] ) , respectively .
we present results for a density imbalance @xmath137 , as to emphasize an interesting behavior of the bosonic momentum distribution function , which occurs only for a sufficiently large imbalance .
one can see , indeed , from fig .
[ mom_distrb ] that the bosonic momentum distribution function vanishes identically at low momenta .
this empty region extends from @xmath138 up to a certain value @xmath139 , which is determined essentially only by the density imbalance ( for the specific case of fig .
[ mom_distrb ] , @xmath140 ) .
the presence of the empty region can be interpreted as follows .
for @xmath141 , most of fermions remain unpaired and fill a fermi sphere of radius @xmath142^{1/3}$ ] , as fig .
[ mom_distrf ] for the fermionic distribution clearly shows . at @xmath72 and
larger couplings , the bosons are instead bound into molecules that , being composite fermions , fill a fermi sphere with a radius @xmath143 .
now , as the region @xmath144 is already occupied by the unpaired fermions , only fermions with @xmath145 participate to the molecule . since the momentum of the molecule is given by the sum @xmath146
, the constraints @xmath147 and @xmath145 then imply that only bosons with @xmath148 participate to the molecule , leaving thus empty the region @xmath149 .
we have verified that the equation @xmath150 reproduces rather accurately the values of @xmath151 obtained numerically .
in particular , the empty region does not exist at small density imbalance , when @xmath152 .
note also that the initial rise of @xmath135 after the threshold @xmath151 is due to the progressive increase of the phase - space volume corresponding to the @xmath153 s satisfying the above constraints at a given @xmath154 . the saturation volume in phase - space
is reached for @xmath155 of the order of @xmath156 , after which @xmath135 starts to decrease , following eventually at sufficiently large wave - vectors a molecular - like internal wave - function @xmath157 , where @xmath158 can be approximated by the two - body normalized wave - function @xmath159 , while the coefficient @xmath160 can be interpreted as the molecular density . the above approximate expression for @xmath161 accounts for
the main difference in the curves calculated at different mass ratios ( namely , the decreasing height of the curves when the mass ratio is lowered ) due to the strong dependence of @xmath72 , and then of @xmath3 , on the mass ratio . note further that the same kind of beaviour is seen in the fermionic distribution function at momenta @xmath145 , as it can be evinced from fig .
[ mom_distrf ] ( b ) .
the comparison between fig .
[ mom_distrf ] ( b ) and fig .
[ mom_distrb ] shows indeed that the bosonic and fermionic distribution functions become identical as the momentum increases ; at large momenta @xmath162 , consistently with the universal large momenta behavior established in ref .
in particular , within our approximation , one can prove by taking the large @xmath163 limit in our expressions that the `` contact '' constant @xmath164 is given by @xmath165 . in the strong - coupling limit , where all bosons are bound into molecules ,
a comparison with the expression for the molecular internal wave - function then leads to the equation @xmath166 .
figure [ contact ] reports the contact constant @xmath164 at @xmath72 normalized to its strong - coupling limit value @xmath167 , as a function of the mass ratio @xmath86 for different values of the density imbalance .
one can see that the constant @xmath164 at @xmath72 is close to its strong - coupling limit expression for all cases considered , with the largest deviations occuring at intermediate values of the mass ratio , as expected , since this is the region where @xmath72 reaches its minimum and consequently the strong - coupling condition is less respected .
in this paper , we have studied a resonant bose - fermi mixture in the presence of a mass and/or density imbalance .
we have first analyzed the finite temperature phase diagram for the specific case of a @xmath0rb @xmath1k mixture .
we have found that the overall shape of the phase diagram is similar to what found previously for equal masses .
the critical temperature starts from the noninteracting value ( @xmath77 ) in the weak - coupling limit and eventually vanishes at a critical coupling @xmath72 , when the effect of the boson - fermion coupling is so strong that the bose - einstein condensate is completely depleted .
the critical temperature presents a weak maximum at intermediate coupling , not found for equal masses .
we have explained this feature as resulting from the balance between the hardening of the boson dispersion due to the boson - fermion interaction , and the condensate depletion due to molecular correlation .
the predominance of one effect over the other one depends on a fine tuning of the physical parameters , thus explaining the different behavior found for different cases .
we have then considered the zero - temperature limit .
we have found that the critical coupling @xmath72 is significally affected by the mass imbalance , in particular for @xmath90 . in this case , for all values of the density imbalance , we have obtained a rapid increase of @xmath72 as @xmath91 .
we have also found that the density imbalance influences only weakly the critical coupling , as previously found for equal masses @xcite . on the other hand ,
a sufficiently large density imbalance produces quite a remarkable effect on the bosonic momentum distribution function .
we have found indeed that this function is completely depleted for momenta with magnitude below a certain value @xmath151 .
this value is set by the difference between the fermi momenta associated with the unpaired fermions and the composite - fermions , respectively .
such a momentum distribution function reflects the constraints on the internal molecular wave - function imposed by the presence of a large fermi surface associated with the unpaired fermions . in this respect ,
the physics underlying this phenomenon is similar to that of the `` sarma phase '' for a highly - polarized fermi gas in the strong - coupling limit of the bcs - bec crossover @xcite . in particular , the bosonic momentum distribution function is similar to the momentum distribution function of the minority species in a polarized fermi gas .
interestingly , the fact that in the present case the molecules can not condense , but rather fill a fermi sphere , smears some features of the momentum distribution function , but still preserves the presence of an empty region in the momentum distribution function .
such an interesting effect could be measured in a bose - fermi mixture in the molecular limit by using the same technique used to measure the momentum distribution of an ultracold fermi gas in the bcs - bec crossover @xcite .
partial support by the italian mur under contract cofin-2009 `` gas quantistici fuori equilibrio '' is acknowledged .
in this appendix we consider the `` one boson '' limit @xmath168 of our equations . in this limit
the fermions become free due to the vanishing boson density .
one has then @xmath169 and @xmath170 .
the boson self - energy @xmath26 remains instead finite and is determined by eq .
( [ selfbzero ] ) , with no simplifications with respect to the case at finite boson density .
the full knowledge of @xmath26 is however not necessary to calculate @xmath82 in the limit @xmath171 .
a study of the analytic structure of @xmath26 in the complex frequency space along the lines of ref .
shows that the limit @xmath171 corresponds , for @xmath172 , to the requirement that the minimum of the composite fermion dispersion @xmath173 occurs exactly at zero frequency . by introducing the retarded composite - fermion propagator via the replacement @xmath174
, the dispersion @xmath173 is determined by the pole of @xmath175 in the complex plane . by assuming the minimum of @xmath176 to occur at @xmath177 , one has then the equation @xmath178 , which determines @xmath82 at a given coupling and mass ratio .
the critical coupling @xmath72 , on the other hand , is determined by the equation @xmath179 , as for finite density . in the limit @xmath171
only the pole of the fermionic green s function contributes to the frequency integral in eq .
( [ selfbzero ] ) , yielding : @xmath180\gamma^{\rm r}(\textbf{p } , \xi_{\rm f}(\textbf{p})),\ ] ] where @xmath181^{-1}\end{aligned}\ ] ] and @xmath182}{\frac{{\bf p}^2}{2m_r}-\frac{m_{\rm b}}{m_{\rm f } } \frac{\textbf{p}^{2}}{2m}+ \vert\mu_{\rm b}\vert}.\ ] ] a calculation of the above integral yields where @xmath184^{1/2}$ ] , while @xmath185,\ ] ] where @xmath186 and @xmath187 . by using the expressions ( [ gammaoneb ] ) to ( [ lr ] ) , the self - energy @xmath188 can then be calculated easily by a simple one - dimensional integral over @xmath189 . at @xmath72 the equation @xmath179
then yields : @xmath190 the simultaneous solution of eqs .
( [ mub ] ) and ( [ pole ] ) allows to obtain @xmath72 and @xmath82 for a given mass ratio .
note finally that , even though we have derived eqs .
( [ mub ] ) and ( [ pole ] ) as a limiting case of our equations for a bose - fermi mixture , they describe also the polaron - to - molecule transition in a fermi - fermi mixture , since in this limit the statistics of the minority species becomes immaterial .
we have verified , indeed , that our results for @xmath72 in this limit agree with the results reported in ref .
@xcite for the `` first - level approximation '' .
[ [ asymptotic - expressions - for - g_c - at - small - and - large - mass - ratios ] ] asymptotic expressions for @xmath72 at small and large mass ratios ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ we conclude this appendix by presenting the derivation of the asymptotic expressions ( [ asysmall ] ) and ( [ asylarge1 ] ) .
we assume @xmath82 to be large and negative as it occurs when @xmath71 is large .
the validity of this assumption is verified by the asymptotic expressions for @xmath72 that are obtained accordingly .
the chemical potential @xmath65 is then also large and negative . by expanding eq .
( [ pole ] ) in powers of @xmath193 one obtains @xmath194 or , equivalently @xmath195 note that the subleading term @xmath196 in eq .
( [ mu ] ) describes the mean - field repulsion between the molecule forming in the strong - coupling limit and the fermions , as it can be seen by casting it in the form @xmath197 , where @xmath198 is the reduced mass of a molecule and one fermion , while @xmath199 is the molecule - fermion scattering length within the born approximation @xcite .
the large and negative value of @xmath82 then implies @xmath200 , as it can be seen more easily directly from eq .
( [ ifapp ] ) .
the expansion of ( [ gammaoneb ] ) for @xmath201 large then yields @xmath202 where we have used eq .
( [ mu ] ) , and @xmath203 . by using eq .
( [ selfbzerozero ] ) we obtain @xmath204\nonumber\\ = \frac{k_{\mu_{\rm f}}}{\alpha \pi m_r^2 a } \left[1+\frac{\sqrt{1+\frac{m_{\rm f}}{m_{\rm b}}}}{2 } \ln\frac{\sqrt{1+\frac{m_{\rm f}}{m_{\rm b}}}-1}{\sqrt{1+\frac{m_{\rm f}}{m_{\rm b}}}+1}\right]\phantom{aaaaaa}\label{sigmab1}\end{aligned}\ ] ] by expanding the expression ( [ sigmab1 ] ) for a _ small _ mass ratio @xmath205 , we obtain @xmath206\nonumber\\ & \approx&-\frac{2^{4/3}}{3\pi } \frac{k_{\rm f}}{a m_{\rm b}}\frac{m_{\rm f}}{m_{\rm b}}\left(1 + \frac{8}{5}\frac{m_{\rm b}}{m_{\rm f}}\right)\end{aligned}\ ] ] the equation @xmath179 then yields @xmath207 where we have kept only the leading term in the expression ( [ mubapp ] ) for @xmath82 ( the term @xmath83 would give a correction of order @xmath208 to the asymptotic expression for @xmath72 , with the last term in ( [ mubapp ] ) contributing an even smaller correction ) . by substituting @xmath24 we obtain then @xmath209 which coincides with the expression ( [ asysmall ] ) reported in sec .
iv , and gives a large value of @xmath72 for small mass ratios , consistently with our starting assumption . the asymptotic expression ( [ a3 ] )
is compared with the full numerical calculation of @xmath72 in fig .
[ asymptotic ] ( a ) .
the expansion of eq .
( [ sigmab1 ] ) for a _
large _ mass ratio @xmath210 yields instead @xmath211 where we have disregarded corrections to ( [ sigmablarge ] ) smaller at least by a factor @xmath212 . by equating ( [ sigmablarge ] ) to ( [ mubapp ] ) we get then @xmath213 yielding @xmath214 from which , by multiplying both sides of eq .
( [ a1 ] ) by @xmath215 , we obtain @xmath216 which shows that , to leading order , @xmath217 at large mass ratios .
the slow log - dependence of @xmath72 on @xmath86 makes however significant to keep also a few subleading corrections to the leading behavior .
the solution of eq .
( [ a2 ] ) by iteration yields then @xmath218 ^ 2}\label{a4}\end{aligned}\ ] ] which ignores corrections of order @xmath219 or higher , and coincides with the expression ( [ asylarge1 ] ) reported in sec .
iv . the asymptotic expression ( [ a4 ] )
is compared with the full numerical calculation of @xmath72 in fig .
[ asymptotic ] ( b ) .
l. viverit , c.j .
pethick , and h. smith , phys .
a * 61 * , 053605 ( 2000 ) .
x. x. yi and c. p. sun , phys .
a * 64 * , 043608 ( 2001 ) .
a. p. albus , s. a. gardiner , f. illuminati , and m. wilkens , phys .
a * 65 * , 053607 ( 2002 ) .
l. viverit and s. giorgini , phys .
a * 66 * , 063604 ( 2002 ) .
r. roth and h. feldmeier , phys .
a * 65 * , 021603(r ) ( 2002 ) .
liu , m. modugno , and h. hu , phys .
a * 68 * , 053605 ( 2003 ) .
m. lewenstein , l. santos , m. a. baranov , and h. fehrmann , phys .
lett . * 92 * , 050401 ( 2004 ) .
s. powell , s. sachdev , and h. p. buchler , phys .
b * 72 * , 024534 ( 2005 ) .
d. b. m. dickerscheid , d. van oosten , e. j. tillema , and h. t. c. stoof , phys .
lett . * 94 * , 230404 ( 2005 ) .
a. v. avdeenkov , d. c. e. bortolotti , and j. l. bohn , phys .
a * 74 * , 012709 ( 2006 ) .
d. c. e. bortolotti , a. v. avdeenkov , and j. l. bohn , phys .
a * 78 * , 063612 ( 2008 ) .
see , e.g. , sec .
29 of a. fetter and j. d.walecka , _ quantum theory of many - particle systems _ , mc - graw hill , new york ( 1971 ) and j. m. luttinger and j. c. ward , phys . rev . * 118 * , 1417 ( 1960 ) .
problems in the replacement of the frequency sums with integrals , in the zero - temperature limit , normally arise in the presence of double ( or higher ) poles of the summands in the complex frequency plane .
we have not met these type of singularities in our case . in any case , the good matching of the finite temperature results with those obtained at zero temperature verify _ a posteriori _ the validity of the procedure .
m. veillette _
et al . _ ,
a * 78 * , 033614 ( 2008 ) .
p. massignan , g. m. bruun , and h. t. c. stoof , phys .
a * 78 * , 031602(r ) ( 2008 ) . c. mora and f. chevy , phys .
a * 80 * , 033607 ( 2009 ) .
m. punk , p. t. dumitrescu , and w. zwerger , phys . rev . a * 80 * , 053605 ( 2009 ) .
r. combescot , s. giraud , and x. leyronas , europhys .
lett * 88 * , 60007 ( 2009 ) . c. j. m. mathy , m. m. parish , and d. a. huse , phys .
lett . * 106 * , 166404 ( 2011 ) j .- l . song and f. zhou , phys .
a * 84 * , 013601 ( 2011 ) .
e. braaten , h. w. hammer , phys . rep .
* 428 * , 259 ( 2006 ) .
m. correggi , g. dellantonio , d. finco , a. michelangeli , and a. teta , arxiv:1201.5740 ( 2012 ) . o. i. kartavtsev and a. v. malykh , j. phys .
b * 40 * , 1429 ( 2007 ) .
m. m. parish , phys .
a * 83 * , 051603 ( 2011 ) . f. chevy , phys .
a * 74 * , 063628 ( 2006 ) . s. tan , ann .
phys . * 323 * , 2952 ( 2008 ) ; _ ibidem _ , 2971 ( 2008 ) .
g. sarma , j. phys .
solids * 24 * , 1029 ( 1963 ) .
p. pieri and g.c .
strinati , phys .
* 96 * , 150404 ( 2006 ) .
d. e. sheehy and l. radzihovsky , ann .
322 * , 1790 ( 2007 ) .
l. radzihovsky and d. e. sheehy , rep .
. phys . * 73 * , 076501 ( 2010 ) . f. chevy and c. mora , rep
. prog . phys . *
73 * , 112401 ( 2010 ) .
p. pieri , d. neilson , and g.c .
strinati , phys . rev .
b * 75 * , 113301 ( 2007 ) ; a. l. subasi , p. pieri , g. senatore , and b. tanatar , phys . rev .
b * 81 * , 075436 ( 2010 ) .
|
we consider a homogeneous bose - fermi mixture , with the boson - fermion interaction tuned by a fano - feshbach resonance , in the presence of mass and density imbalance between the two species . by using many - body diagrammatic methods
, we first study the finite - temperature phase diagram for the specific case of the mass - imbalanced mixture @xmath0rb @xmath1k for different values of the density imbalance .
we then analyse the quantum phase transition associated with the disappearance at zero temperature of the boson condensate above a critical boson - fermion coupling .
we find a pronounced dependence of the critical coupling on the mass ratio and a weak dependence on the density imbalance .
for a vanishingly small boson density , we derive , within our approximation , the asymptotic expressions for the critical coupling in the limits of small and large mass ratios .
these expressions are relevant also for the polaron - molecule transition in a fermi mixture at small and large mass ratios .
the analysis of the momentum distribution functions at sufficiently large density imbalances shows an interesting effect in the bosonic momentum distribution due to the simultaneous presence of composite fermions and unpaired fermions .
|
the @xmath2-scheme now becomes the standard renormalization scheme for the definition of the strong coupling constant .
the measured coupling constant @xmath3 in some experiment at relatively high energy is converted to @xmath4 at some representative scale by perturbation theory .
the current world average of such estimates gives @xmath5 .
lattice qcd calculations , on the other hand , have a potential ability to determine the the strong coupling constant from the experimental inputs at low ( hadronic ) energy scale . in order to compare the coupling constant obtained at low energy by the lattice calculations with @xmath4 obtained at high energy ,
the schrdinger functional scheme has been proposed by the alpha collaboration @xcite , and the scheme is shown to be successful . at present , the results on the running coupling constant of two massless flavor qcd are reported @xcite . in the real world
there are three light quarks .
qcd simulations including three dynamical quark effects are thus required to understand the low energy qcd dynamics .
our ultimate goal is to estimate @xmath4 from @xmath6 qcd simulations .
recently , cp - pacs / jlqcd collaborations have started the @xmath6 qcd simulation employing an exact fermion algorithm developed for odd number of quark flavors @xcite .
in particular , noteworthy results have been obtained in @xcite : there exists strong lattice artifacts associated with the phase transition in the @xmath6 qcd simulation with the combination of the plaquette gauge action and o(@xmath0 ) improved wilson quark action , while such lattice artifacts are absent for the renormalization group ( rg ) improved gauge action .
hence , the collaborations have decided to adopt the combination of the rg improved gauge action and o(@xmath0 ) improved wilson quark action for the @xmath6 qcd simulations to obtain @xmath4 . as a first step of our program
, we study a schrdinger functional coupling with improved gauge actions in perturbation theory .
in particular we perturbatively calculate the o(@xmath0 ) boundary improvement coefficients at 1-loop level with improved gauge actions for the pure su(3 ) gauge theory . combining this with the 1-loop results for the fermionic sector @xcite , we determine the o(@xmath0 ) boundary improvement coefficients at 1-loop , which can be used for the dynamical quark simulations in the future .
the rest of this paper is organized as follows . in sect 2 , after a brief reminder of the schrdinger functional scheme and its extension to improved gauge actions , we specify an action used for latter calculations , and discuss the o(@xmath0 ) boundary counterterm . in sect 3
the schrdinger functional coupling constant is defined and formula for a determination of the o(@xmath0 ) boundary improvement coefficients are given .
the 1-loop computation is outlined in sect 4 , and the results of the o(@xmath0 ) boundary improvement coefficients are summarized in sect 5 . our conclusion is given in the last section , together with a discussion on the lattice artifact of the step scaling function .
it has been shown by alpha collaboration that the schrdinger functional ( sf ) scheme is a powerful tool to probe energy evolutions of some physical quantities and to compute improvement coefficients as well as renormalization constants . in the sf scheme
, the theory is defined on a finite box of size @xmath7 with periodic boundary conditions in the spatial directions and dirichlet boundary conditions in the time direction . in the pure su(@xmath8 )
gauge theory with wilson plaquette action @xmath9 $ ] , the partition function in the sf scheme ( in the case that @xmath10 ) is given by @xmath11 e^ { - s[u ] } , \ ] ] where link variables @xmath12 for gauge fields satisfy boundary conditions @xmath13 here @xmath0 is lattice spacing , and @xmath14 , @xmath15 are diagonal traceless matrices , which depend on background field parameters @xmath16 and @xmath17 @xcite .
it is shown@xcite that the minimum of @xmath9 $ ] is given by the lattice background field @xmath18 , where @xmath19 with @xmath20 .
\end{aligned}\ ] ] this background field represents a constant electric field .
an extension of the sf scheme to improved gauge actions was first considered by klassen @xcite .
the transfer matrix construction @xcite was adopted in the discussion . in this formulation ,
each boundary consists of two time slices , to achieved the tree - level o(@xmath21 ) improvement . in this paper
, however , we adopt the formulation proposed by aoki , frezzotti and weisz @xcite , where each boundary consists of only one time slice and the tree - level o(@xmath0 ) improvement is achieved : dynamical variables to be integrated over are independent on the form of the action , plaquette or improved ones , and are given by the spatial link variables @xmath22 with @xmath23 and temporal link variables @xmath24 with @xmath25 on a cylinder with volume @xmath26 .
this formulation is implemented more easily in numerical simulations . the background field in eq.([eqn : v ] )
gives the local minimum of the theory in both cases @xcite and @xcite .
it has not been theoretically proved , however , that eq.([eqn : v ] ) is the absolute minimum for the improved gauge actions . this is checked only numerically @xcite .
our improved action includes plaquette and rectangle loops , and is given by @xmath27 = \frac{1}{g^2_0 } \sum_{{\cal c } \in { \cal s}_0 } w_0({\cal c},g^2_0 ) { 2{\cal l(c)}}+ \frac{1}{g^2_0 } \sum_{{\cal c } \in { \cal s}_1 } w_1({\cal c},g^2_0 ) { 2{\cal l(c ) } } , \label{eqn : impaction}\ ] ] with @xmath28,\ ] ] where @xmath29 is the ordered product of the link variables along loop @xmath30 contained in the set @xmath31(plaquette ) or @xmath32(rectangular ) . @xmath31 and @xmath32 consist of all loops of the given shape which can be drawn on the cylindrical lattice with volume @xmath26 .
the loops involve only the `` dynamical links '' in the sense specified above , and spatial links on the boundaries at @xmath33 and @xmath34 .
in particular , rectangles protruding from the boundary of the cylinder are not included .
one has to choose appropriate boundary weights to achieve the 1-loop level o(@xmath0 ) improvement for observables involving derivative with respect to the boundary . among various choices to achieve this ,
ours is given as follows .
@xmath35 @xmath36 with @xmath37 where coefficients @xmath38 and @xmath39 of the improved gauge action are normalized such that @xmath40
. we call @xmath41 and @xmath42 o(@xmath0 ) boundary improvement coefficients .
so far , 1-loop coefficients @xmath43 and @xmath44 are independent each other .
weight factors @xmath45 , which include loop corrections , becomes the choice b of @xcite in the weak coupling limit .
the choice b achieves the tree - level o(@xmath0 ) improvement and , at the same time , the lattice background field @xmath46 in eq.([eqn : v ] ) satisfies the equation of motion obtained by the variation of dynamical links .
incidentally , we discuss the o(@xmath0 ) boundary counterterm from a different point of view . here
, it is assumed that the plaquette loops and rectangle loops which lie completely on the cylinder @xmath26 are included in the action , and that each boundary consists of one time slice only . as explained in @xcite , at order @xmath0 in the pure gauge theory ,
there are two possible boundary counterterms , @xmath47 and @xmath48 , each of which are summed over the @xmath33 or @xmath34 hyper plane .
since the latter boundary term vanishes in the case of abelian constant boundary field , in the following , we consider only the former . in this case , we have three candidates for o(@xmath0 ) boundary counterterms which respect lattice symmetries , 1 . spatial sum of time - like plaquette loop that just touches one of the boundaries , 2
. spatial sum of rectangle loop that has exactly 2 links on a boundary , 3 . spatial sum of rectangle loop that has exactly 1 link on a boundary , to satisfy one condition , the o(@xmath0 ) improvement condition .
therefore , for simplicity , we can take a trivial weight for the term 3 , and we still have one degree of freedom for the choice of the boundary terms . at the tree level , however , the background field given in eq.([eqn : v ] ) must satisfy the equation of motion , so that one has to take @xmath49 and @xmath50 ( choice b ) . since no such an extra constraint exists for the 1-loop boundary terms , we can freely set a relation between @xmath43 and @xmath44 , which will be given in the next section .
the sf with the improved gauge action is given by @xmath51 e^ { - s[u ] } , \label{eqn : sfimp}\ ] ] where @xmath9=s_{\rm imp}[u]$ ] .
we require the same boundary condition , eq.([eqn : cc ] ) , for the link variables as in the case of the wilson plaquette action . in perturbative calculations ,
there are two main concerns to note : one is whether the background field given by eq .
( [ eqn : v ] ) corresponds to the absolute minimum of the action , and the other is the gauge fixing . for the latter
, we used the covariant gauge fixing procedure outlined in @xcite .
the former statement is positively proved in @xcite for wilson plaquette action .
unfortunately the statement has not been proved yet in the case of the improved gauge actions , since the proof in @xcite is not applicable to these cases . in @xcite , however , it has been numerically checked that the background field given in eq .
( [ eqn : v ] ) corresponds to the minimum for a large class of improved actions , hence we assume this in our perturbative calculations . in a neighborhood of the background field @xmath46
, any link variables @xmath52 can be parametrized by @xmath53 where @xmath54 are quantum fields . the sf coupling is defined through the free energy @xmath55 in eq .
( [ eqn : sfimp ] ) @xmath56 where @xmath57 is the derivative with respect to @xmath16 .
@xmath58 is a normalization constant @xmath59 \right|_{g^2_0=0 } \nonumber \\ & = & \frac{\partial}{\partial \eta } \left .
\left [ c_0 \left\ { \sum_{p_{\rm{t } } } { 2{\cal l(c)}}+ \sum_{p_{\rm{other } } } { 2{\cal l(c)}}\right\ } + c_1 \left\ { \frac{3}{2 } \sum_{r^2_{\rm{t } } } { 2{\cal l(c)}}+ \sum_{r_{\rm{other } } } { 2{\cal l(c)}}\right\ } \right ] \right|_{u = v } \nonumber \\ & = & 12 \left ( \frac{l}{a } \right)^2 \left [ c_0 { \left ( \mbox{\hspace{-1 mm } } \frac { } { } \mbox{\hspace{-1 mm } } \right.}\sin 2 \gamma + \sin \gamma { \left
. \mbox{\hspace{-1 mm } } \frac { } { } \mbox{\hspace{-1 mm } } \right)}+ 4 c_1 { \left ( \mbox{\hspace{-1 mm } } \frac { } { } \mbox{\hspace{-1 mm } } \right.}\sin 4 \gamma + \sin 2 \gamma { \left .
\mbox{\hspace{-1 mm } } \frac { } { } \mbox{\hspace{-1 mm } } \right)}\right ] , \end{aligned}\ ] ] where @xmath60 is given in appendix [ sec : inversepropagator ] .
let us discuss the perturbative expansion of the sf coupling .
if we require that @xmath43 and @xmath44 satisfy ( this is possible by using the last degree of freedom as mentioned in the end of the previous section ) @xmath61 and introduce @xmath62 @xmath63 then one obtains the relation between @xmath64 $ ] and @xmath65 @xmath66 = \left ( \frac{1}{g^2_0 } + \frac{2a}{l } c^{(1)}_{\rm{t } } \right ) \gamma_0 + o(g^2_0 ) .
\label{eqn : sgamma}\ ] ] using the eq.([eqn : sgamma ] ) , the perturbative expansion of @xmath67 is given by @xmath68 with @xmath69 where @xmath70 is the 1-loop correction to the sf coupling , calculated with the tree - level o(@xmath0 ) boundary coefficients , and the detail of the calculation will be given in the next section .
the value of the 1-loop coefficient @xmath71 is determined by the condition that the dominant part of the scaling violation of @xmath72 should be proportional to @xmath73 , and then @xmath43 and @xmath44 are uniquely given by eq.([eqn : relation ] ) and eq.([eqn : ct ] ) .
in the following , we choose lattice unit ( i.e. @xmath74 ) . according to the unpublished note @xcite
, we have used @xmath75 @xmath76 as a basis of lie algebra of su(3 ) in the presence of the background field .
their explicit form can be found in @xcite .
decomposing in a basis @xmath77 @xmath78 the quantum fields @xmath54 are fourier transformed with respect to spatial momenta as @xmath79 where the phase @xmath80 is given in appendix [ sec : inversepropagator ] . in terms of @xmath81 , the quadratic part of the improved gauge action eq.([eqn : impaction ] ) takes the form @xmath82 with the condition @xmath83 the explicit form of the inverse propagator @xmath84 is given in @xcite , and also in appendix [ sec : inversepropagator ] .
the 1-loop correction @xmath85 is given by @xmath86 \right|_{\eta=\nu=0 } , \label{eqn : impm1}\ ] ] where the determinant for the quantum field sector ( the first term in the right hand side of the equation ) is taken with respect to the spatial momentum * p * , the time @xmath87 , the lie algebra sector @xmath0 and lorentz index @xmath88 .
the second term in the right hand of eq.([eqn : impm1 ] ) represents a contribution from the ghost sector . here
we will exclusively consider the quantum field sector , since the contribution from the ghost sector to the 1-loop correction is same as in the case of wilson plaquette action .
our boundary condition for the temporal component @xmath89 is different from that in @xcite , so that the `` non - uniform '' contribution in the gauge fixing term remains in the inverse propagator @xmath90 ( see appendix [ sec : inversepropagator ] ) .
|
we determine o(@xmath0 ) boundary improvement coefficients up to 1-loop level for the schrdinger functional coupling with improved gauge actions including plaquette and rectangle loops .
these coefficients are required to implement 1-loop o(@xmath0 ) improvement in full qcd simulations for the coupling with the improved gauge actions . to this order ,
lattice artifacts of step scaling function for each improved gauge action are also investigated .
in addition , passing through the sf scheme , we estimate the ratio of @xmath1-parameters between the improved gauge actions and the plaquette action more accurately .
|
the accretion induced collapse of a rapidly rotating white dwarf can result in the formation of a rapidly and differentially rotating compact object .
it has been suggested that such rapidly rotating objects could emit a substantial amount of gravitational radiation @xcite , which might be observable by the gravitational wave observatories such as ligo , virgo and geo .
it has been demonstrated that if the collapse is axisymmetric , the energy emitted by gravitational waves is rather small @xcite .
however , if the collapsed object rotates rapidly enough to develop a non - axisymmetric ` bar ' instability , the total energy released by gravitational waves could be @xmath0 times greater than the axisymmetric case @xcite .
rotational instabilities of rotating stars arise from non - axisymmetric perturbations of the form @xmath1 , where @xmath2 is the azimuthal angle .
the @xmath3 mode is known as the _ bar mode _ , which is often the fastest growing unstable mode .
there are two kinds of instabilities .
a _ dynamical _ instability is driven by hydrodynamics and gravity , and develops on a dynamical timescale , i.e. the time for sound waves to travel across the star .
a _ secular _
instability is driven by dissipative processes such as viscosity or gravitational radiation reaction , and the growth time is determined by the dissipative timescale .
these secular timescales are usually much longer than the dynamical timescale of the system .
an interesting class of secular and dynamical instabilities only occur in rapidly rotating stars
. one convenient measure of the rotation of a star is the parameter @xmath4 , where @xmath5 is the rotational kinetic energy and @xmath6 is the gravitational potential energy . dynamical and secular instabilities set in when @xmath7 exceeds the critical values @xmath8 and @xmath9 respectively .
it is well known that @xmath10 and @xmath11 for uniformly rotating , constant density and incompressible stars , the maclaurin spheroids @xcite .
numerous numerical simulations in newtonian theory show that @xmath8 and @xmath9 have roughly these same values for differentially rotating polytropes with the same specific angular momentum distribution as the maclaurin spheroids @xcite . however , the critical values of @xmath7 are smaller for polytropes with some other angular momentum distributions @xcite . and
general relativistic simulations also suggest that the critical values of @xmath7 are smaller than the classical maclaurin spheroid values @xcite .
most of the stability analyses to date have been carried out on stars having simple _ ad hoc _ rotation laws .
it is not clear whether these rotation laws are appropriate for the nascent neutron stars formed from the accretion induced collapse of rotating white dwarfs .
new - born neutron stars resulting from the core collapse of massive stars with realistic rotation laws were studied by mnchmeyer , janka and mller ( mnchmeyer & mller 1988 ; janka & mnchmeyer 1989ab ) , and zwerger and mller .
the study of mnchmeyer et al .
shows that the resulting neutron stars have @xmath12 .
zwerger and mller carried out simulations of 78 models using simplified analytical equations of state ( eos ) .
they found only one model having @xmath13 near core bounce .
however , @xmath7 remains larger than 0.27 for only about one millisecond , because the core re - expands after bounce and slows down .
the pre - collapse core of that model is the most extreme one in their large sample : it is the most rapidly and most differentially rotating model , and it has the softest eos . in addition , they found three models with @xmath14 .
rampp , mller and ruffert subsequently performed 3d simulations of three of these models .
they found that the model with @xmath15 shows a non - linear growth of a non - axisymmetric dynamical instability dominated by the bar mode ( @xmath3 ) .
however , no instability is observed for the other two models during their simulated time interval of tens of milliseconds , suggesting that they are dynamically stable .
their analysis does not rule out the possibility that these models have non - axisymmetric secular instabilities , because the secular timescale is expected to range from hundreds of milliseconds to few minutes , much longer than their simulation time .
the aim of this paper is to improve zwerger and mller s study by using realistic eos for both the pre - collapse white dwarfs and the collapsed stars . for the pre - collapse white dwarfs , we use the eos of a zero - temperature degenerate electron gas with electrostatic corrections .
a hot , lepton rich protoneutron star is formed as a result of the collapse .
this protoneutron star cools down to a cold neutron star in about 20 s ( see e.g. burrows & lattimer 1986 ) , which is much longer than the dynamical timescle .
so we adopt two eos for the collapsed stars : one is suitable for protoneutron stars ; the other is one of the standard cold neutron - star eos . instead of performing the complicated hydrodynamic simulations , however
, we adopt a much simpler method .
we assume ( 1 ) the collapsed stars are in rotational equilibrium with no meridional circulation , ( 2 ) any ejected material during the collapse carries a negligible amount of mass and angular momentum , and ( 3 ) the neutron stars have the same specific angular momentum distributions as those of the pre - collapse white dwarfs .
the justifications of these assumptions will be discussed in section [ col - mod ] .
our strategy is as follows .
first we build the equilibrium pre - collapse rotating white dwarf models and calculate their specific angular momentum distributions
. then we construct the resulting collapsed stars having the same masses , total angular momenta and specific angular momentum distributions as those of the pre - collapse white dwarfs .
all computations in this paper are purely newtonian . in the real situation ,
if a dynamical instability occurs , the star will never achieve equilibrium .
however , our study here can still give a useful clue to the instability issue .
the paper is organized as follows . in the next section we present equilibrium models of pre - collapse , rapidly and rigidly rotating white dwarfs . in section [ col - mod ] ,
we construct the equilibrium models corresponding to the collapse of these white dwarfs .
the stabilities of the collapsed objects are discussed in section [ stability ] .
finally , we summarize our conclusions in section [ conclusion ] .
as mass is accreted onto a white dwarf , the matter in the white dwarf s interior is compressed to higher densities .
compression releases gravitational energy and some of the energy goes into heat @xcite . if the accretion rate is high enough , the rate of heat generated by this _ compressional heating _ is greater than the cooling rate and the central temperature of the accreting white dwarf increases with time .
the inner core of a carbon - oxygen ( c - o ) white dwarf becomes unstable when the central density or temperature becomes sufficiently high to ignite explosive carbon burning .
carbon deflagration releases nuclear energy and causes the pressure to increase .
however , electron capture behind the carbon deflagration front reduces the temperature and pressure and triggers collapse .
such a white dwarf will either explode as a type ia supernova or collapse to a neutron star .
which path the white dwarf takes depends on the competition between the nuclear energy release and electron capture @xcite .
if the density at which carbon ignites is higher than a critical density of about @xmath16 @xcite , electron capture takes over and the white dwarf will collapse to a neutron star .
however , if the ignition density is lower than the critical density , carbon deflagration will lead to a total disruption of the whole star , leaving no remnant at all .
more recent calculations by bravo & garca - senz , taking into account the coulomb corrections to the eos , suggests that this critical density is somewhat lower : @xmath17 .
the density at which carbon ignites depends on the central temperature .
the central temperature is determined by the balance between the compressional heating and the cooling and so strongly depends on the accretion rate and accretion time . for zero - temperature c - o white dwarfs , carbon ignites at a density of about @xmath18 @xcite , which is higher than the above critical density .
if the accreting white dwarf can somehow maintain a low central temperature during the whole accretion process , carbon will ignite at a density higher than the critical density , and the white dwarf will collapse to a neutron star . the fate of an accreting white dwarf as a function of the accretion rate and the white dwarf s initial mass is summarized in two diagrams in the paper of nomoto ( see also nomoto & kondo 1991 )
. roughly speaking , low accretion rates ( @xmath19 ) and high initial mass of the white dwarf ( @xmath20 ) , or very high accretion rates ( near the eddington limit ) lead to collapse rather than explosion . under certain conditions , an accreting oxygen - neon - magnesium ( o - ne - mg )
white dwarf can also collapse to a neutron star @xcite .
the collapse is triggered by the electron captures of @xmath21 and @xmath22 at a density of @xmath23 .
electron captures not only soften the eos and induce collapse , but also generate heat by @xmath24-ray emission .
when the star is collapsed to a central density of @xmath18 , oxygen ignites @xcite . at such a high density , however , electron captures occur at a faster rate than the oxygen burning and the white dwarf collapses all the way to a neutron star . in this section
, we explore a range of possible pre - collapse white dwarf models .
we assume that the white dwarfs are rigidly rotating .
this is justified by the fact that the timescale for a magnetic field to suppress any differential rotation , @xmath25 , is short compared with the accretion timescale .
for example , @xmath26 if the massive white dwarf has a magnetic field @xmath27 .
we construct three white dwarf models using the eos of a zero - temperature degenerate electron gas with coulomb corrections derived by salpeter .
all three white dwarfs are rigidly rotating at the maximum possible angular velocities .
_ model i _ represents a c - o white dwarf with a central density of @xmath28 , the highest @xmath29 a c - o white dwarf can have before carbon ignition induces collapse .
_ model ii _ is also a c - o white dwarf but has a lower central density , @xmath30 .
this is the lowest central density for which a white dwarf can still collapse to a neutron star after carbon ignition .
_ model iii _ is an o - ne - mg white dwarf with @xmath31 , that is the density at which electron captures occur and induce the collapse .
since the densities are very high , the pressure is dominated by the ideal degenerate fermi gas with electron fraction @xmath32 that is suitable for both c - o and o - ne - mg white dwarfs .
coulomb corrections , which depend on the white dwarf composition through the atomic number @xmath33 , contribute only a few per cent to the eos at high densities , so the three white dwarfs are basically described by the same eos .
we treat the equilibrium rotating white dwarfs as rigidly rotating , axisymmetric , and having no internal motion other than the motion due to rotation .
the lichtenstein theorem ( see tassoul 1978 ) guarantees that a rigidly rotating star has reflection symmetry about the equatorial plane .
we also neglect viscosity and assume newtonian gravity . under these assumptions
the equilibrium configuration is described by the static euler equation : @xmath34 where @xmath35 is pressure ; @xmath36 is density ; @xmath37 is the gravitational potential , which satisfies the poisson equation @xmath38 where @xmath39 is the gravitational constant .
the fluid s velocity @xmath40 is related to the rotational angular frequency @xmath41 by @xmath42 , where @xmath43 is the distance from the rotation axis and @xmath44 is the unit vector along the azimuthal direction .
the eos we use is barotropic , i.e. @xmath45 , so the euler equation can be integrated to give @xmath46 where @xmath47 is a constant .
the enthalpy ( per mass ) @xmath48 is given by @xmath49 and is defined only inside the star .
the boundary of the star is the surface with @xmath50 .
the equilibrium configuration is determined by hachisu s self - consistent field method @xcite : given an enthalpy distribution @xmath51 , we calculate the density distribution @xmath52 from the inverse of equation ( [ enthalpy ] ) and from the eos .
next we calculate the gravitational potential @xmath53 everywhere by solving the poisson equation ( [ poisson ] )
. then the enthalpy is updated by @xmath54 with @xmath55 and @xmath56 determined by two boundary conditions . in hachisu s paper , the axis ratio , i.e. the ratio of polar to equatorial radii , and the maximum density are fixed to determine @xmath55 and @xmath56 .
however , we find it more convenient in our case to fix the equatorial radius @xmath57 and central enthalpy @xmath58 , so that @xmath59 \ , \end{aligned}\ ] ] where @xmath60 and @xmath61 are the gravitational potential at the centre and at the star s equatorial surface respectively .
the procedure is repeated until the enthalpy and density distribution converge to the desired degree of accuracy .
we used a spherical grid with @xmath62 radial spokes and @xmath63 evenly spaced grid points along each radial spoke .
the spokes are located at angles @xmath64 in such a way that @xmath65 correspond to the zeros of the legendre polynomial of order @xmath66 : @xmath67 .
because of the reflection symmetry , we only need to consider spokes lying in the first quadrant .
poisson s equation is solved using the technique described by ipser and lindblom .
the special choice of the angular positions of the radial spokes and the finite difference scheme make our numerical solution equivalent to an expansion in legendre polynomials through order @xmath68 @xcite .
although the white dwarfs we consider here are rapidly rotating , the equilibrium configurations are close to spherical , as demonstrated in the next subsection .
so a relatively small number of radial spokes are adequate to describe the stellar models accurately .
we compared the results of @xmath69 with @xmath70 and find agreement to an accuracy of @xmath71 .
the accuracy of the model can also be measured by the virial theorem , which states that @xmath72 for any equilibrium star ( see e.g. tassoul 1978 ) . here
@xmath5 is the rotational kinetic energy ; @xmath6 is the gravitational potential energy and @xmath73 .
we define @xmath74 all models constructed in this section have @xmath75 .
we constructed three models of rigidly rotating white dwarfs .
all of them are maximally rotating : material at the star s equatorial surface rotates at the local orbital frequency .
models i and ii are c - o white dwarfs with central densities @xmath28 and @xmath30 respectively ; model iii is an o - ne - mg white dwarf with @xmath31 .
the properties of these white dwarfs are summarized in table [ wdmodels ] .
we see that the angular momentum @xmath76 decreases as the central density @xmath29 increases , because the white dwarf becomes smaller and more centrally condensed .
although @xmath5 increases with @xmath29 , @xmath77 increases at a faster rate so that @xmath4 decreases with increasing @xmath29 .
we also notice that the mass does not change much with increasing @xmath29 .
the reason is that massive white dwarfs are centrally condensed so their masses are determined primarily by the high density central core . here
the degenerate electron gas becomes highly relativistic and the coulomb effects are negligible , so the composition difference is irrelevant .
hence the white dwarf behaves like an @xmath78 polytrope , whose mass in the non - rotating case is independent of the central density .
the masses of our three models are all greater than the chandrasekhar limit for non - rotating white dwarfs .
a non - rotating c - o white dwarf with @xmath28 has a radius @xmath79 and a mass @xmath80 .
when this white dwarf is spun up to maximum rotation while keeping its mass fixed , the star puffs up to an oblate figure of equatorial radius @xmath81 and polar radius @xmath82 , and its central density drops to @xmath83 .
this peculiar behaviour is caused by the soft eos of relativistic degenerate electrons , which makes the star highly compressible and also highly expansible .
when the angular velocity of the star is increased , the centrifugal force causes a large reduction in central density , resulting in a dramatic increase in the overall size of the star .
[ cols="^,^,^,^,^,^,^,^,^,^ " , ] table [ tab : prop ] shows some properties of the collapsed objects resulting from the collapse of the three white dwarfs in section [ ini - mod ] .
we define the radius of gyration , @xmath84 , and the characteristic radius , @xmath85 , of the star by @xmath86 we see that @xmath84 and @xmath85 that result from the same initial white dwarfs are insensitive to the neutron - star eos , while there is a dramatic difference in the central density @xmath29 and the ratio of rotational kinetic energy to gravitational potential energy @xmath7 .
the collapsed stars with the hot eos have smaller @xmath29 and @xmath7 than those with the cold eos .
in fact , the central densities of these hot stars are less than nuclear density .
it is well - known that a non - rotating star can not have a central density in the sub - nuclear density regime ( @xmath87 ) because the eos is too soft to render the star stable against gravitational collapse .
it has been suggested that if rotation is taken into account , a star with a central density in this regime can exist .
such stars are termed ` fizzlers ' in the literature @xcite . however
, these so - called fizzlers in our case can exist for only about 20 s before evolving to rotating cold neutron stars . in order to build a stable cold model in the sub - nuclear density regime
, the collapsed star has to rotate much faster , which is impossible unless the pre - collapse white dwarf is highly differentially rotating .
we mention in section [ introduction ] that zwerger and mller performed 2d hydrodynamic simulations of axisymmetric rotational core collapse .
their pre - collapse models are rotating stars with @xmath78 polytropic eos , which is close to the real eos of a massive white dwarf .
all of their pre - collapse models have a central density of @xmath18 ( see their table 1 ) .
the model a1b3 in their paper is the fastest ( almost ) rigidly rotating star , but its total angular momentum @xmath76 and @xmath7 are respectively 22 per cent and 40 per cent less than those of our model i of the pre - collapse white dwarf , though both have the same central density .
this suggests that the structure of a massive white dwarf is senstive to the eos .
zwerger and mller state in their paper that no equilibrium configuration exists that has @xmath88 for the ( almost ) rigidly rotating case .
this assertion is confirmed by our numerical code .
zwerger and mller adopt a simplified analytical eos for the collapsing core . at the end of their simulations , the models a1b3g1-a1b3g5 , corresponding to the collapsed models of a1b3 , have values of @xmath7 less than 0.07 , far smaller than the @xmath7 s of our collapsed model i ( see table [ tab : prop ] ) , indicating that the eos of the collapsed objects also play an important role on the final equilibrium configurations ( or that their analysis violates one of our assumptions ) . figures [ deni]-[deniii ] show the density contours of the collapsed objects .
we see that the contours of the dense central region look like the contours of a typical rotating star .
as we go out to the low density region , the shapes of the contours become more and more disk - like . eventually , the contours turn into torus - like shapes for densities lower than @xmath89 . in all cases ,
the objects contain two regions : a dense central core of size about 20 km and a low density torus - like envelope extending out to 1000 km from the rotation axis . since we truncate the @xmath90-curve , we can not determine accurately the actual boundary of the stars .
the contours shown in these figures have been checked to move less than one per cent as the cutoff @xmath91 is changed from 0.7 to 0.9 .
this little change is hardly visible at the displayed scales .
figure [ freq ] shows the rotational frequency @xmath92 as a function of @xmath43 , the distance from the rotation axis .
we see that the cores of the cold models are close to rigid rotation .
the rotation periods of the cores of the cold neutron stars are all about 1.4 ms , slightly less than the period of the fastest observed millisecond pulsar ( 1.56 ms ) .
a further analysis reveals that @xmath93 in the region @xmath94 , where @xmath95 for the cold models and @xmath96 for the hot models . to gain an insight into the structure of the envelope
, we define the kepler frequency @xmath97 at a given point on the equator as the angular frequency required for a particle to be completely supported by centrifugal force , i.e. @xmath97 satisfies the equation @xmath98 , where @xmath99 is the magnitude of gravitational acceleration at that point . figure [ freq - ratio ] plots @xmath100 as a function of @xmath43 along the equator . for
the cold models , the curves increase from 0.5 at the centre to a maximum of about 0.95 at @xmath101 , then decrease to a local minimum of about @xmath102 , and then gradually increase in the outer region .
the curves of the hot models , on the other hand , increase monotonically from about 0.4 at the centre to over 0.7 in the outer region . in all cases
, centrifugal force plays an important role in the structure of the stars , especially in the low density region .
figure [ fig : mpm ] plots the cylindrical mass fraction @xmath103 as a function of @xmath43
. in all cases the cores contain most of the stars mass .
material in the region @xmath104 occupies only a few per cent of the total mass , but it is massive enough that its self - gravity can not be neglected in order to compute the structure of the envelope accurately .
the envelope can be regarded as a massive , self - gravitating accretion torus .
the same structure is also observed in the core collapse simulations of janka & mnchmeyer and fryer & heger ( private communication with fryer ) .
figure [ tow ] shows @xmath105 as a function of @xmath43 , where @xmath106 and @xmath107 are the rotational kinetic energy and gravitational potential energy inside the cylinder of radius @xmath43 , i.e. @xmath108 the values of @xmath109 approach @xmath7 when @xmath110 for the cold eos models and when @xmath94 for the hot eos models .
this suggests that material in the region @xmath94 contains negligible amount of kinetic energy , and any instability developed in this region could not produce strong gravitational waves .
we first consider axisymmetric instabilities , i.e. axisymmetric collapse .
this stability is verified when we construct the models . recall that we start from the model of a non - rotating spherical star which is stable .
then we use it as an initial guess to build a sequence of rotating stellar models with the same specific angular momentum distribution but different total masses and angular momenta .
if the final model we end up with is unstable against axisymmetric perturbations , there must be at least one model in the sequence such that @xmath111 which signals the onset of instability @xcite . here
@xmath112 is the total mass and @xmath29 is the central density .
the partial derivative is evaluated by keeping the total angular momentum @xmath76 and specific angular momentum distribution @xmath90 fixed .
we have verified that all of our equilibrium models in the sequencies satisfy @xmath113 .
hence they are all stable against axisymmetric perturbations .
we next consider non - axisymmetric instabilities .
we have @xmath114 for the cold eos models and @xmath115 for the hot eos models ( table [ tab : prop ] ) .
the hot models are probably dynamically stable but may be secularly unstable .
however , since they are evolving to cold neutron stars in about 20 s and their structures are continually changing on times comparable to the secular timescale , we shall not discuss secular instabilities of these hot models here .
the values of @xmath7 for the three cold neutron stars are slightly less than the traditional critical value for dynamical instability , @xmath10 .
this critical value is based on simulations of differentially rotating polytropes having the @xmath90-distribution of maclaurin spheroids .
however , recent simulations demonstrate that differentially rotating polytropes having other @xmath90-distributions can be dynamically unstable for values of @xmath7 as low as 0.14 @xcite .
the equilibrium configurations of some of those unstable stars also contain a low density accretion disk like structure in the stars outer layers .
this feature is very similar to the equilibrium structure of our models .
hence a more detailed study has to be carried out to determine whether the cold models are dynamically stable .
the subsequent evolution of a bar - unstable object has been studied for the past 15 years @xcite .
it is found that a bar - like structure develops in a dynamical timescale .
however , it is still not certain whether the bar structure would be persistent , giving rise to a long - lived gravitational wave signal , or material would be shed from the ends of the bar after tens of rotation periods , leaving an axisymmetric , dynamically bar - stable central star . even if the cold neutron stars are dynamically stable , they are subject to various secular instabilities .
the timescale of the gravitational - wave - driven bar - mode instability can be estimated by @xmath116 @xcite . in our case ,
@xmath117 ( see figure [ tow ] ) , @xmath118 and @xmath119 , so @xmath120 .
gravitational waves may also drive the r - mode instability @xcite .
the timescale is estimated by @xmath121 for the @xmath122 r - mode at low temperatures @xcite , where @xmath123 is the average density .
inserting @xmath123 for the inner 20-km cores of the cold stars , we have @xmath124 . the evolution of the bar - mode secular instability has only been studied in detail for the maclaurin spheroids .
these objects evolve through a sequence of deformed non - axisymmetric configurations eventually to settle down as a more slowly rotating stable axisymmetric star @xcite .
it is generally expected that stars having more realistic eos will behave similarly .
we have constructed equilibrium models of differentially rotating neutron stars which model the end products of the accretion induced collapse of rapidly rotating white dwarfs .
we considered three models for the pre - collapse white dwarfs .
all of them are rigidly rotating at the maximum possible angular velocities .
the white dwarfs are described by the eos of degenerate electrons at zero temperature with coulomb corrections derived by salpeter .
we assumed that ( 1 ) the collapsed objects are axisymmetric and are in rotational equilibrium with no meridional circulation , ( 2 ) the eos is barotropic , ( 3 ) viscosity can be neglected , and ( 4 ) any ejected material carries negligible amounts of mass and angular momentum .
we then built the equilibrium models of the collapsed stars based on the fact that their final configurations must have the same masses , total angular momenta and specific angular momentum distributions , @xmath90 , as the pre - collapse white dwarfs .
two eos have been used for the collapsed objects .
one of them is one of the standard cold neutron - star eos .
the other is a hot eos suitable for protoneutron stars , which are characterized by their high temperature and high lepton fraction .
the equilibrium structure of the collapsed objects in all of our models consist of a high density central core of size about 20 km , surrounded by a massive accretion torus extending over 1000 km from the rotation axis .
more than 90 per cent of the stellar mass is contained in the core and core - torus transition region , which is within about 100 km from the rotation axis ( see figure [ fig : mpm ] ) .
the central densities of the hot protoneutron stars are in the sub - nuclear density regime ( @xmath125 ) .
the structures of these protoneutron stars are very different from those of the cold neutron stars , which the protoneutron stars will evolve to in roughly 20 s. the protoneutron stars have lower central densities , rotate less rapidly , and have smaller values of @xmath7 . on the other hand ,
the structures of the three cold neutron stars are similar .
their central densities are around @xmath126 and their central cores are nearly rigidly rotating with periods of about 1.4 ms , slightly less than the fastest observed millisecond pulsar ( 1.56 ms ) .
zwerger and mller performed 2d simulations of the core collapse of massive stars .
the major difference between their models and ours is that they used rather simplified eos for both the pre - collapse and the collapsed models .
when compared with their fastest rigidly rotating model , alb3 , we found their pre - collapse star has less total angular momentum and smaller @xmath7 than the pre - collapse white dwarf of our model i , although both have the same central density .
the differences between their final collapsed models ( a1b3g1-a1b3g5 ) and ours are even more significant .
the values of @xmath7 of our collapsed objects are much larger than theirs , suggesting that the eos plays an important role in the equilibrium configurations of both the pre - collapse white dwarfs and the resulting collapsed stars .
the values of @xmath7 of the cold neutron stars are only slightly less than the traditional critical value of dynamical instability , 0.27 , frequently quoted in the literature .
the cold neutron stars may still be dynamically unstable and a detailed study is required to settle the issue .
even if they are dynamically stable , they are still subject to various kinds of secular instabilities .
a rough estimate suggests that the gravitational - wave driven bar - mode instability dominates .
the timescale of this instability is about 0.1 s.
|
equilibrium models of differentially rotating nascent neutron stars are constructed , which represent the result of the accretion induced collapse of rapidly rotating white dwarfs .
the models are built in a two - step procedure : ( 1 ) a rapidly rotating pre - collapse white dwarf model is constructed ; ( 2 ) a stationary axisymmetric neutron star having the same total mass and angular momentum distribution as the white dwarf is constructed .
the resulting collapsed objects consist of a high density central core of size roughly 20 km , surrounded by a massive accretion torus extending over 1000 km from the rotation axis .
the ratio of the rotational kinetic energy to the gravitational potential energy of these neutron stars ranges from 0.13 to 0.26 , suggesting that some of these objects may have a non - axisymmetric dynamical instability that could emit a significant amount of gravitational radiation .
= = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = # 1 latexl-.36em.3ex-.15em t-.1667em.7ex-.125emx [ section ] [ firstpage ] stars : neutron stars : rotation stars : interiors white dwarfs
instabilities
|
the calculation of precise radial density profiles for a localized radiation source ( _ i.e. , _ soft x - ray emission by hot plasmas ) has been a long standing problem in both plasma physics and astrophysics @xcite . for a sufficiently tenuous plasma in which radiation emission processes dominate over absorption processes ,
the symmetric density profile of the plasma can be obtained from the projected radiation intensity by abel s inversion .
another example of the application of abel s inversion is the determination of density profiles of an imploding spherical targets driven by laser - beams @xcite .
these two seemingly different diagnostic methods are mathematically identical and rely on the solutions of abel s integral equation @xcite . before proceeding further ,
it is important to point out that there have been extensive studies on the absorption of a soft x - ray by the abel s integral equation for medical applications by cormack @xcite who apparently was not aware of previous work on the x - ray diagnostics employed in astrophysics @xcite . despite numerous studies over many years @xcite ,
there still exists considerable debate over the importance of a numerical method for solving the integral equation .
the numerical algorithm by barr @xcite has attracted considerable interest , particularly with the observation of a diffuse background due to compton scattering . in all x - ray diagnostics ,
the x - rays are normally produced by a high - energy electron accelerator and depend on the bremsstrahlung process in a high z - material .
in particular , the phermex electron accelerator at los alamos that is a 50 mhz , 30 mev standing - wave device which operates at 500 a with 600 kv injector .
it has been relatively a simple task to focus the high current electron beam by injecting a neutral plasma by means of self - focusing by a self- pinching mechanism @xcite and thus creates a point - like high energy x - ray ( or @xmath0-ray ) source .
the x - rays are produced by the scattering of the high - energy electron beam at a target material ( usually high z - material such as tungsten ) .
the electrons lose their energy by two basic process : collision and radiation .
the collision process leaves the atoms in the material either excited states or the ionized states .
the ejected electrons have a small energy that is deposited locally .
on the other hand the x - rays ( radiation energy ) generated by bremsstrahlung process is uniformly distributed among secondary photons of all energies up to the energy of the primary electron energy itself . at the energy of phermex electron beam
, the bremsstrahlung process dominates over the collision process . for an optimal radiography
, it is essential to have a pencil beam to create a point source of the x - rays at a distance far from the test object .
this provides approximately a parallel x - ray beam as shown in figure [ fig : fig1 ] .
moreover , the spot size of a point source determines the spatial resolution in an x - ray film .
this optimal condition allows one to study a density profile of a test object by means of abel s integral equation @xcite . in all previous numerical calculations based on the volterra solution of abel integral equation @xcite , expressing the data as a well - behaved function was an essential step to avoid the noise amplification .
when the data , which contain small random errors due to a photon cascade and large fluctuations by the enhanced backscattering at the impurities , are smoothed by polynomial fits , the calculations tend to obscure the basic processes .
moreover , the smoothing leads to a hybridization of random errors and fluctuations such that the computed absorption coefficients do not provide any meaningful information about absorption coefficients . worse yet , the results are often misleading . here
we report a study , by flash x - ray photography , of the absorption coefficients of metallic cylinders of high z - material ( @xmath1 ) with a high concentration of impurities .
first we discuss a new numerical solution to abel s integral equation and its application to a cylindrically symmetric object using experimental data .
we then discuss the results , their relevance , and connection to other work .
last we discuss optimal experimental conditions with regard to feasibility of x - ray radiography as a quantitative tool for x - ray diagnostics .
x - ray absorptions at the surface of condensed matter in photon energies below 40 kev has been extensively studied .
experimental measurements of the absorption coefficients are in good agreement with detailed theoretical analysis @xcite
. however , the hard x - ray absorption mechanism at the surface has not been fully explored .
we show , by specific examples , that since compton scattering and pair production are two competing processes in x - ray absorption in the photon energy up to 5 mev @xcite , where the upper limit of the photon energy depends on the z - number ( the cross - section of pair - production becomes equal to that of compton scattering at 5 mev for pb ; it is 15 mev for al ) , compton scattering leads to a partial reflection of x - rays at a sharp boundary with a curvature .
furthermore , we show that since the basic x - ray transmission equation remains correct to the single - scattering approximation , the mean free - path of an incident photon in a given material must be sufficiently large to neglect the effect of multiple scattering . a solid cylinder with different holes in size at both ends
is ideally suited for studying the effect of a partial reflection of x - rays at a sharp boundary on the absorption coefficients .
a single exposure of an x - ray beam to the cylinder provides a data for a solid cylinder as well as for a cylinder with a void .
moreover , the circular curvature allows one to study the angular dependence of scattered photons at the surface , where the scattering centers are located .
an analytic solution to abel s integral equation that is least sensitive to local random errors is the starting point of our study .
we show , by an analytic method , that abel s integral equation admits a different form of a solution than that obtained by volterra @xcite .
this modified form of a solution not only reduces the computing time by an order of magnitude , but also allows one to analyze the effect of intensity fluctuations on the absorption coefficients .
first we review the basic x - ray absorption processes in a high - z metallic test object and discuss the relative importance of the processes at a given energy of an x - ray .
next we review abel s integral equation and derive its solutions in sec .
we then study numerical methods of data analysis based on the new solution to abel s integral equation in sec .
iv and compare the numerical results with the known absorption coefficients of material .
the application of the numerical code is discussed in sec .
v. in sec .
vi we discuss an optimal experimental condition for x - ray diagnostics and the basic requirements of an x - ray source for the condition .
lastly in sec .
vii , the feasibility of quantitative x - ray diagnostics is discussed .
pulsed high - energy x - ray sources ( _ e.g. _ phermex ) have provided a way to obtain x - ray snap shots of dynamical processes such as imploding spherical targets in an icf program .
these snap shots provide the transmitted x - ray intensity profiles with a diffuse background . in phermex facility at los alamos ,
the high - energy @xmath0-rays are produced by short pulsed high - energy electron beam .
the high - energy electron beam strikes the target material in which the energy loss of the beam by bremsstrahlung is fairly uniformly distributed among the secondary photons of all energies from zero up to the energy of the primary electron itself ( 30 mev ) , for which the bremsstrahlung process dominates over the collision loss @xcite . at some electron energy ,
the radiation loss is equal to the collision loss .
this energy coincides with the critical energy of the material , a parameter that also plays an important role in the shower theory of rossi @xcite .
thus at high - energies a large fraction of the electron energy is converted to high - energy @xmath0-rays in the target which in turn may interact with test object .
the theory of high - energy photon absorption mechanism ( in the order of @xmath2 mev ) has been extensively studied .
the basic physics is fairly easy to understand and is fully described in refs .
@xcite .
next we discuss the radiation absorption process in a test object ; at high - energies , the radiation absorption by pair - production ( electron - positron ) dominates over compton scattering .
hence the high - energy @xmath0 rays create an electromagnetic cascade shower . the photoelectric effect and the multiple coulomb scattering of the secondary electrons by atoms
perturb the cascade shower .
this coupled with compton scattering creates a diffuse background in a radiographic film . in a thick test object ,
the energy of a low - energy secondary electron in the beam is dissipated by either excitation or the ionization of atoms .
we will discuss further the basic processes in the radiation absorption below . in general ,
as a beam of x - rays penetrates an absorbing material , its intensity shows an exponential decay .
it is because a photon in the beam is removed from its original direction of penetration by either absorption or scattering in _ a single event_. since the number of photons removed in a distance @xmath3 is proportional to the beam intensity @xmath4 , the intensity of a monochromatic beam passing through the matter of a test object must decrease exponentially and is given by @xmath5 , \label{absorption}\ ] ] where the absorption coefficient @xmath6 represents the average number of absorption and scattering processes a single photon undergoes per @xmath7 in its path and @xmath8 is the beam intensity at @xmath9 . neglecting the effect of multiple scattering and photon cascade , the cross - section for the primary processes ( _ i.e. , _ photoelectric effect , compton scattering , and pair production ) by which x - rays or @xmath0-rays are absorbed is given by the sum : @xmath10 .
the absorption coefficient can be defined as @xmath11 , where @xmath12 is the number of atoms of the absorbing matter per cubic centimeter .
briefly we now digress to summarize the salient points of the photon absorption processes in condensed matter for the application of radiography : at photon energies less than 0.5 mev , x - ray interaction with the material is dominated by photo - absorption in which an x - ray photon is completely absorbed by a bound electron in atoms .
consequently the bound electron jumps from its ground state to an excited state or into the continuum ; the latter process is an ionization of the atom , which is usually more probable than the excitation .
the cross - section for photoelectric effect is derived first by heitler @xcite .
when the photon energy is not too close to the absorption edge , the cross - section for the photo - emission of a single electron from the k shell is given by @xmath13 , where @xmath14 is the cross - section thompson scattering , @xmath15 the fine - structure constant , z the atomic number , and @xmath16 the photon energy .
thus in the region of photon energies of the order of a few kev , the cross - section for the photoelectric effect decreases rapidly as @xmath17 . for high - energy photons , @xmath18 , the cross - section for the photoelectric effect
is given by @xmath19 .
this shows that at high - energies the cross - section decreases as @xmath20 but photoelectric effect can still contribute to x - ray absorption for high - z material up to 5 mev ( e.g. , pb ) although its effect becomes increasingly small . at
photon energies higher than 1 mev , the cross - section for the photo - absorption becomes negligible , however , compared to that of compton scattering .
scattering of photons by free elctrons plays an important role in x - ray or @xmath0-ray absorption in condensed matter .
the cross - section for compton scattering by relativistic corrections was first derived by klein and nishina @xcite .
we note that since the binding energy of the k shell electrons in atoms is of the order of kev , the scattering a high energy photon in mev by a bound electron is essentially equivalent to compton scattering by a free electron . for a cylindrical test object
, the scattering center is well defined at the surface and the angular dependence of scattered photons by the surface can be studied by the klein - nishina formula . for high - energy photons
, the formula shows that the angular distribution of scattered photons tends to peak in the forward direction , which agrees well with the observation .
the conservation of energy and momentum give @xmath21 , where @xmath22 and @xmath23 are the energy of the incident and scattered photons respectively , and @xmath24 the angle in the units of @xmath25 , @xmath26 the velocity of light .
if the energy of an incident photon is large compared with the rest mass of an electron , the effect of compton scattering can be studied in two extreme cases : for a very small scattering angle , the change of photon energy is negligible , and at large scattering angles , the scattered photon energy is always of the order of @xmath27 .
thus the effect of compton scattering in this case reduces to the classical thompson scattering and the change in the photon energy becomes negligible .
an important point to emphasize is that , at high energies , the cross - section for compton scattering @xmath28 varies as @xmath29 $ ] , where @xmath30 is the classical electron radius and @xmath22 is the energy of a photon in the units of @xmath25 .
thus the cross - section for compton scattering decreases with the increasing energy of an incident photon .
this is precisely the reason why the penetrating power of x - rays ( or @xmath0-rays ) increases with increasing energy so long as no other absorption processes are important . in the coulomb field of a nucleus
, a high - energy photon creates an electron - positron pair ( i.e. , @xmath31 .
no pair production can occur if the photon energy is less than @xmath32 or in the absence of the nuclear coulomb field required for the conservation of energy and momentum @xcite .
the cross - section for the pair - production increases rapidly with the photon energy and reaches an asymptotic value in the relativistic limit . if the photon energy @xmath33 , then the cross - section takes the asymptotic form : @xmath34 $ ] , where the effect of atomic electrons through their screening of the nuclear charge is taken into account .
this limiting form shows that the cross - section increases with the square of the z - number .
we also note that , for high energy photons , most of the electrons and positrons are emitted in the forward direction .
since both the photoelectric effect and compton scattering decrease as the photon energy increases , the pair production is the dominant absorption mechanism for very high - energy photons . furthermore the critical energy above which pair production dominates in photon absorption decreases with increasing z number .
for example , the cross - section for pair production becomes equal to that of compton scattering at about 5 mev for pb . in elementary quantum field theory , there is a useful , general substitution rule which is valid to arbitrary orders and which relates processes of the type , @xmath35 , to the processes @xmath36 , where @xmath37 denotes the antiparticle of b. this substitution rule relates compton scattering to the annihilation of an electron - positron pair into two photons ( _ i.e. , _ @xmath38
@xcite . in a similar manner ,
the pair production process in condensed matter is closely related to bremsstrahlung process by high energy electrons @xcite . in condensed matter
, @xmath0-rays can undergo multiple scatterings , a sequence of scattering of the incident photon by bound electrons , if the mean free path of the photon is small compared to the size of a medium , which is usually the case for high - z material @xcite .
thus the critical parameter determining the effect of multiple scattering becomes negligible if the size of a medium is less than a few times of the mean free path of a photon , which is given as @xmath39 , where @xmath6 is the absorption coefficient at a given energy in material ( _ e.g. , _
@xmath40 cm at @xmath41 for pb ) .
we notice from figure 32 of bethe and ashkin @xcite that the absorption coefficient for any given material goes through a minimum at some energy @xmath42 of an incident photon energy .
the nature of the transmitted x - rays depends on whether the energy of an incident photon is greater or less than @xmath43 .
for example , if the energy of a primary photon is less than @xmath44 , the primary photon is more penetrating than the secondary photon .
thus in a test object whose dimension is about the same as the mean free path of a photon , the effect of multiple scattering is negligible .
it would be difficult to carry out a systematic treatment of the multiple scattering involving photoelectric effect , compton scattering , and pair production . in spite of this complexity
, an elaborate analysis of x - ray transmission in thick targets has been carried out analytically that shows how to correct the simple transmission rate eq .
( 1 ) , beyond its simplifying approximations .
fortunately , we now have an excellent monte carlo code ( egs4 ) @xcite developed at slac , which can be used to test its accuracy against the experimental data .
furthermore , this test helps us understand the detailed absorption and production mechanism of @xmath0-rays for a given electron beam from the accelerator , such as phermex facility at los alamos . at high - energies @xmath45 , where absorption by compton scattering dominates
, one may treat the multiple scattering as a series of small angle compton scattering in which the photon energy loss is small for @xmath0-rays that penetrate a large distance in the absorbing medium . with this approximation
, the asymptotic solution for the transmission rate was obtained .
it contains a factor @xmath46 , a build - up factor , where @xmath47 is the smallest absorption coefficient for a given material ( _ e.g. , _
@xmath47 is the absorption coefficient at 2.5 mev for pb ) .
the build - up factor depends on whether the incident photon energy is greater or smaller than the energy at which the absorption coefficient is minimum @xcite .
it is also well known that at very high energies ( @xmath48 ) , absorption by pair production dominates and thus the secondary compton scattered photons may become more penetrating . under this condition ,
the spectral equilibrium is controlled by the hardest secondary photons and the transmission takes a form @xmath49 , where @xmath50 is the minimum value at some wavelength ( or energy ) of an incident photon for any given material @xcite . here
the energy of an x - ray photon is related to its wavelength by @xmath51 , where @xmath52 is planck s constant . in laboratory units , @xmath53 , where @xmath54 is the wavelength in angstrom . at extremely high energies ,
x - ray absorption by pair production becomes increasingly important . in this process
, a photon creates an electron - positron pair which in turn can produce secondary photons by bremsstrahlung processes ; the pair can also annihilate into two photons ( i.e. , @xmath55 ) .
the secondary photons can also undergo photo - absorption , compton scattering or pair production , a phenomenon known as a photon cascade .
the effects on the absorption coefficient of multiple scattering and photon cascades can be appreciable , depending on the range of an incident photon energy @xcite . in connection with compton scattering of soft photons ,
we discuss the reflection of an x - ray beam at a sharp boundary by treating the x - ray beam as an incident electromagnetic wave .
a semi - classical analysis of the reflection of x - rays gives the index of refraction for x - rays as @xmath56 , where @xmath57 and @xmath58[z+\delta f^{'}\lambda^{2}]$ ] .
here @xmath59 is the avogadro s number , @xmath60 the density , and @xmath61 the atomic number weight , @xmath62 is small away from the absorption edges @xcite . since the index of refraction is less than 1 ,
the reflection can occur if the incident glancing angle is sufficiently small . from snell s law , the critical angle below which the total reflection can occur
is determined by @xmath63 , where we have neglected the absorption term .
this semi - classical analysis is often useful in identifying the reflection of soft x - rays from a bremsstrahlung source at a sharp boundary .
for an x - ray wavelength of @xmath64 , the critical angle @xmath65 . as the incident glancing angle increases from the critical angle , the reflection intensity decreases .
an important refinement of the x - ray absorption process can be made from the observation that in a disordered medium the transmitted beam intensity shows a local fluctuation .
given the enormous amount of effort that has been devoted to the x - ray absorption mechanism , remarkably little is known about the nature of intensity fluctuations .
the concept of localization of electrons in a solid was invoked in 1958 by anderson to explain the electron transport phenomena in the presence of impurities @xcite .
the concept is usually associated with a disordered medium in which interference of multiple scattering of electrons by impurities give rise to enhanced backscattering .
we can see from eq .
( [ absorption ] ) that apart from the @xmath66 and @xmath67 absorption edges the total cross section of the primary processes is a smooth function of an incident photon energy .
the transmission rate eq .
( [ absorption ] ) was derived based the single scattering approximation , which is valid as long as the mean free path of an incident photon for any given material is not small compared with the dimension of a test object .
this implies that an independent , identical n atom model can not explain the fluctuations of beam intensity .
one final comment on the fluctuation of a transmitted x - ray intensity is note - worthy .
anderson localization was observed in experiments with a monochromatic light source @xcite .
unfortunately , the wide spectrum of a bremsstrahlung source can be a contributing factor to the intensity fluctuations but it is difficult to assess this effect .
we can test our qualitative analysis of beam transportation in a material in each step by using the code egs4 @xcite which incorporates many different absorption processes which are almost impossible to carry out a detailed analysis in practice .
we now consider a symmetric cylinder whose material composition is uniform in the @xmath68-direction perpendicular to the incident x - ray beam and focus on an infinitely thin disk of the cylinder that is sufficiently far from the x - ray source .
this allows one treat the x - ray source as a spatially uniform beam , which is an essential requirement for the derivation of abel s integral equation @xcite .
furthermore we assume that a detector accepts only those x - rays which traverse a narrow slice from the source ( _ i.e. , _ a narrow beam approximation ) . for a pure absorption or a single scattering ,
the flux of an x - ray beam emerging from the disk , @xmath69 along the slice is given by @xmath70,\ \label{flux}\ ] ] where @xmath8 is the flux of an incident monochromatic beam [ see figure [ fig : fig1 ] ] and is assumed to be independent of @xmath71 , since we assume a uniform beam as an x - ray source . in figure
[ fig : fig2 ] , @xmath72 and @xmath73 , where @xmath74 is the radius of the disk .
we have also assumed the absorption coefficient @xmath75 is cylindrically symmetric and hence our analysis is applicable to an one - dimensional problem , which is often misunderstood in practice . before displaying abel s integral equation , which is the principal object of this paper , it is appropriate to comment on the narrow beam approximation . in an ideal situation
, the transmitted x - ray intensity @xmath69 must remain nearly monochromatic as the incident x - ray beam , since the scattered x - rays are assumed not to reach a detector ( _ i.e. , _ a narrow beam approximation ) . in the presence of compton scattering
the narrow beam approximation breaks down if the scattering angle is large .
this is the reason why we observe a diffuse background in an x - ray film .
if we define @xmath76 , then eq . ( [ flux ] ) becomes @xmath77 here @xmath78 gives the sum of absorption or scattering along the slice .
this quantity appears naturally in the integral equation for all observable which , like an x - ray absorption , represents strictly local interactions of x - rays with atoms .
we shall therefore give a brief discussion of @xmath78 itself , before returning to the details of abel s integral equation . in slab geometry ,
@xmath78 remains positive or become zero if there is no absorption . in the presence of compton scattering
, @xmath78 tends to give a value less than the expected value because of smear from the adjacent slices .
although the effect of compton scattering is negligible in slab geometry ( _ i.e. , _ a white noise ) , its effect becomes serious in a test object with a geometrical structure ; @xmath78 can give a negative value which signals a breakdown of a narrow beam approximation .
isolating @xmath78 from the background in an x - ray transmission then requires the subtraction of the structure - dependent quantity , which is unknown _ a priori _ , a process requiring very accurate data analysis .
let us now return to the derivation of the integral equation .
if we eliminate the variable @xmath79 in favor of @xmath80 [ see figure [ fig : fig2 ] ] , eq .
( [ abs2 ] ) can be written as @xmath81 this is a form of abel s integral equation @xcite . with abel s transformation ,
volterra s solution to eq .
( [ abel ] ) is given as @xcite @xmath82 where the prime on @xmath78 denotes a differentiation with respect to its argument .
equation ( [ abel ] ) is our fundamental expression for the material absorption coefficient .
its equivalence to the x - ray emission from a hot plasma is clear from the derivation of eq .
( [ abel ] ) but can be made explicit by replacing @xmath83 and @xmath84 in abel s integral equation .
thus eq .
( [ abel2 ] ) is applicable to both the radiation absorption and emission processes @xcite . in practice , @xmath78 is given as a set of discrete data taken from a detector or an x - ray film rather than as an analytic function .
normally , experimental data also contain a diffuse background ( _ i.e. , _ random errors ) due to compton scattering or photon cascades . in the presence of fluctuations of a beam intensity
, it is difficult to eliminate these errors either by fourier - filtering or other numerical methods .
the differentiation on the data points introduces spikes in noise in the integral .
thus a numerical representation of the data poses a difficult problem that must be treated with care .
the numerical integration of eq .
( [ abel2 ] ) is best avoided entirely , if the numerical accuracy is desired .
bockasten @xcite , and recently wing @xcite also recognized this point , but their calculations of polynomial fits of data points or expansion of the kernel of integral equation limit the accuracy of the results .
as an alternative to the use of eq .
( [ abel2 ] ) , barr @xcite obtained a solution of eq .
( [ abel ] ) by means of abel s transformation .
his solution allows one to devise an efficient smoothing method for data having a small random errors .
his numerical algorithm eliminates the random errors after the integration , but before differentiation .
this procedure ensures smooth results for a gaussian type of density profiles .
however , his solution is numerically unstable near the center . to overcome this difficulty we look for a different solution to eq .
( [ abel ] ) that is numerically more stable .
we start with the identity @xcite , @xmath85 where @xmath86 . by making a transformation @xmath87
, we obtain @xmath88 for @xmath89 , eq . ( [ abel3 ] ) reduces to @xmath90 next we form a product integral [ abel5 ] @xmath91^{1/2}}\\ & = & \int_{z}^{r}r\mu(r)dr\left[\int_{z}^{r}\frac{y dy}{[(y^{2}-z^{2})(r^{2}-y^{2})]^{1/2}}\right]\\ & = & \pi\int_{z}^{r}r\mu(r)dr , \label{abel6}\end{aligned}\ ] ] where we have used dirichlet s formula @xcite and eq .
( [ abel4 ] ) . differentiating both sides of eq .
( [ abel6 ] ) , we obtain @xmath92 which can be rewritten as @xmath93 here we notice that , in the limit @xmath94 , @xmath75 becomes numerically unstable , since @xmath75 is indeterminate at the limit @xcite .
we now derive a different form of a solution to abel s integral equation eq .
( [ abel ] ) , which is equivalent to eq .
( [ abel2 ] ) but is less sensitive to random errors , and removes the difficulty in barr s solution eq .
( [ abel7 ] ) . as before
, we form a product integral @xmath95 \int_{y}^{r}\frac{\mu(r)r dr}{(r^{2}-y^{2})^{1/2}}. \label{new}\end{aligned}\ ] ] by means of dirchlet s formula @xcite , the integral can be rearranged as @xmath96^{1/2}}\right].\ ] ] with the transformation @xmath97^{1/2}=\eta(r^{2}-y^{2})$ ] .
the integral in the bracket @xmath98 $ ] can be evaluated by an elementary method , yielding @xmath99 thus we have @xmath100 using eq ( [ new1 ] ) we derive the following equation by taking the advantage of the near continuous nature of the function @xmath78 and with the similar transformation : [ last ] @xmath101 some features of eqs .
( [ last ] ) are similar to those of barr , eq . ( [ abel7 ] ) @xcite . while he starts with abel s transformation , we start , however , from dirichlet s formula @xcite and carry out the final integral by an elementary method . an important point about this approach
is that it avoids the direct differentiation of data points .
instead , we integrate over the data points , which , with proper choice of mesh points , smooths random errors in the data .
next we show that abel s integral equation admits three different forms of a solution , which are all equivalent as shown below : [ test ] @xmath102 differentiating both sides of eq .
( [ test ] ) with respect to @xmath80 , we obtain @xmath103 thus we have @xmath104 which shows that the solution by barr @xcite is equivalent to eq .
( [ new1 ] ) . to show that eq .
( [ abel4 ] ) is equivalent to volterra s solution , we integrate @xmath105 by parts and obtain @xmath106 where the boundary condition @xmath107 was used .
next we differentiate both sides of eq .
( [ test2 ] ) and obtain @xmath108 hence we have @xmath109 which completes the proof that there are indeed three different solutions to eq .
( [ abel ] ) .
we may choose any one of the three solutions .
the simplest numerical method , in principle , and the one that is least sensitive to small random errors , and the most computationally efficient method provide the criteria for selecting a particular form of the solution : eq .
( [ last ] ) meets all of these criteria .
before we proceed with the numerical computation , it is necessary to study the analytic structure of eq .
( [ abel ] ) and eq .
( [ int2 ] ) .
we see at once that both eq .
( [ abel ] ) and eq .
( [ int2 ] ) possess an end - point singularity .
therefore , the numerical accuracy will depend on how we treat the end - point singularities in these equations .
we may proceed with our numerical calculations by either polynomial expansion of data points similar to bockasten s method @xcite or by carrying out the numerical integration for a given set of data by extrapolation ( or interpolation ) .
however , in the presence of fluctuations in the x - ray intensity , the former approach tends to introduce errors which are too large to neglect and leads to erroneous results .
the latter approach suffers , of course , from the difficulty of treating the end - point singularity .
fortunately the numerical algorithm by bickley @xcite has been very useful for the evaluation of singular functions at the integral limits .
we carry out the numerical integrations of eq .
( [ abel ] ) and eq .
( [ int2 ] ) using the numerical scheme of jeffereys and jeffereys @xcite , which employ bickley s formula near the integral limits and carry out the remaining integral by the gregory formula .
a more elaborate but similar end - point formula has been also developed by young @xcite .
however , we prefer that of jeffereys and jeffereys @xcite , because it is easier to adapt to different sets of experimental data .
next we discuss the steps to reduce the raw experimental data to properly extracted data for numerical analysis .
there are two aspects in the processing experimental data : a selection of necessary data points and optimum background removal .
the selection of data points to carry out the numerical integration and differentiation depends on the numerical algorithm . in most experiments , data taken from an x - ray film
provide a sufficient number of data points . by three point interpolation or extrapolation
, we set the data at the cell boundaries for the integrations and carry out the differentiation at the cell centers .
this process effectively smoothes the data in the short wavelength regime , but it preserves the fluctuation in the long wavelength regime . in particular , this numerical scheme is useful for any type of background removal ( _ e.g. , _ white noise ) , and to make certain that there are no oscillations in the background whose frequencies are much higher than those of fluctuations in the data .
( [ abel ] ) and eq .
( [ int2 ] ) play critical roles in the present work . to check the self - consistency of the numerical code ,
we have calculated @xmath78 based on a theoretically model given by eq .
( [ abel ] ) , which are in turn , used to calculate the absorption coefficients from eq .
( [ int2 ] ) . in the absence of appreciable noise in @xmath78 ,
the code calculates the precise profiles of the absorption coefficients ( see figures [ fig : fig5 ] and [ fig : fig6 ] ; and figures [ fig : fig9 ] and [ fig : fig10 ] ) .
it should be noticed that eq .
( [ abel ] ) has the same analytic structure as eq .
( [ int2 ] ) , and that the same subroutine in the code can be used to compute both @xmath78 and @xmath110 .
the advantage of using the theoretical values @xmath78 ( _ i.e. , _ instead of a true experimental value in slab geometry ) is that they can be calculated without including effects of wide spectrum of an x - ray source and compton scattering of photons with electrons in a metal .
moreover they can be used to see if a significant amount of noise is present in the experimental data .
thus this intermediate step of analyzing the experimental data is extremely useful for identifying the large angle compton scattering effects .
within the limits of our approximation on the x - ray absorption mechanism ( _ i.e. , _ neglecting the effects of multiple scattering and photon cascades since these effects are extremely small in the present experiment ) , we may calculate the absorption coefficients from the experimental data .
however there are several other important effects which we must address : first we have assumed the x - ray source is monochromatic in the derivation eq .
( [ absorption ] ) to calculate @xmath78 .
this is perhaps the most serious assumption of the approximations made in the derivation of abel s integral equation .
this assumption together with the spatial uniformity of the x - ray beam is essential for the determination of absorption coefficients for any given material by means of abel s inversion , since the absorption coefficients for any given material depend on the photon energy of x - ray source .
the spatial uniformity requires a point source of an x - ray which is the major challenge for an accelerator physicist because the point source requires a pencil beam of a high energy electron in a high current accelerator ( _ e.g. , _ phermex at los alamos ) @xcite .
hence the basic assumption of a monochromatic x - ray source is hard to achieve , since the bremsstrahlung source has a wide spectrum ; the x - ray spectrum spans the entire energy range from an incident electron beam energy of an accelerator to zero ( _ e.g. , _
@xmath111 mev to zero in phermex ) . in other words ,
the bremsstrahlung spectrum is fairly flat function of photon energy .
if the x - ray intensity at a great depth from a bremsstrahlung source depends on multiple scattering , the beam intensity tends to reach an equilibrium state ( _ i,.e .
, _ the progressive formation and decay of the hardest secondary soft photons ) .
the reason for this process is that the progressive filtering of an x - ray beam from a bremsstrahlung source tends to make the attenuation curve steeper than @xmath112 , but the secondary photons from compton scattering can compensate for the steepening if the beam penetrates to a great depth to allow the multiple scattering .
for example , the mean free path of the average photon energy of 1.2 mev in @xmath1 ia approximately @xmath113 , which is roughly equal to the size of test object , the effect of the progressive filtering of an x - ray beam by the test object can not be compensated for by any simple physical mechanism .
next we turn to a discussion of an x - ray transmission in a material with a void whose radius is much larger than the coherence length of a disordered medium . the effect of the void on the absorption coefficients is similar to the surface absorption in that incident x - rays see a concave surface where a partial reflection of the x - rays can occur @xcite and the primary photons may lose its energy by compton scattering at a large scattering angle near the surface ( _ i.e. , _ a reflection of x - rays at the sharp boundary )
. it differs , however , in that the net effect of progressive filtering and compton scattering of photon depends on a geometrical structure factor .
the presence of a sharp boundary leads to multiple reflections of soft photons in a void such that an assumed free propagation of photons without a reflection is no longer appropriate , _
i.e. _ , a radiation trap .
hence the spatial analysis given in cormack s paper @xcite requires a proper correction at the interface between aluminum and lucite because of a sharp boundary between them .
in addition , the calculation is invalid because of the ill - defined gamma ray source which should be assumed to be spatially uniform to make use of abel s inversion .
although the spectrum of @xmath114 source is not as wide as bremsstrahlung source , the gamma ray beam still experience a partial reflection at a sharp boundary .
the present work was undertaken to study the above undesirable effects experimentally .
in particular , to investigate a possible beam intensity fluctuations , we have selected metallic @xmath1 cylinder , which contains a small fraction of impurities that simulate a random , disordered medium . the schematic picture of the experimental arrangement is given in figure [ fig : fig1 ] .
the test object is a cylinder of @xmath1 , 2 _ in _ long and 0.5 _ in _ diameter with holes on both ends is exposed to a high energy x - ray source with the average energy of photons @xmath115 .
the x - ray beam emerging from the test object enters a photographic plate through 0.25 _ in _ thick lead filter , 3 m down stream from the x - ray source .
the data from the x - ray film were then taken , with a microdensitometer digitized , and then processed by a minicomputer to eliminate the background .
experimental data sets for both solid and hollow disks are given in figures [ fig : fig3 ] and [ fig : fig4 ] , to which a digital noise reduction technique has not been applied . in figures [ fig : fig5 ] and [ fig : fig6 ] , we show @xmath78 as a function of @xmath71 by eq .
( [ abel ] ) with @xmath116 for a cylinder of @xmath1 . in order to see the details of discrepancy between the experimental data and the theoretical values ,
we have also computed @xmath117 from the experimental data , which is displayed in figures [ fig : fig7 ] and [ fig : fig8 ] .
the quality of @xmath117 should be critically examined by comparing it with @xmath78 .
first we see from figures [ fig : fig5 ] and [ fig : fig7 ] that @xmath117 shows a long tail extending beyond the outer boundary .
crucial to any x - ray diagnostics is the calibration of the measured transmitted beam and source intensities .
this can be done either with the theory using a measured absorption coefficient from a slab geometry or empirically from the measured standards .
the reason for this is the difficulty of estimating the incident flux intensity @xmath8 which varies from shot to shot in a bremsstrahlung source .
the basic goal in determining @xmath8 is to match as closely as possible the ideal situation of measurements in the slab geometry .
the long tail in figure [ fig : fig7 ] clearly suggests that the calibration for @xmath117 was seriously in error , because @xmath117 must be zero at the outer surface and become negative beyond the boundary due to compton scattering at the surface .
next we observe that @xmath118 exhibits the intensity fluctuation which , we believe , is due to enhanced backscattering in a disordered medium ( anderson localization @xcite ) . in the experimental data on carbon reported by several different groups @xcite ,
one notices a wide range of disagreement in the absorption coefficients . in an experiment with carbon
, one must be concerned with x - ray absorption in a disordered medium ( _ i.e. , _ due to a random distribution of small voids ) .
we believe the disagreement can be attributed to anderson localization , which implies a breakdown of a model of @xmath12 independent , identical atoms for a medium .
this is a macroscopic realization of quantum many - body effects in a disordered medium .
furthermore , the recent experiments of light propagation in a disordered medium @xcite provide direct evidence of enhanced backscattering with the light intensity fluctuation which are very sensitive to the pattern of speckles in the medium .
these measurements are consistent with our data , which suggests that the x - ray intensity fluctuations are due mostly to impurities in @xmath1 and partly by the broad spectrum of the bremsstrahlung source .
if we compare figures [ fig : fig5 ] and [ fig : fig7 ] , we can clearly see the effects of progressive filtering and large angle compton scattering ; the breakdown of the narrow beam approximation is particularly evident .
an important question concerning the use of x - ray radiography for quantitative determination of a density profile is how accurately both the progressive filtering and compton scattering effects can be determined and their contribution eliminated . if they can be accurately accounted for , then , only then
, the remaining data can provide a density profile within the limit of the narrow beam approximation .
we observe from figures [ fig : fig5 ] and [ fig : fig7 ] that the breakdown of the transmission rate , eq .
( [ absorption ] ) , increases as @xmath80 approaches the center of a test object .
we also see appreciable effects of compton scattering at the edge , which produce a diffuse boundary .
this shows that one can not hope to isolate these effects by a simple numerical method or background removal , and thereby obtain precise absorption coefficients by abel s inversion .
these effects must be corrected by filtering technique that can keep the spectrum width @xmath119 as small as possible ; this eliminates the progressive filtering by a test object . by placing an appropriate filter in front of a film
, we may further reduce the effect of compton scattering at the edge .
again , what has been said about the compton scattering effect also applies to the curved surface of a test object in which an oblique incidence of an x - ra beam leads to a partial reflection of the beam .
this can be corrected only by an analytic method .
comparing figures [ fig : fig6 ] and [ fig : fig8 ] we observe similar effects in @xmath117 for a hollow cylinder , but a large discrepancy between @xmath117 and @xmath78 exists near the center extending to the boundary of the void . in the absence of dense material at the center of the test object , the peculiarly strong absorption of x - rays at the void can be explained only in terms of the multiple reflection of soft x - rays at the the sharp boundary ( i.e. , a radiation trap ) , as will be discussed later .
the absorption coefficient @xmath6 in the numerical calculation of @xmath78 was measured in the slab geometry in which the surface scattering effects are subtracted by adjusting @xmath120 where @xmath121 is a correction term .
although considerable accuracy of the experimental values has been achieved from such measurements , similar measurements using an object with a geometrical structure do not provide the same degree of accuracy , partly because we are not able to find a reasonably accurate analytical form for @xmath121 .
contrary to common belief , the significant discrepancy between @xmath117 and @xmath78 in the presence of a void in materials does not involve spatial resolution in connection with the spot size of an x - ray source , but is instead due to both compton scattering and the progressive filtering .
this is evident because we can still locate the boundaries of a test object whose dimension is less than the spot size .
finally we have calculated the absorption coefficients for both a sold cylinder figure [ fig : fig9 ] and a cylinder with a void at the center figure [ fig : fig10 ] using the numerical code developed based on the eqs .
( [ last ] ) .
the results clearly show that eqs .
( [ last ] ) are indeed a correct solution to the abel s integral equation eq .
( [ abel ] ) and the numerical accuracy of the code is indeed impressive . by means of abel s inversion eqs .
( [ last ] ) , the absorption coefficients are computed from the experimental data with the code and are shown in figures [ fig : fig11 ] and [ fig : fig12 ] . analyzing the data from the computed absorption coefficients has the advantage by comparing with figures [ fig : fig9 ] and [ fig : fig10 ] in that one clearly sees the effects of enhanced backscattering and compton scattering .
the main result obtained from this numerical computation of the absorption coefficients is the demonstration of a strikingly accurate numerical algorithm that treats the random noise without any sign of amplification .
moreover , the calculated absorption coefficients show a consistency with the experimental data .
figure [ fig : fig11 ] shows the calculated absorption coefficient as a function of radius for the solid cylinder of @xmath1 .
several important qualitative conclusions can be drawn from the figure .
fist of all , the variation of the calculated absorption coefficient exhibits irregular patterns containing a smooth oscillation in each isolated region , a phenomenon that can only arise as a result of enhanced backscattering in a disordered medium .
moreover , the calculated absorption coefficient reflects a small but pronounced local fluctuation in the x - ray intensity in the medium . as mentioned earlier
, such an effect has also been observed in the experiments on the multiple scattering of visible light by an aqueous suspension of sub - micrometer size mono - disperse polystyrene spheres @xcite .
these experiments confirmed the enhanced backscattering of light waves by disordered scattering centers in a medium .
a qualitative picture of the dependence of the correction term @xmath121 on the glancing angle can be inferred , for the energy region of our interest , from the surface reflection phenomenon at x - ray energies less than 10 kev @xcite .
the shape of @xmath121 sensitively depends on the incident photon energy and glancing angles less than the critical angle ( above which there is no reflection ) . by adjusting the correction term to give the theoretical values @xmath78 in closer agreement with the experimental data
, one learns which physical processes are responsible for the various peculiarities in the data . for the same cylinder with a void
, the absorption coefficient calculated from the experimental data is shown in figure [ fig : fig12 ] .
the effect of a partial reflection of x - rays at the boundary of a void can be clearly seen .
the error in the absorption coefficient is about 52 % at the center , which appears to suggest the presence of radiation absorbing material .
this significant disagreement between the computed value and the expected value ( figure [ fig : fig12 ] ) reflects the neglect , in this simple form of absorption mechanism , of the fundamental physics treating a void in a medium or suggests an error in the calibration of the measured transmitted beam and source intensities , or both . as explained above , the calibration of the experimental data is clearly in serious error .
a proper calibration would have reduced the error in the absorption coefficient to less than 30 % .
any simple physical process that may have been involved in this x - ray transmission is not abstruse , since one notices that the reflection of an incident x - ray beam at the outer and inner surfaces can reduce the transmitted beam intensity significantly .
the maximum reflection takes place at the edge of a void where the glancing angle becomes nearly equal to the critical glancing angle @xcite .
as the glancing angle increases , the effect of x - ray reflections reduces . it should be emphasized that a bremsstrahlung source is far from being monochromatic as we will discuss later .
thus soft x - rays , whose critical glancing angles are much larger than those of hard x - rays , will reflect more readily at the surfaces . for high - energy photons ,
the angular dependence of scattered photons must be analyzed by the klein - nishina formula @xcite .
it is straightforward to assess the effect of compton scattering at a sharp boundary for a monochromatic source , but it becomes prohibitively difficult for a bremsstrahlung source .
thus we can see that compton scattering at a sharp boundary leads the breakdown of the basic assumptions of x - ray transmission in the void for which egs4 can be utilized to see the effect of the bremsstrahlung source . in passing we note that , in compton scattering experiment , a target of circular cross section such as cylinder is often used for the study of angular dependence ; the scattering center at the surface is better defined even if the beam is poorly collimated .
however , as we have seen in our data analysis , accurate evaluation of the flux intensity at the target is difficult .
we are thus faced with a question : how does the x - ray reflection at the surface of a void differ from that of the outer surface ? moreover , in the presence of photo - emitted electrons , how are we to describe the x - ray transmission in the void ? to answer these questions we must focus on the nature of interactions of photons at the surface and in the free space of the void . the incident x - rays see the concave surface , where a partial reflection of the primary and secondary photons can take place .
a fraction of the primary and secondary photons are ultimately absorbed near the boundary of a void , giving rise to enhanced photoemission of electrons in the void .
we also note that , in the absence of a strong coulomb field of nucleus , no pair - production can occur in the void .
furthermore , compton scattering of primary and reflected photons by photo - emitted electrons in the void is small compared with that of condensed matter , but it can be a contributing factor to the discrepancy , since the trapped photon intensity is much larger than the primary x - ray beam intensity for a small void . in general , as expected , the effect of the x - ray reflections on the transmitted beam intensity is less severe as the radius of a void increases .
this can be easily understood by the fact that the x - ray reflection takes place near the edge .
an important point to notice is that near the edge of the void we see a significant dip in the absorption coefficient which shows the reflection of x - rays . in summary ,
_ we have shown that a numerical calculation of density profile by abel s inversion has its limitation and can not be used as a quantitative tool for a hydrodynamic test ! _ the conventional approach to examine the film by experienced eyes will provide a better diagnosis for a hydrodynamic test .
the noise is inherently too large in its magnitude to neglect .
in addition , a three - dimensional density profile of a test object can not be obtained in an environment of a dynamical test of hydrodynamics as discussed below .
in order to obtain reasonably accurate absorption coefficients from the experimental data by means of abel s inversion , the most optimal experimental condition must be realized so that the data to be analyzed are free of significant distortion . to derive the abel s integral equation eq .
( [ abel ] ) , several simplifying assumptions are made .
we examine the assumptions to minimize the errors that they introduce and to find the optimal experimental conditions .
the most serious approximation made was to assume that a bremsstrahlung source is monochromatic .
the potential difficulty introduced by this assumption can be appreciated if we recall that a bremsstrahlung source spans the entire spectrum exhibiting a relatively flat intensity profile with respect to photon energy @xcite .
furthermore an absorption coefficient for any given material depends on photon energy and z - number .
thus it is important to limit the width of the spectrum to the smallest possible value .
there are several technique for limiting the wavelength range @xmath122 available , with each its advantages and disadvantages .
the simplest method is to use a crystal monochromater .
however this would be impractical in our x - ray diagnostics , because the wavelength of our x - rays is too short compared with a lattice parameter of available material . from fig .
32 of bethe and ashkin @xcite we see that there is a region of maximum transparency for all material ( mtm ) which lies between 2.5 mev and 10 mev .
this is admittedly a rather loosely defined spectral range , its limit being in principle different for each metal , but one may think of the spectral range as the window of maximum transparency for heavy materials . in the language of scattering theory , one can say that while at high energies ( _ i.e. , _
@xmath123 ) the mean free path is so small that for a thick target a photon must undergo multiple scattering to reach a detector .
as the photon energy is lowered into the region of mtm the photon mean path is sufficiently large that the effect of multiple scattering can be negligible . as discussed earlier , the transmission rate defined in eq .
( [ absorption ] ) does not admit multiple scattering , which can be a serious effect for a thick targets . because the x - rays in the region of mtm can penetrate to a great depth without significant multiple scattering , a varying photon energy @xmath124 can cover many blemishes in abel s integral equation .
thus the solution of the integral equation can give a reasonably reliable absorption coefficient apart from the large angle compton scattering which can be eliminated without difficulty .
this discussion has an important implication with respect to a design of an x - ray source for an x - ray diagnostics .
the basic goal in selecting @xmath125 is to match as closely as possible the upper limit of the region of mtm with the smallest possible spectral width .
since the maximum photon energy in a bremsstrahlung source is nearly equal to an incident electron beam energy hitting a tungsten target , the upper limit can be set by selecting a specific electron energy .
a better technique than a monochromater for reducing the spectral width is to use filtering materials by which the photons with energy less than 5 mev are eliminated and compton effects are minimized .
it is important to note that since the same absorption coefficient can be measured at two different energies in this region , we must block the low energy photons before they reach a test object .
this is possible because the x - rays to be filtered are of higher energy than the fluorescent signal . by choosing an appropriate filter
, its absorption edge can be placed between the photon energy ( to be filtered ) and fluorescent energies , preferentially absorbing the former .
the major difficulty with this method of x - ray filtering is that absorbed x - rays can be re - emitted as fluorescent radiation that is relatively un - attenuated by filters and strike film to increase the fog level .
this degrades the sensitivity to low level signals .
it should be emphasized , however , that by eliminating the soft photons we reduce both progressive filtering and compton effects at the surface of a test object and that the filtering improves the quality of data considerably .
on the other hand , the filtering reduces the source brightness , which is defined as the number of photons emitted per unit solid angle .
empirically it is found that the bremsstrahlung radiation intensity in terms of the electron beam energy in the forward direction is given @xmath126 , where @xmath127 is the electron beam intensity and @xmath128 is the electron kinetic energy in unit of @xmath27 @xcite . here
the target material was gold .
we can appreciate from the above empirical formula and fig .
30 of koch and motz @xcite that , for example , at a given electron beam energy it is necessary to increase the current intensity in order to maintain the same effective source brightness .
this requirement poses a difficult problem for which the accelerator community is in search of an answer - a new acceleration mechanism capable of generating an intense relativistic beam .
fortunately , a novel acceleration developed by friedman @xcite has been tested in experiments demonstrating that a beam of electron current of 70 ka can be accelerated to a energy of 7 mev .
however , it was found that the acceleration mechanism is not scalable to higher energy due to various instabilities .
the unique feature of the auto - acceleration mechanism is that the electromagnetic field for the acceleration is created by the interaction of the first electron beam with a cavity .
the electromagnetic field energy is then transferred to the second electron beam in the same structure , producing a short pulsed ( _ e.g. , _
20 ns ) , intense electron beam .
moreover , the auto - accelerator circumvents the two major problems in pulsed power accelerators - the fast switches and insulators ; the accelerator is automatically self - synchronized and thus eliminates the jitter due to the switches .
it should be stressed , however , that the efficiency of the auto - acceleration mechanism depends on the ratio of the first beam intensity to that of the second . since the extracted electromagnetic field energy from the first beam is the only source of the acceleration for the second beam , the larger the ratio , therefore , the more efficient the auto - acceleration is bound to become .
consideration of the stability of an intense relativistic electron beam in a structure led friedman @xcite to the modification of previous experiments . in a recent experiment @xcite ,
the first beam was modulated by using an external rf source to enhance the interaction between the beam and the modified structure .
hence the accelerated beam and the modulated beam exchange energy through the modified structure while moving on different trajectories .
friedman has demonstrated that the accelerated beam of 200 attains the kinetic energy of 60 mev with the beam radius of 0.1 cm .
this result strongly suggest that there exists a domain of parameters in which we can select an appropriate accelerated beam intensity for an optimally useful x - ray source .
we now examine in some detail the characteristics of of an intense electron beam for an x - ray source for the radiography : first , we study the efficiency of bremsstrahlung production for a given beam energy , which is defined as the ratio of the total bremsstrahlung radiation emitted to the total energy of an incident electron beam and is given by @xmath129 where @xmath130 is the kinetic energy of an incident electron @xcite .
the efficiency depends on the target material and the kinetic energy of an incident electron beam .
for example , the efficiency @xmath131 % at @xmath132 but it is 41% at @xmath133 for pt targets . note also that the spectrum shape of bremsstrahlung intensity is given by @xmath134}$ ] , where @xmath135 , @xmath136 is the frequency of a photon emitted in the forward direction , and @xmath52 is plank s constant @xcite .
thus for 10 mev electron beam incident on a thick tungsten target , the efficiency @xmath137 and about 35 % of photons emitted in the bremsstrahlung process have their energy higher than 5 mev . in the x - ray transmission , unfiltered soft photons
are absorbed through progressive filtering in a test object and do not reach a detector .
assuming the spectrum width @xmath138 , which is equivalent to 1 mev , the optimally useful x - rays are , therefore , 15 % to 25 % of the total photons emitted in the bremsstrahlung processes . at extremely high electron beam energies like the one in the phermex facility ( the electron beam current @xmath139 and
the electron energy ( @xmath140 ) , the radiation cone from a bremsstrahlung source is extremely narrow and x - ray photons are emitted mostly in the forward direction .
thus spectrum shape can be described by the above analytic expression which shows the peak intensity in the energy range of 5 mev to 6 mev .
it should be remarked , however , that although the efficiency for this beam is @xmath141 % , hard @xmath0-rays from a bremsstrahlung process tend to cascade in a test object and the x - ray transmission is controlled mainly by the photons with energy in the neighborhood of the peak intensity and hardest secondary photons are produced in multiple scattering .
these examples are given to show certain features that provide a qualitative picture to identify an optimally useful x - ray source .
again , one may try to use egs4 code to check if the above qualitative picture holds for a target design of a bremsstrahlung source .
the above analysis shows that the high - current , high - energy electron beam is obviously preferred for an optimally useful x - ray source and the only remaining question that concerns if the photon energy falls in the region of mtm .
the second issue is the more fundamental problem , which is common in all high - current accelerators ; it is important to focus an electron beam to obtain uniformly , diverging x - ray beam to satisfy the assumption of a parallel x - ray beam .
although it is extremely difficult , one should be able to achieve a small spot size by the combined use of a quadrupole magnet and ion focusing @xcite .
we also note that the shape of radiation cone from bremsstrahlung processes depends on the incident electron beam energy and that , with an electron beam energy in the region of mtm , the radiation cone is sufficiently large that the x - ray beam diverges uniformly in space .
this is a critical factor for the spatial uniformity of the x - ray beam , which is the basic assumption in deriving abel s integral equation . in the light of the above problem ,
it is essential to develop a simple , reliable acceleration mechanism such as friedman s auto - acceleration mechanism .
indeed , the simplicity and the efficiency of the auto - acceleration mechanism hold out the hope that one might develop an x - ray source that meets the requirements .
clearly the auto - acceleration mechanism shows definite advantages over a more conventional rf linac for generating a high - current , high - energy electron beam .
moreover , this type of accelerator is compact and inexpensive to build compared to an rf linac .
we would like to emphasize , at the end of this detailed discussion on the optimal condition for an x - ray source , that the physics and experimental technique elaborated here provide a starting point from which we can develop an optimally useful x - ray source with narrow spectrum width and adequate beam intensity .
although our analysis of the experimental data is qualitative in nature , we have attempted to quantify the discrepancy from the form of transmission rate @xmath142 , by using abel s inversion eqs .
( [ last ] ) and its numerical solution .
given the large uncertainties due to the lack of detailed understanding of the physical processes in the x - ray transmission for a given test object with a structure as well as the limited data , our interpretation should be taken as only suggestive .
one of the major problems that confronted us in this experiment was the complexity of the acquisition and interpretation of the experimental data .
abel s inversion is conceptually simple and mathematically straightforward .
qualitative though this picture may be , as it has to be , since a simplified transmission of @xmath0-rays in material can not hold true even in its first - order analysis .
the lack of detailed agreement between the theory based on the a simple x - ray transmission rate and the experimental data , however , shows a complexity of transmission of radiation in high z - material .
since the x - ray transmission itself is inherently complex a phenomenon , there seems little hope to achieve the reliability required for a quantitative method in x - ray diagnostics in hydrodynamics .
_ thus it is highly unlikely that the quantitative diagnostic by the radiographic method in hydrodynamic test can be meaningful .
_ in practice , it can be misleading . that the experimental measurements are not always reliable themselves is not an important issue , since the most optimal condition can be realized , and perhaps reliable data can be obtained .
even under the best condition , the data can not still satisfy the basic assumption on abel s inversion .
in general , several different corrections must be made to the directly observed transmitted x - ray intensity in order to obtain a reasonably credible absorption coefficient by abel s inversion .
it should be emphasized , however , that , in most experiments , @xmath8 and @xmath4 are read simultaneously from a film , and that the incident and the transmitted x - ray beam intensities are assumed to be nearly monochromatic , which is far from being true .
this requirement of an x - ray source is essential for the measurement of absorption coefficients , which depend on the average energy of an incident photons .
numerous simplifying approximations ( _ e.g. , _ the narrow beam approximation , _ etc .
_ ) have been made for reasons of both availability of an x - ray source and expediency .
there is no practical method to produce monochromatic x - ray beam .
the final effects of these approximations remain unknown .
application of abel s inversion in x - ray diagnostics is not as straightforward as for plasma diagnostics since we do not know how to make corrections to the transmitted x - ray intensity .
sometimes one might get the impression that the correction to the transmitted beam intensity , @xmath143 , is something of an art . for the worst case
this is probably true .
one must therefore exercise the scientific prudence by checking every step with the monte carlo code egs4 @xcite to insure that the corrections are well justified and are in the order of procedures weighted by the importance of the processes : a ) an incident x - ray beam intensity ; b ) removal of background due to photon cascade ; c ) compton scattering at a sharp boundary ; d ) progressive filtering due to a finite spectrum width ; e ) a narrow beam approximation .
the relative importance of each of these corrections , as well as the details of correction procedure , depends on the x - ray beam energy and the particular material composition of a test object .
unfortunately , little discussion of these corrections has been attempted in most of the published measurements of x - ray transmission .
this has greatly impeded progress in x - ray diagnostics . a crucial question concerning the application of abel s inversion for quantitative x - ray radiography
is how accurately the compton effects such as partial reflection of x - rays at a sharp boundary and x - ray scattering at the edge can be determined and can be corrected .
unfortunately the most difficult quantity to measure in x - ray radiography is an accurate incident x - ray beam intensity @xmath8 .
a variety of effects can distort the incident beam intensity profile severely while leaving the energy spectrum relatively undisturbed .
yet it is the incident beam intensity that is critical factor in determining the absorption coefficients by abel s inversion .
as explained earlier , the reflected x - ray beam intensity at a sharp boundary varies with the incident glancing angles larger than the critical angle @xmath144 .
this implies that @xmath145 with a test object having a curved surface depends on the geometrical structure factor which is extremely difficult to determine .
any background removal method has the potential of distorting the data .
it is therefore important to isolate each contribution to the correction term by combined use of analytic method and egs4 code .
for example , some of the strongest compton scattering effects occur at the outer edge , where they are caused by the total reflection of an x - ray beam .
occasionally the simplest remedy for this compton effect is , albeit difficult in some instances , to place an appropriate filter in front of a film .
one can see the effect with the clarity of an improved image .
on the other hand , with the data having contributions from progressive filtering of photons , it may be difficult to find a simple remedy that can isolate the filtering effects .
one notices from the absorption curves of metals ( see fig
. 32 of ref .
@xcite ) that one should not use an average photon energy for the calculation of absorption coefficients unless the spectrum width of an x - ray source is small .
one should have expected that , in addition to the progressive filtering , the narrow beam approximation [ eq . ( [ absorption ] ) ] will break down in this most unfavorable case , _
i.e. , _ a bremsstrahlung source without filtering .
the difference between the absorption coefficients calculated from experimental data taken for a cylindrical test object and the experimental value determined in slab geometry is quite significant , suggesting that the major effect ( _ e.g. , _ compton scattering by free and bound electrons at large scattering angles ) in the presence of a sharp boundary was not taken into account . because of this
, it is essential to find the method that is consistent and systematic in the selection of parameters for @xmath143 .
this can be achieved by analytic study of the reflection and absorption of x - rays at the surface of a test object and systematic experimental tests for various test objects .
it is imperative , albeit tedious , to make several attempts at background removal with different sets of parameters , follow each set of experimental data through the entire analysis for different test objects , and test the results for consistency .
although the calculation of @xmath146 is an intermediate step in abel s inversion , a great deal of information may be gained by direct inspection . for test objects with small amount of impurities , the theoretical value @xmath78 calculated from eq .
( [ abs2 ] ) using the experimental value @xmath6 determined in slab geometry can be used for a comparison with @xmath146 .
this comparison will show the fluctuations in the transmitted x - ray intensity and diffuse boundaries .
this exercise is useful as an overview of the data , yielding the information about a possible form of correction to the incident x - ray intensity .
inspection of @xmath146 is also useful as a check of the x - ray beam profile .
the fluctuations in x - ray beam intensity , which should not be confused with statistical fluctuation of noise @xcite , are intrinsic phenomena which one should not apply a smoothing technique in any data analysis ; the fluctuations reflect the property of a material , for example anderson s localization by impurities .
next , it should be stressed that abel s integral equation derived from the simple x - ray transmission rate eq .
( [ absorption ] ) does not admit multiple scattering which can modify the transmission equation significantly . as a general guide in the design of experimental x - ray transmission measurements
, we may find it useful to check if the dimension of a test object is less than @xmath147 , the mean free path of a photon at a given energy .
for a thin target , the multiple scattering can be safely neglected .
one may not able to neglect the multiple scattering , however , in a thick target ( _ e.g. , _ a size larger than @xmath148 ) .
the error introduced in eq .
( [ absorption ] ) by the multiple scattering will enter into the later calculation of the absorption coefficients of materials by abel s inversion .
such a check can be easily made by referring to a published x - ray absorption coefficient table for metals and a systematic test with egs4 code .
finally , we have shown how we may obtain one - dimensional density profile by computing the absorption coefficient with abel s inversion eq .
( [ last ] ) from two - dimensional x - ray film for a cylindrically symmetric object in one x - ray exposure for which an x - ray source is located in a sufficiently far distance from a test object to insure a parallel beam .
there is little need to see a two - dimensional density profile in a radiography , since we can easily obtain the two - dimensional view in one x - ray exposure .
a question naturally arises , however , how one may obtain a three - dimensional density profile that is quite useful in non - invasive diagnosis such as in medicine . since the wave number @xmath149 in cylindrical coordinates
, the spatial resolution requires that _ the @xmath150 exposure by rotating the x - ray source about the axis ( @xmath68-direction ) to obtain the three - dimensional density profile .
_ this requirement was misunderstood in his study of a line integral for application of to radiological application by cormack @xcite and thus his analysis is invalid .
the @xmath150 different views provide the necessary information on density profile in each direction which can be used to construct the three - dimensional density profile as in a three - dimensional radiographic machine . for a given mode number
, the best spatial resolution will be achieved near the axis of symmetry .
a rule of thumb on the spatial resolutions is : the larger the number of exposures @xmath150 is , the more reliable the three - dimensional density profile will be in a numerical reconstruction @xcite .
this can be achieved either by rotating an x - ray source or rotating a test object , which is not feasible in any dynamic hydrodynamic test .
this analysis does clear up the myth that one may be able to build a 3-d hydrodynamic test facility with the two x - ray sources which is known as the dual axis radiographic hydrodynamic test facility or darht project at los alamos @xcite .
h. hrman , z physik * 97 * , 539 ( 1935 ) and v. volterra , _ lecons sur les equations integrals et les equations integro - differentielles ( paris , 1913 ) _ p34 h. maecker , z. physik * 136 * , 119 ( 1953 ) ; * 139 * , 448 ( 1954 ) .
j. friedlich , ann . phys . * 3 * , 327 ( 1959 ) .
k. bockasten , j. opt .
soc . of am . *
51 * , 943 ( 1961 ) .
r. n. bracewell , aust .
j. physics * 9 * , 198 ( 1956 ) .
j. chang , _ et al .
39 * , 279 ( 1981 ) .
n. c. luhman and w. a. peebles , rev .
55 * , 279 ( 1984 ) ; chul park and d. moore , nasa tn dw-5677 ( 1970 ) .
w. l. barr , j. opt .
soc . of am .
* 52 * , 885 ( 1962 ) .
a. m. cormack , j. applied physics , * 34 * , 2722 ( 1963 ) ; * 35 * , 2908 ( 1964 ) . s. j. han ( la - ur-93 - 2146 ) ; steven weinberg , j. math .
* 8 * , 614 ( 1967 ) .
one should be aware of the assumption of a paralell incident x - ray beam which is produced at a far distant by a point source .
hence one can not apply a fourier decomposition on the angular dependence in cormack s equation 2 .
d. c. koningsberger and r. prins , _ x - ray absorption : principles , applications , technique of exafs , sexafs , and xanes _
( john wiley , new york , 1988 ) and the references therein .
b. rossi , _ high - energy particles _ , ( prentice - hall , inc .
englewood cliffs , new jersey , 1952 ) h. w. koch and j. w. motz , rev .
31 * , 920 ( 1959 ) .
w. heitler , _ the quantum theory of radiation _ , ( oxford univ . press , oxford 1954 )
j. d. bjorken and s. d. drell , _ relativistic quantum mechanics _ , ( mcgraw - hill , new york , 1953 ) p133 .
bethe and j. ashkin , _ experimental nuclear physics _ , ( john wiley , new york , 1964 ) vol .
i , p166 - 357 .
a. c. melissinos , _ experiments in modern physics _
p150 and p252
. this optimal condition does not allow the fourier decomposition of abel s integral equation as was carried out by cormack .
hence comack s analysis is not valid for an optimal condition of radigraphy .
a. b. migdal , phys . rev . * 103 * , 1811 ( 1956 ) .
j. w. motz , h. a. olsen , and h. w. koch , rev .
phys . * 69 * , 581(1969 ) .
walter r. nelson , hideo hirayama , and david w. o. rogers , _ the egs4 code system_(slac report-265 , 1985 ) . j. o. hirschfelder , j. j. magee , and m. h. hull , phys . rev . *
73 * , 852(1948 ) .
j. o. hirschfelder and e. n. adams ii , phys . rev .
* 73 * , 863 ( 1948 ) .
u. fano , phys . rev . * 76 * , 739 ( 1949 ) .
g. c. wick , phys . rev . *
75 * , 738 ( 1949 ) .
b. rossi and k. greisen , rev .
* 13 * , 240 ( 1941 ) .
l. g. parratt , phys . rev . * 95 * , 359 ( 1954 ) .
p. w. anderson , phys .
b * 109 * , 1492 ( 1958 ) .
e. t. whittaker and g. n. watson , _ a course of modern analysis _ ( cambridge univ
. press , 1953 ) p77 and p229 .
h. jefferey and b. jefferey , _ methods of mthematical analysis _
( cambridge univ .
press , 1953 ) p288 .
p. r. karr and j. c. lamkin , phys . rev . *
76 * , 1843 ( 1949 ) .
l. v. spencer and f. jenkins , phys . rev . * 76 * , 1885 ( 1949 ) .
u. fano , h. hurwitz , jr . and
l. v. spencer , phys . rev . * 77 * , 425 ( 1950 ) .
y. s. tsai , rev .
phys . * 46 * , 815 ( 1974 ) .
m. alvada and a. lagendijk , phys .
lett . * 55 * , 2692 ( 1985 ) .
p. e. wolf and g. maret , phys .
* 55 * , 2696 ( 1985 ) .
this assumption is critical to the derivation of abel s inversion ( cat scan ) .
hence the three dimensional cat scan is feasible only if a test object is rotated with sufficiently small angular resolution or the x - ray source is rotated about the axis of symmetry .
one may apply the fourier decomposition with n - different views , since each view provides a one dimensional density profile as shown below .
e.t . whittaker and g. n. watson , _ a course of modern analysis _
( cambridge univ . press , 1953 ) p77 and 229 .
g. m. wing , la - ur-84 - 1234 .
h. jeffereys and b. jeffereys , _ methods of mathematical physics _ ( cambridge univ .
press , 1953 ) p288 .
a. young , phil . mag .
trans . a , 552 ( 1954 ) .
d. c. creagh , nucl .
inst . and methods in phys .
* a225 * , 1 ( 1987 ) . s. j. han , la - ur-93 - 2146 . _ stability of a relavistic electron beam in a resistive plasma channel_. m. friedman , phys .
* 32 * , 92 ( 1974 ) .
m. friedman , _ et al .
* 63 * , 2468 ( 1989 ) and references therein . m. v. nezlin , plasma physics,*10 * , 337 ( 1968 ) . m. e. rose and m. m. shapiro , phys .
rev . * 74 * , 1853 ( 1948 ) .
r. w. hamming , _ numerical methods for scientists and engineers _ ( mcgraw - hill book company , inc , 1962 ) chapter 22 .
|
the hard x - ray ( @xmath0-ray ) absorption by cylindrically symmetric @xmath1 test objects is studied by means of @xmath0-ray transmission measurements . to make a precise comparison between the theoretically modelled values and the absorption coefficients calculated from the experimental data , we have developed a highly accurate numerical code based on a new solution of abel s integral equation .
it is shown that progressive filtering , surface reflections by compton scattering , and the enhanced backscattering due to impurities can explain much of the observed discrepancy .
we also discuss optimal experimental conditions with regard to the feasibility of quantitative radiography for @xmath0-ray diagnostics .
|
strong / weak - coupling duality , or s - duality , exists in a number of quantum systems . in 4d gauge theories , it takes the form of electromagnetic duality , which inverts the gauge coupling and exchanges the roles of elementary charged particles and magnetic monopoles @xcite .
it is realized in the simplest manner in maximally supersymmetric yang - mills ( sym ) theory @xcite . in this case , the spectrum of dyons in the coulomb phase was shown to exhibit @xmath5 duality @xcite , providing a robust evidence of s - duality .
s - duality in quantum field theories has also been a cornerstone of developing string dualities @xcite . in both qft and string theory ,
s - duality provides valuable insights on the strongly coupled regions of the systems .
s - duality of maximal sym has many implications .
in particular , this duality is related to the existence of 6d superconformal field theories called @xmath2 theories @xcite .
4d maximal sym theories with gauge groups @xmath6 ,
@xmath7 , @xmath8 are realized by compactifying 6d @xmath2 theories on small @xmath3 .
the @xmath5 duality originates from the modular transformation on @xmath3 .
on one side , this relation highlights the far - reaching implications of the 6d cfts to challenging lower dimensional systems .
on the other hand , the 6d cfts lack microscopic definitions , so that s - duality can provide useful clues to better understand the mysterious 6d cfts . in this paper , we study the s - duality of the 6d @xmath2 theories compactified on finite @xmath3 , and use it to explore some interesting properties of these systems .
our key observable of the 6d @xmath2 theory is the partition function of the d0-d4 system .
more precisely , we study the witten index of the quantum mechanical @xmath9 gauge theory for @xmath10 d0-branes bound to @xmath11 separated d4-branes and fundamental open strings , and also study their generating function for arbitrary @xmath10 . in m - theory
, this system is made of @xmath11 m5-branes wrapping @xmath0 .
the d0-d4 systems describe the 6d @xmath2 cft on m5-branes in the sectors with nonzero kaluza - klein momenta . from the viewpoint of 5d super - yang - mills theory on d4-branes , this partition function
is also known as nekrasov s instanton partition function @xcite .
although the 5d yang - mills description of its instanton solitons is uv incomplete , the d0-d4 system provides a uv complete description for computing the instanton partition function .
the results in our paper rely only on the uv complete d0-d4 quantum mechanics .
the d0-d4 index was explored in @xcite . from the 5d viewpoint , this is a partition function on @xmath12 , where @xmath0 is the temporal circle for the witten index . however , with d0-branes ( yang - mills instantons ) providing the kaluza - klein ( kk ) modes of m - theory , there are evidences that this index ( multiplied by a 5d perturbative part ) is a partition function of the 6d @xmath2 scft on @xmath13 .
see @xcite for the @xmath14 theories , and @xcite for the @xmath15 theories . regarding the d0-d4 index as a 6d partition function on @xmath13 , one can naturally ask if it transforms in a simple manner under the s - duality acting on @xmath3 . in this paper
, we establish the s - duality of the prepotential of this index , finding a simple anomaly of s - duality which we can naturally interpret with 6d chiral anomalies .
note that the prepotential @xmath16 is the leading coefficient of the free energy @xmath17 when the so - called omega deformation is taken to zero , @xmath18 .
@xmath19 and @xmath20 are coulomb vev and 5d @xmath21 mass parameter , explained in section 2 .
the anomaly of s - duality takes the following form .
the prepotential @xmath16 can be divided into two parts , @xmath22 , where @xmath23 is related to its s - dual prepotential by a 6d generalization of the legendre transformation .
( see section 2 for details . )
@xmath24 is a simple function which does not obey s - duality , thus named anomalous part .
we find @xmath24 in a closed form in section 2 , which in particular is independent of the coulomb vev @xmath19 .
this finding has two major implications .
firstly , similar result was found for the prepotential @xmath25 of the 4d @xmath26 theory @xcite , related to our prepotential by taking the small @xmath3 limit .
@xmath25 is also given by the sum @xmath27 , where @xmath28 is self s - dual . since @xmath29 is independent of @xmath19 , @xmath25 is s - dual in the seiberg - witten theory , which only sees @xmath19 derivatives of @xmath25 . in our 6d uplift ,
@xmath16 appearing in the seiberg - witten theory also does not see @xmath30 for the same reason .
secondly , the partition function @xmath31 itself is a witten index of the 6d theory on @xmath32 .
so the full prepotential @xmath33 including the @xmath19 independent @xmath24 is physically meaningful , as the leading part of the free energy @xmath34 when @xmath35 are small . at this stage
, we note our key discovery that @xmath24 contains a term proportional to @xmath4 in a suitabe large @xmath11 limit , to be explained in section 2 .
in particular , we further consider the limit in which the kk modes on the circle become light .
this amounts to taking the chemical potential @xmath36 conjugate to the kk momentum ( d0-brane charge ) to be much smaller than the inverse - radius of @xmath0 .
the small @xmath36 limit is the strong coupling limit of the 5d yang - mills theory , or the limit in which the sixth circle decompactifies .
this is the regime in which 6d cft physics should be visible .
the prepotential in this limit can be computed from our anomalous s - duality , since it relates the small @xmath36 ( strong coupling ) regime to the well - understood large @xmath36 ( weak coupling ) regime .
@xmath24 determines the small @xmath36 limit of the free energy , and makes it scale like @xmath4 .
we also show that the term in the asymptotic free energy proportional to @xmath4 is related to the chiral anomaly of the 6d @xmath2 theory , using the methods of @xcite .
these findings show that the number of 5d kk fields for d0-brane bound states grows like @xmath4 , as we decompactify the system to 6d .
the rest of this paper is organized as follows . in section 2.1 and 2.2
, we develop the anomalous s - duality of the prepotential and test it either by expanding @xmath16 in the 5d @xmath21 mass @xmath20 , or by making the ` m - string ' expansion @xcite . in section 2.3 , we discuss the 6d @xmath2 theories of @xmath15 and @xmath8 types . in section 3 ,
we study the high temperature free energy and show that it scales like @xmath4 in a suitable large @xmath11 limit . in section 3.1 ,
we test our result for @xmath37 theory . in section 3.2
, we account for the imaginary part of the asymptotic free energy from 6d chiral anomalies .
section 4 concludes with comments and future directions .
we shall study the witten index of the d0-d4 quantum mechanics , consisting of @xmath10 d0-branes and @xmath11 d4-branes .
this system is a quantum mechanical @xmath9 gauge theory with @xmath38 supersymmetry and @xmath6 global symmetry .
see , for instance , @xcite for the details of this system . here
, we shall only explain some basic aspects .
the bosonic variables consist of four hermitian @xmath39 matrices @xmath40 , two complex @xmath41 matrices @xmath42 , five hermitian @xmath39 matrices @xmath43 , and a quantum mechanical @xmath9 gauge field @xmath44 . here , @xmath45 is the vector index on @xmath46 for the spatial worldvolume of the d4-branes . @xmath47 and @xmath48 are doublet indices of @xmath49 and @xmath50 respectively , which form @xmath51 rotation of @xmath46 .
@xmath52 is the vector index on @xmath53 transverse to the d4-branes .
when @xmath43 are all diagonal matrices , their eigenvalues are interpreted as d0-brane positions transverse to d4-branes .
similarly , when @xmath54 are all diagonal , their eigenvalues are interpreted as d0-brane positions along d4-brane worldvolume . @xmath42
represent internal degrees of freedom .
the bosonic potential energy is given by @xmath55 ^ 2 + \left|\varphi^iq_{\dot\alpha}\right|^2-\frac{1}{4}[\varphi^i,\varphi^j]^2\ , \ ] ] where traces are assumed if necessary , @xmath56 is the @xmath50 triplet index .
@xmath57 are given by @xmath58 \right)\ , \ ] ] where @xmath59 are the pauli matrices .
this system flows in ir to two branches .
classically , they are described by two branches of moduli space satisfying @xmath60 , or @xmath61 , @xmath62=0 $ ] , @xmath63 and @xmath64=0 $ ] .
the first branch is obtained by taking @xmath65 , and @xmath54 , @xmath43 to be diagonal matrices .
the @xmath10 sets of eigenvalues of @xmath66 represent the positions of @xmath10 d0-branes on @xmath67 , unbound to the d4-branes .
the second branch is obtained by taking @xmath68 , and @xmath42 , @xmath54 to satisfy @xmath61 .
after modding out by the @xmath9 gauge orbit , one can show that this branch is described by @xmath69 real parameters .
the two branches meet at @xmath68 , @xmath65 . far away from this intersection ,
each branch is described by a nonlinear sigma model ( nlsm ) on its moduli space .
we are interested in the second branch , describing 6d cft on m5-branes in the sector with @xmath10 units of kk momentum .
the witten index of the second branch can be computed easily by deforming the system by a fayet - iliopoulos ( fi ) parameter , shifting @xmath57 in ( [ adhm ] ) by three constant @xmath70 .
after this deformation , the first branch becomes non - bps , since @xmath65 can not solve @xmath61 with @xmath71 .
so the witten index acquires contributions only from the second branch .
one can understand the second branch from the low energy field theory of d4-branes , the 5d maximal sym theory .
d0-branes are realized in yang - mills theory as instanton solitons , classically described by finite energy stationary solutions of the following bps equation , @xmath72 the finite energy solutions are labeled by the instanton number @xmath10 , defined by @xmath73 we shall consider d0-branes rather than anti - d0-branes , with @xmath74 , equivalently ( [ self - dual ] ) with @xmath75 sign .
@xmath10 corresponds to the rank of the @xmath9 gauge group of the quantum mechanics .
the solutions of @xmath61 , modded out by the @xmath9 gauge orbit , provides the moduli space of self - dual instantons .
the nlsm on the second branch can be obtained by the moduli space approximation of the instanton solitons . however
, this moduli space is known to be singular , having the so - called small instanton singularities . due to these singularities ,
the nlsm description is incomplete , and needs a uv completion .
this is naturally interpreted as inheriting the uv incompleteness of the 5d yang - mills description .
although we do not know how to uv complete the full 5d sym , the nlsm can be uv completed to the @xmath9 quantum mechanics .
we shall study the d0-d4 system in the coulomb phase , with scalar vacuum expectation value ( vev ) . for @xmath6 theories ,
the vev is parametrized by constant @xmath76 diagonal matrices @xmath77 . in the d0-d4 system
, @xmath78 deforms the bosonic potential @xmath79 as @xmath55 ^ 2 -\frac{1}{4}[\varphi^i,\varphi^j]^2+\left| \varphi^iq_{\dot\alpha}-q_{\dot\alpha}\phi^i\right|^2\ .\ ] ] the @xmath11 eigenvalues of @xmath78 correspond to the positions of @xmath11 d4-branes on @xmath53 .
we shall separate the d4-branes along a line , giving nonzero vev to @xmath80 only . in this setting
, we shall study the bps bound states of the d0-branes and the fundamental open strings stretched along the @xmath80 direction , suspended between a pair of d4-branes .
the bound states preserve @xmath81 hermitian supercharges . in 6d @xmath2 theory , we compactify a spatial direction on a circle with radius @xmath82 .
the bps states saturate the bound @xmath83 , where @xmath84 is the energy , and @xmath85 is the quantized momentum on @xmath0 which is @xmath10 in the d0-d4 system .
@xmath86 are the @xmath11 eigenvalues of the scalar @xmath80 , and @xmath87 s are the @xmath88 electric charges in the coulomb branch , satisfying @xmath89 . from the 6d viewpoint , they are the self - dual strings with charges @xmath87 coming from open m2-branes , with @xmath85 units of momenta on them .
we also define @xmath90 , which is the ( dimensionless ) energy on the self - dual strings .
the 6d index is defined by @xmath91\ .\ ] ] here , @xmath92 are two cartans rotating the two 2-planes of @xmath46 , @xmath93 are the cartans of @xmath94 unbroken by the vev of @xmath80 .
the measure is chosen so that it commutes with @xmath95 of the @xmath81 hermitian supercharges preserved by the bps states , or a complex supercharge @xmath96 and its conjugate @xmath97 . see @xcite for the details .
one also finds that @xmath98 .
since only the states saturating the bps bound @xmath99 contribute to the index , @xmath31 is independent of @xmath100 . with @xmath101 understood , the factor @xmath102 weights the bps states with the momentum @xmath85 along the circle .
so @xmath31 can be written as @xmath103 where @xmath104 , and @xmath105 by definition .
@xmath31 can be computed in the weakly coupled type iia regime , in which d0-branes are much heavier than the stretched fundamental strings .
@xmath106 is computed as the nonperturbative witten index of the d0-d4 system with fixed @xmath10 .
@xmath107 comes from the zero modes at @xmath108 , the perturbative open string modes on the d4-branes .
this factor can also be understoood as coming from the perturbative partition function of the 5d maximal sym .
since we are in the weakly coupled regime , @xmath107 can be computed unambiguously from the quadratic part of the yang - mills theory .
although we compute @xmath107 and @xmath106 in this special regime , we naturally expect the result to be valid at general type iia coupling , since this is a witten index independent of the continuous coupling .
@xmath106 and @xmath107 are known for classical gauge groups . for @xmath6
, @xmath106 is given by @xcite @xmath109 where @xmath110 , and @xmath111 the summation is made over @xmath11 young diagrams @xmath112 with total number of boxes @xmath10 , and @xmath113 runs over all boxes of the young diagram @xmath112 .
@xmath114 is the distance from @xmath113 to the right end of the young diagram @xmath112 , and @xmath115 is the distance from @xmath113 to the bottom end of the young diagram @xmath116 .
see @xcite for the details .
one often calls @xmath117 the instanton partition function .
@xmath107 is given by @xcite @xmath118^{\frac{1}{2}}\ ] ] where @xmath119 , and @xmath120 is the barnes gamma function . as noted in @xcite , @xmath47 in the adjoint representation
includes cartans , @xmath121 , for which ` @xmath122 ' in the denominator would diverge . for these @xmath47
, one replaces ` @xmath123 ' factors by @xmath124 $ ] .
see @xcite for more details . for @xmath125 , @xmath126 ,
@xmath107 can be rewritten as @xmath127^{\frac{1}{2}}\ , \ ] ] where prime here again means excluding the zero modes at @xmath128 for the cartans @xmath121 .
the overall factor @xmath129 is given for gauge group @xmath130 by @xcite @xmath131\nonumber\\ \hspace*{-1.2cm}&=&-\frac{\pi i}{2}\sum_{\alpha(v)\in{\bf adj } } \frac{\epsilon_+^2-m^2}{2\epsilon_1\epsilon_2}(1 - 2\alpha(v ) ) = \frac{\pi i(m^2-\epsilon_+^2)}{4\epsilon_1\epsilon_2}|g|\ , \end{aligned}\ ] ] where @xmath132 is the barnes zeta function .
when @xmath133 for all @xmath134 , @xmath107 is rewritten as @xmath135\ , \ ] ] where @xmath136\equiv\exp\left[\sum_{n=1}^\infty\frac{1}{n } f(nx , ny , nz,\cdots)\right]$ ] , @xmath137 , and @xmath138 is the rank of gauge group which is @xmath139 for @xmath6 .
the term @xmath140 in @xmath141 comes from excluding @xmath138 fermionic zero modes for the cartans .
one may multiply an alternative perturbative factor @xmath142^n\hat{z}_{\rm pert}$ ] to @xmath143 , where @xmath144^n$ ] is the perturbative partition function for the @xmath11 cartans , @xmath145 is defined by @xmath146\ , \ ] ] and @xmath147 here all sums are over positive roots @xmath148 .
this expression is well defined when all @xmath149 are positive for positive roots and larger than @xmath150 .
this expression will be useful when studying s - duality from the m - string viewpoint , in section 2.2 .
@xmath107 and @xmath151 are different in subtle ways , which shall not affect the studies of prepotential in this paper but has implications on the s - duality of @xmath31 , which we comment on in section 2.2 .
( [ hat - zpert ] ) has a more natural interpretation as the witten index of charged w - bosons in the coulomb phase @xcite .
however , as an abstract partition function , @xmath107 is more natural as it is manifestly weyl - invariant
. it will also be useful to know the simple structures of the abelian partition function , @xmath152 .
firstly , the perturbative @xmath37 partition function can be written as @xmath153=e^{-\frac{\pi i(m^2-\epsilon_+^2)}{4\epsilon_1\epsilon_2 } } pe\left[\frac{1}{2}\frac{\sinh\frac{m+\epsilon_-}{2 } \sinh\frac{m-\epsilon_-}{2}}{\sinh\frac{\epsilon_1}{2}\sinh\frac{\epsilon_2}{2 } } \right]\ ] ] at @xmath154 , @xmath155 , by following the discussions till ( [ zpert ] ) for @xmath156 .
the instanton part can be written as @xcite @xmath157\ , \ ] ] after summing over all young diagrams in ( [ zk - young ] ) .
given @xmath158 , or @xmath159 , one can write this partition function as @xmath160 \equiv\exp\left[\sum_{n=1}^\infty \frac{1}{n}\frac{f(n\tau , nm , n\epsilon_{1,2},nv ) } { 2\sinh\frac{n\epsilon_1}{2}\cdot 2\sinh\frac{n\epsilon_2}{2}}\right]\ .\ ] ] or a similar expression for @xmath161 using @xmath162 .
the expression appearing in @xmath141 is called the single particle index , containing all the information on the bps bound states .
the coefficients of @xmath163 in fugacity expansion are also called gopakumar - vafa invariants @xcite .
the factor @xmath164 comes from the center - of - mass zero modes of the particle on @xmath46 , which would have caused the path integral for @xmath31 to diverge at @xmath165 .
so @xmath166 also plays the role of ir regulators .
@xmath167 takes into account the relative degrees of freedom of the bound state , in which @xmath166 are just chemical potentials .
in particular , @xmath168 limit is smooth in @xmath163 . in this paper
, we shall mostly discuss the limit @xmath18 . in this limit
, one finds @xmath169 \ \ , \ \
z_{\rm inst}\sim \exp\left[-\frac{f_{\rm inst}(\tau , v , m)}{\epsilon_1\epsilon_2}\right]\ ] ] from ( [ single - particle ] ) .
@xmath170 is the prepotential .
@xmath171 can be obtained from ( [ zk - young ] ) after a straightforward but tedious calculation .
@xmath172 can be obtained from ( [ zpert ] ) , which is given by @xmath173 where @xmath174 for @xmath175 , and can be continued to the complex @xmath176 plane with a branch cut . the first term coming from @xmath129
will play no role in this paper .
one way of obtaining ( [ fpert ] ) is to first take @xmath177 to be purely imaginary , to guarantee convergence of the sum in ( [ zpert ] ) , and take the limit @xmath168 to obtain ( [ fpert ] ) .
then , ( [ fpert ] ) can be analytically continued to complex @xmath177 .
one may alternatively start from @xmath151 and obtain its prepotential , @xmath178 @xmath179 is the set of positive roots . here , from the identity @xmath180 for @xmath181 and @xmath182 , where @xmath183 s are bernoulli polynomials , one finds @xmath184 for @xmath185 and @xmath186 .
so for simplicity , let us assume @xmath187 for all positive roots @xmath47 , and also @xmath188 is chosen such that all @xmath189 are within the range @xmath190 $ ] for positive roots .
then one finds @xmath191 = -\frac{\pi im^2}{2}|\delta_+|\ ] ] where we used @xmath192 .
so at least in this setting , @xmath172 and @xmath193 differ only by a trivial constant independent of @xmath194 .
the last constant will play no role in this paper .
it will be helpful to consider the prepotential of the @xmath37 theory separately . from ( [ abelian - pert ] ) and ( [ abelian - inst ] ) , the prepotential @xmath195 for the @xmath37 theory is given by @xmath196 for studying the s - duality of this prepotential , it will be useful to make an expansion of @xmath197 in @xmath20 .
one first finds that the instanton part is given by @xmath198 where @xmath199 is the euler function , and we used the identity @xmath200 for the eistenstein series @xmath201 .
@xmath202 are the bernoulli numbers : @xmath203 , @xmath204 , @xmath205 and so on . the perturbative prepotential can be expanded in @xmath20 by using @xmath206 at small @xmath207 , with @xmath208 .
one finds @xmath209 combining all , one obtains @xmath210 this will be useful later for understanding @xmath211 , as a part of the @xmath6 prepotential .
one can understand the chemical potentials from the viewpoint of the 4d effective action in the coulomb branch .
the dimensionless variables @xmath150 , @xmath194 take the form of @xmath212 where @xmath213 is the radius of the temporal circle of @xmath12 .
@xmath214 is the mass deformation parameter of the 4d @xmath26 yang - mills theory , or the 5d @xmath21 theory .
( more precisely , @xmath214 is @xmath215 times the mass . )
@xmath216 are the omega deformation parameters which have dimensions of mass .
@xmath19 is the coulomb vev of the scalar field @xmath80 .
@xmath217 is identified as @xmath218 where @xmath82 is the radius of the sixth circle .
this is the inverse gauge coupling in 4d .
@xmath217 can be complexified with a real part , given by the rr 1-form holonomy of type iia theory on @xmath0 .
the 4 dimensional limit of the partition function is obtained by taking @xmath219 with fixed @xmath220 . from ( [ zk - young ] )
, one finds that all @xmath221 functions of @xmath222 are replaced by linear functions of @xmath223 , and the @xmath213 dependences cancel between numerator and denominator . as a result
, the 4d limit @xmath224 of the instanton partition function is given by a rational function of @xmath225 of degree @xmath226 .
this makes @xmath227 and @xmath228 to enjoy a simple scaling property , @xmath229 this will be used in section 2.1 to provide two interpretations of the 4d s - duality , and extend one version to 6d .
as for the perturbative part @xmath172 , one can use ( [ li - expand ] ) to obtain @xmath230 .
one finds @xmath231\ ] ] where the first term independent of the coulomb vev is unphysical in the seiberg - witten theory .
the perturbative prepotential satisfies the following pseudo - scaling property , @xmath232 which is homogenous and degree @xmath95 up to a coulomb vev independent shift .
@xmath143 or @xmath171 are only known as @xmath233 expansion when @xmath234 , or @xmath235 .
this is useful when the ` temperature ' is much smaller than the kaluz - klein scale @xmath236 , when the kk modes are ` heavy . '
however , to study 6d scft , it is more interesting to explore the regime @xmath237 , or @xmath238 , in which case the circle effectively decompactifies .
the two regimes are weakly coupled and strongly coupled regimes , respectively .
so if there is s - duality for the partition function on @xmath13 , it will be helpful to study the interesting decompactifying regime from the well - understood region @xmath235 .
developing the s - duality of the prepotential @xmath16 is the goal of this section .
( in section 2.2 , we also comment on the s - duality of the full partition function . ) following @xcite , we review the basic set up for studying the s - duality of 4 dimensional prepotential , and extend it to the 6d theory on @xmath3 .
the prepotential @xmath16 of general 4d @xmath239 gauge theory determines the effective action in the coulomb branch .
the magnetic dual description uses the dual coulomb vev @xmath240 and the dual prepotential @xmath241 , defined by the following legendre transformation ,
@xmath242(a)\equiv f(a)-2\pi ia_da = f - a\frac{\partial f}{\partial a}\ .\ ] ] for theories with higher rank @xmath243 , @xmath19 has many components , @xmath244 with @xmath245 .
expressions like @xmath246 should be understood with contracted @xmath247 indices , i.e. @xmath248 , whose sum structures will not be explicitly shown to make the notations simpler . for
generic @xmath239 theories , @xmath249 depend on other parameters like hypermultiplet masses and the coupling constant ( or the dynamically generated scale @xmath250 instead of the coupling ) . for 4d @xmath26 theory
, the prepotential @xmath25 ( to be distinguished with the 6d prepotential @xmath16 which we shall consider later ) depends on the microscopic coupling constant @xmath217 and the adjoint hypermultiplet mass @xmath214 .
the prepotential can be divided into the classical , perturbative , and instanton contributions , @xmath251 where @xmath252 , and @xmath253 .
@xmath254 is the quantum prepotential . to study self s - dual theories , it is convenient to define @xmath255 as a function of the dual coupling @xmath256 . for the 4d @xmath26 theory ,
@xmath255 is defined by @xmath257(\tau , a , m ) = f^{\rm 4d}(\tau , a , m)-a\frac{\partial f^{\rm 4d}}{\partial a}(\tau , a , m)\ .\ ] ] then , self s - duality exists if @xmath255 and @xmath25 are same function , @xmath258 .
this s - duality has been tested in detail in @xcite .
more precisely , it was found that @xmath259 where @xmath260 satisfies @xmath261 and @xmath29 is an anomalous part of s - duality , depending on @xmath262 but is independent of the coulomb vev @xmath19 @xcite . since the coulomb branch effective action is obtained by taking @xmath19 derivatives of @xmath25 , @xmath25 and @xmath28 are identical in the seiberg - witten theory .
this establishes the s - duality of the 4d @xmath26 theory in the coulomb branch effective action .
let us rephrase the 4d s - duality in a way that is suitable for 6d extension .
@xmath263 satisfies the scaling property ( [ 4d - scaling ] ) . combining the perturbative part
, one finds @xmath264 applying this to @xmath265 , one obtains @xmath266 so the left hand side of ( [ 4d - s - dual ] ) can be written as @xmath267 let us consider the structure of @xmath268 . since the prepotential has mass dimension @xmath95 , one may think that its @xmath214 dependence is simply @xmath269 .
however , the perturbative part ( [ 4d - fpert ] ) shows that there is a term @xmath270 in @xmath25 which scales in an odd manner . in the computational framework of @xcite , which we shall explain below in our 6d version
, @xmath260 is by construction taken to be a series expansion in @xmath269 .
this means that the odd term @xmath270 should have been put in @xmath268 .
therefore , had one been doing the calculation of @xcite using ( [ 4d - fpert ] ) as the perturbative part , one would have found that @xmath271 , where @xmath272 only depends on @xmath217 . using this structure , ( [ f - s - dual - scaling ] )
can be rewritten as @xmath273 so defining @xmath274 one finds that @xmath275 satisfies @xmath276 instead of ( [ 4d - s - dual ] ) .
to summarize , by trivially redefining @xmath260 and @xmath29 by the last term of ( [ f - s - dual - redefine ] ) , one can reformulate the standard s - duality ( [ 4d - s - dual ] ) as ( [ 4d - s - dual-2 ] ) . only (
[ 4d - s - dual-2 ] ) will naturally generalize to the s - duality on @xmath13 .
now we seek for the s - duality of the 6d prepotential .
note that in 4d , ( [ 4d - s - dual ] ) and ( [ 4d - s - dual-2 ] ) are equivalent by making a minor redefinition of @xmath29 , using ( [ homogeneous ] ) . in 6d
, a property like ( [ homogeneous ] ) does not hold . before making a quantitative study of the 6d s - duality
, we first explain that ( [ 4d - s - dual-2 ] ) is more natural in 6d . to discuss the 6d prepotential , it is convenient to work with the dimensionless parameters @xmath277 .
firstly , in the 6d theory compactified on @xmath3 , the complex mass parameter @xmath20 is simply the holonomy of the background gauge field for @xmath278 global symmetry , along the two sides of @xmath3 .
then after making an s - duality of the torus , exchanging two sides of @xmath3 , one naturally expects @xmath279 .
let us briefly review this by taking a rectangular torus , for simplicity . in this case ,
the complex structure @xmath217 of the torus is purely imaginary .
@xmath217 is related to the two radii of @xmath3 by @xmath218 where @xmath82 is the radius of the circle which compactifies the 6d theory to 5d sym , and @xmath213 is the radius of another circle which compactifies the 5d theory to 4d .
the s - duality transformation exchanges @xmath280 .
so the dual complex structure is @xmath281 .
more precisely , s - duality rotates the torus by @xmath282 degrees on a plane .
it also transforms the two @xmath278 holonomies along the two circles .
let @xmath283 be the holonomy on the circle with radius @xmath213 , and @xmath284 that on the circle with radius @xmath82 . under s - duality
, one finds @xmath285 , @xmath286 .
so one finds @xmath287 . in @xmath16
, @xmath214 appears in the dimensionless combination @xmath288 , which transforms as @xmath289 the final result holds for complex @xmath217 .
similar property holds for @xmath290 , i.e. @xmath291 .
this makes the appearance of @xmath292 to be more natural on the left hand side of ( [ 4d - s - dual-2 ] ) .
secondly , let us discuss how @xmath19 should transform .
in 4d , we already stated that @xmath293 naturally appears on the left hand side of ( [ 4d - s - dual ] ) . for simplicity ,
let us discuss these variables in the limit of large coulomb vev , @xmath294 , @xmath295 .
the second term can be ignored in this limit , yielding the semi - classical result @xmath296 . in this limit
, we shall discuss what is the natural s - dual variable using the abelian 6d @xmath2 theory .
in 4d , @xmath296 is a natural aspect of s - duality being electromagnetic duality .
also , it makes sense to multiply @xmath19 by a complex number @xmath217 , since @xmath19 is a complex variable living on a plane .
however , in 6d cft on @xmath3 , @xmath19 lives on a cylinder .
the real part of @xmath19 is the vev of the real scalar in the 6d self - dual tensor multiplet , which is noncompact . on the other hand ,
the imaginary part of @xmath19 comes from the holonomy of the 2-form tensor field @xmath297 on @xmath3 , implying that it is a periodic variable .
so it does not good make sense to rotate @xmath19 living on a cylinder by complex @xmath217 .
more precisely , the 6d scalar @xmath298 and the 5d scalar @xmath19 are related by @xmath299 .
so one finds @xmath300 where @xmath301 and @xmath95 denote two directions of @xmath3 .
thus , @xmath302 is invariant under @xmath280 , meaning that it makes more sense to set @xmath303 in the limit @xmath304 . using the dimensionful variables ,
this requires one to use @xmath305 as the dual variable , instead of @xmath306 .
this does not rotate the variable @xmath19 by a complex number , so makes better sense in 6d .
incidently , we have already found the alternative ( but equivalent ) statement ( [ 4d - s - dual-2 ] ) of s - duality which uses @xmath307 as the dual variable , instead of @xmath308 .
note that the usage of @xmath309 is valid even beyond the semi - classical limit @xmath310 .
thus , in the 6d uplift , it is natural and consistent to regard @xmath311 . as the dual variable .
appearing on the right hand side should have been @xmath312 .
however , we shall define the prepotential as the coefficient of the dimensionless @xmath313 , @xmath314 , rather than @xmath315 that is conventional in the seiberg - witten theory , making @xmath163 dimensionless .
namely , @xmath316 in 6d is related to the conventionally normalized prepotential by @xmath317 . ]
so it appears natural to seek for a 6d generalization of ( [ 4d - s - dual-2 ] ) rather than ( [ 4d - s - dual ] ) .
this is what we shall establish in the rest of this section .
namely , we shall find that the 6d prepotential is divided into two , @xmath318 where @xmath319 , @xmath320 , and @xmath24 is independent of the coulomb vev .
@xmath23 satisfies @xmath321 we have some freedom to choose @xmath24 , by adding / subtracting @xmath194 independent s - dual expressions to @xmath24 , @xmath23 .
we shall explain that one can choose @xmath24 as @xmath322 where @xmath197 is the @xmath37 prepotential ( [ abelian - f ] ) .
the first term @xmath211 comes from the @xmath11 6d abelian tensor multiplets in @xmath88 , which has their own s - duality anomaly .
the second term of @xmath24 is one of the key findings of this paper , which comes from the charged part of the partition function . after replacing @xmath323 , and multiplying @xmath324 to the above @xmath24 to get to the conventionally normalized prepotential
, one can take the 4d limit of @xmath24 .
the second term proportional to @xmath325 vanishes in the 4d limit @xmath219 , as it is proportional to @xmath326 . with the motivations and results given , we now properly set up the calculation and show the claims made above . as in 4d
, we decompose the 6d prepotential as @xmath327 where @xmath328 .
the prepotential is s - dual if it satisfies @xmath329 we first study the structures of this equation , before showing that it is satisfied by our @xmath23 .
firstly , replacing @xmath16 by @xmath330 , one can check that s - duality trivially holds at the classical level : @xmath331 = -\pi i\tau v^2 = f_{\rm cl}(\tau , v)-v\frac{\partial f_{\rm cl}}{\partial v}(\tau , v)\ , \ ] ] where @xmath332 is replaced by its classical value @xmath333 ( formally at @xmath334 ) .
now we subtract ( [ s - dual-5d ] ) by ( [ s - dual - classical ] ) to find the following condition for the quantum prepotential @xmath163 : @xmath335 we are going to study the last equation .
note again that the effective action in the coulomb branch only contains @xmath194 derivatives of @xmath16 , or @xmath163 .
thus , in seiberg - witten theory , @xmath163 is ambiguous by addition of @xmath194 independent functions , possibly depending on @xmath217 and @xmath20 .
however , the s - duality requirement ( [ s - dual - requirement ] ) is sensitive to the value of @xmath163 , including the @xmath194 independent part .
so when one tries to establish the s - duality of the coulomb branch effective action , one should have in mind that one may have to add suitable coulomb vev independent terms to @xmath163 computed microscopically from @xmath31 .
following @xcite , we shall establish the s - duality ( [ s - dual - requirement ] ) and its anomaly ( [ s - dual - anomaly ] ) by expanding @xmath163 in the mass @xmath20 when it is small enough .
we shall still get an exact statement ( [ s - dual - anomaly ] ) , which we check for certain orders in @xmath20 .
one should however have in mind that the exact statement ( [ s - dual - anomaly ] ) may be valid only within a finite region of @xmath336 in the complex planes . in section 2.2
, we shed more lights on the exactness of ( [ s - dual - anomaly ] ) , by making an m - string expansion @xcite .
as studied in the 4d limit @xcite , there is a natural way of achieving the s - duality requirement ( [ s - dual - requirement ] ) .
this is to require that @xmath163 is expanded in quasi - modular forms of suitable weights . to precisely explain its meaning , we first expand @xmath163 in @xmath20 as @xmath337
this series makes sense as follows .
firstly , the @xmath338 limit exhibits enhanced maximal supersymmetry .
so at @xmath339 , the classical prepotential @xmath340 acquires no quantum corrections , meaning that @xmath163 vanishes at @xmath339 .
also , the prepotential is an even function of @xmath20 , which restrict the expansion as above . in the perturbative part , which is easiest to see from the 4d limit ( [ 4d - fpert ] ) .
however , we shall expand @xmath341 as ( [ m - series ] ) , while the term @xmath342 is moved to @xmath24 . ] then , following @xcite , we require that @xmath343 is a quasi - modular form of weight @xmath344 , which means the following .
quasi - modular forms are polynomials of the first three eisenstein series @xmath345 , @xmath346 , @xmath347 , where each series has weight @xmath348 respectively under s - duality in the following sense : @xmath349 more concretely , they are given by @xmath350 higher eisenstein series @xmath351 are polynomials of @xmath352 with weight @xmath353 . to study the quasi - modular property , it is helpful to decompose their dependence on @xmath217 into the dependence through @xmath345 and the dependence through @xmath352 .
we thus write @xmath354 , where the @xmath217 dependence through @xmath345 is explicitly shown . a weight @xmath344 quasi - modular form @xmath343 satisfies @xmath355 where @xmath356 . in terms of @xmath163 ,
this is equivalent to @xmath357 we now investigate how quasi - modularity is related to the s - duality ( [ s - dual - requirement ] ) .
one can make ( [ quasi - modular ] ) to be equivalent to ( [ s - dual - requirement ] ) by specifying the @xmath345 dependence of @xmath163 , which we now turn to .
let us first try to find the desired @xmath345 dependence , by requiring both ( [ s - dual - requirement ] ) and ( [ quasi - modular ] ) .
by applying ( [ quasi - modular ] ) to @xmath358 , one obtains @xmath359 where again recall that @xmath356 . combining this with ( [ s - dual - requirement ] ) , one obtains @xmath360 we want to make this equation to hold , by specifying a particular @xmath345 dependence of @xmath163 .
@xcite showed that the desired @xmath345 dependence is @xmath361 for the sake of completeness , we repeat the logics presented in @xcite and expand it to make a proof . in fact , we shall make a stronger claim than needed .
namely , we need to find the @xmath345 dependence of @xmath163 which guarantees ( [ s - dual - v2 ] ) only at @xmath362 .
however , we shall show that ( [ modular - anomaly ] ) guarantees ( [ s - dual - v2 ] ) for arbitrary independent parameter @xmath363 , and then set @xmath362 later . as a warm - up , we follow @xcite to make a series expansion of the left hand side of ( [ s - dual - v2 ] ) in small @xmath363 , to see how ( [ modular - anomaly ] ) guarantees ( [ s - dual - v2 ] ) at low orders .
one finds that @xmath364 so at @xmath365 and @xmath366 orders , one finds that it agrees with the right hand side if ( [ modular - anomaly ] ) is met . now assuming ( [ modular - anomaly ] )
, we consider whether ( [ s - dual - v2 ] ) is satisfied in full generality . to this end , we take @xmath363 derivative of both sides of ( [ s - dual - v2 ] ) , at fixed @xmath367 , @xmath368 where for simplicity , we defined @xmath369 note that at the second step of the first equation in ( [ derivative - s - dual ] ) , we used @xmath370 which follows from ( [ modular - anomaly ] ) . if ( [ s - dual - v2 ] ) holds for general @xmath363 , its first derivative would yield @xmath371 which one can show by using ( [ derivative - s - dual ] ) . on the other hand , ( [ derivative - s - dual - v2 ] ) together with the @xmath372 component of ( [ s - dual - v2 ] )
is equivalent to ( [ s - dual - v2 ] ) , since the @xmath372 component is the only information lost by taking @xmath363 derivative .
however , we have already shown around ( [ zeroth - first ] ) that the @xmath372 component of ( [ s - dual - v2 ] ) is satisfied . therefore , showing ( [ derivative - s - dual - v2 ] ) will be equivalent to showing ( [ s - dual - v2 ] ) .
so will show ( [ derivative - s - dual - v2 ] ) by assuming ( [ modular - anomaly ] ) .
we take @xmath363 derivative of @xmath373 at fixed @xmath374 . again using ( [ modular - anomaly ] ) , one obtains @xmath375 so if @xmath373 is zero at a particular value of @xmath363 , ( [ 2nd - derivative ] ) guarantees that it is zero at different values of @xmath363 .
since we already checked around ( [ zeroth - first ] ) that ( [ s - dual - v2 ] ) is true up to @xmath376 , we have shown that ( [ derivative - s - dual - v2 ] ) holds at @xmath372 , or that @xmath377 at @xmath378 .
this establishes that ( [ modular - anomaly ] ) implies ( [ derivative - s - dual - v2 ] ) , and in turn that ( [ modular - anomaly ] ) implies ( [ s - dual - v2 ] ) . finally , we insert @xmath362 . to summarize till here , ( [ s - dual - v2 ] ) holds if @xmath163 satisfies ( [ modular - anomaly ] ) .
but ( [ s - dual - v2 ] ) and ( [ quasi - modular ] ) implies the s - duality relation ( [ s - dual - requirement ] ) .
therefore , s - duality requirement ( [ s - dual - requirement ] ) is satisfied if @xmath163 satisfies the quasi - modular property ( [ quasi - modular ] ) and the modular anomaly equation ( [ modular - anomaly ] ) . in the rest of this subsection
, we shall discuss the last two equations .
following and extending @xcite , we show that the prepotential @xmath163 obeys the two properties ( [ quasi - modular ] ) , ( [ modular - anomaly ] ) , up to an anomalous part which is independent of the coulomb vev @xmath194 . again following @xcite
, our strategy is to first find a prepotential @xmath341 in a series of @xmath379 which satisfies both ( [ quasi - modular ] ) and ( [ modular - anomaly ] ) .
then we show that @xmath380 is independent of @xmath194 .
we expand @xmath341 like ( [ m - series ] ) , @xmath381 .
( [ modular - anomaly ] ) is given in terms of @xmath343 by @xmath382 for @xmath383 .
this equation can be used to recursively compute @xmath343 .
namely , once we know @xmath384 for @xmath385 , one can integrate the right hand side of ( [ modular - anomaly - rec ] ) with @xmath345 to get @xmath343 , up to an integration constant independent of @xmath345 .
the integration constant is a polynomial of @xmath346 and @xmath347 with modular weight @xmath344 , whose coefficients depend only on @xmath194 .
these integration constants depending on @xmath194 can be fixed once we know a few low order coefficients of @xmath163 in @xmath233 expansion . also , to start the recursive construction , the first coefficient @xmath386 at @xmath379 should be known .
it will turn out that this can be also fixed by the known perturbative part @xmath387 @xcite .
this way , one can recursively generate the coefficients of @xmath341 from ( [ quasi - modular ] ) , ( [ modular - anomaly ] ) and the knowledge of the few low order coefficients of @xmath163 in @xmath233 expansion .
we emphasize here that our purpose of making a recursive construction of @xmath341 is to show that the coulomb vev dependent part of @xmath163 is s - dual .
therefore , while fixing the integration constants and @xmath386 in @xmath341 by using the low order @xmath233 expansion coefficients of @xmath163 , it suffices to use @xmath163 up to the addition of any convenient expression independent of @xmath194 .
so for technical reasons , we shall fit these integration constants and @xmath386 by comparing @xmath341 with @xmath388 rather than @xmath163 itself .
note that @xmath211 is the prepotential contribution from @xmath88 cartan part , coming from d0-branes bound to d4-branes but unbound to w - bosons which see @xmath194 .
one reason for comparing with @xmath389 is that @xmath197 does not admit a power series expansion in @xmath379 like ( [ m - series ] ) .
the s - duality anomaly of @xmath211 can be calculated separately from ( [ anom - series ] ) . with these understood ,
we start the recursive contruction by determining @xmath386 .
this can be fixed solely from the perturbative part of @xmath390 @xcite .
namely , when instantons are bound to w - bosons , there are fermion zero modes which provide at least a factor of @xmath391 in @xmath163 .
this means that @xmath379 term @xmath386 should come from the perturbative part only .
this fact can also be straightforwardly checked from the microscopic calculus .
so one finds @xmath392 where @xmath393 is the set of roots of @xmath6 .
one can then compute @xmath394 using ( [ modular - anomaly - rec ] ) at @xmath395 , @xmath396 one can integrate it with @xmath345 , to obtain @xmath397 there is no integration constant at weight @xmath95 . to proceed
, we study the properties of the @xmath6 roots .
@xmath393 consists of vectors of the form @xmath398 , @xmath399 , @xmath400 , where @xmath401 are @xmath11 orthogonal unit vectors .
@xmath402 takes following possible values , @xmath403 where @xmath404 is given for @xmath405 by @xmath406 for a given @xmath47 , there are @xmath407 elements of @xmath404 .
using this , one finds @xmath408\nonumber\\ \hspace*{-1cm}&=&-\frac{e_2(\tau)}{96}\left [ \sum_{\alpha\in\delta}\frac{4}{(1-e^{\alpha(v)})^2 } -\sum_{\alpha\in\delta}\frac{2(n-1)}{1-e^{\alpha(v ) } } + \sum_{\alpha\in\delta}\sum_{\beta\in\psi(\alpha ) } \frac{2}{(1-e^{\alpha(v)})(1-e^{\beta(v)})}\right]\nonumber\end{aligned}\ ] ] where we used @xmath409 . using @xmath410 ,
the first term can be rewritten so that @xmath411\ .\ ] ] here , one can simplify the second term by using @xmath412 also , using @xmath413 , one finds @xmath414\nonumber\\ & & \hspace{1cm}=\frac{1}{3}\sum_{i\neq j\neq k \neq i}\left[(i , j , k)+(j , k , i)+(k , i , j)\right ] = \sum_{i\neq j\neq k\neq i}\frac{2}{3}=\frac{2}{3}n(n-1)(n-2)\ , \end{aligned}\ ] ] where at the second step we symmetrized the summand by making a cyclic permutation of @xmath415 .
this simplifies the third term .
one thus finds @xmath416\ ] ] at @xmath417 order . before proceeding to higher order coefficients
@xmath343 with @xmath418 , let us first discuss @xmath394 that we computed by requiring s - duality of @xmath341 . note that at @xmath391 , we have obtained an all order result in the instanton expansion , coming from @xmath419
so from the microscopic instanton calculus , one can expand @xmath420 in small @xmath20 , and we can compare @xmath163 and @xmath341 at @xmath391 order .
we find that @xmath421 which we checked till @xmath422 order for general @xmath11 , and till @xmath423 for @xmath424 .
therefore , we find that the microscopic prepotential is compatible with s - duality at @xmath391 order , up to the addition of an ` anomalous ' term on the right hand side independent of the coulomb vev
. one can make further recursive calculations of @xmath343 for @xmath418 , using ( [ modular - anomaly - rec ] ) , and test the consistency of @xmath341 with our microscopic @xmath163 .
the next recursion relation of ( [ modular - anomaly - rec ] ) is @xmath425 knowing @xmath426 , one can integrate ( [ f3-rec ] ) to obtain @xmath427+c_3(v)e_4(\tau)\ .\ ] ] the integration constant @xmath428 can be determined by expanding @xmath429 in @xmath233 , and comparing the @xmath430 order with the perturbative contribution @xmath390 at @xmath431 order .
one obtains @xmath432 inserting this @xmath428 in ( [ f3-result ] ) , one can further study the higher order coefficients of @xmath429 in @xmath233 expansion , against the microscopic result @xmath163 .
we find that @xmath433 which we checked till @xmath422 order for general @xmath11 , and till @xmath423 order for @xmath424 . integrating ( [ modular - anomaly - rec ] ) to get higher @xmath343 s
, the integration constants take the following form , @xmath434 more concretely , one would get @xmath435 and so on . to fix the coefficients @xmath436 , one should use some low order data of @xmath389 .
if there are @xmath437 independent @xmath438 s , one should use up to @xmath10 instanton coefficients of @xmath389 to fix them . then from @xmath437 or higher instantons
, one can test @xmath343 by comparing with @xmath389 . for general @xmath11
, we tested @xmath341 against @xmath389 till @xmath431 and @xmath422 orders , as already reported above . for @xmath439 , we tested it till @xmath440 ( i.e. @xmath441 ) and @xmath423 orders . for @xmath442 , we tested it till @xmath443 ( i.e. @xmath444 ) and @xmath423 orders .
the result is that there is no further difference between @xmath341 and @xmath389 from @xmath431 and higher orders .
namely , we find that @xmath445 the s - duality transformation of @xmath197 can be separately derived from ( [ anom - series ] ) , which is @xmath446+\frac{m^4}{48\pi i\tau}\ , \end{aligned}\ ] ] where @xmath447 .
although we checked ( [ s - dual - summary ] ) in a serious expansion in small @xmath20 , we believe it is an exact property , valid for finite @xmath20 . in section 2.2 , we shall provide another test of ( [ s - dual - summary ] ) by keeping @xmath20 finite , but expanding @xmath163 as a series of @xmath448 at positive coulomb vev @xmath194 .
also , in section 2.3 , we extend ( [ s - dual - summary ] ) to all ade theories . in this subsection
, we provide another derivation of the s - duality of prepotential
. the analysis here will provide more evidence for our s - duality anomaly at finite @xmath20 .
it may also provide strong hints towards the s - duality of the partition function @xmath449 at finite @xmath166 , but we postpone more extensive discussions on @xmath31 to a separate project .
we use an alternative partition function @xmath450 .
this differs from @xmath31 by a different perturbative partition function @xmath151 . for the purpose of studying the prepotential in the limit @xmath168
, we already saw that they differ only by a trivial term , as shown in ( [ pert - difference ] ) in a suitable regime on the complex @xmath194 plane .
the last term of ( [ pert - difference ] ) is separately invariant under the s - duality transformation @xmath451 , and is independent of @xmath194 , so it is completely fine to drop this constant factor when discussing the s - duality of prepotential .
@xmath161 can be expanded in @xmath452 when @xmath453 are positive and sufficiently large , where @xmath454 ( @xmath455 ) are the simple roots of @xmath456 .
the expansion takes the form of @xcite @xmath457 @xmath458 is for the d0-branes unbound to the w - bosons , just bound to one of the @xmath11 d4-branes , which is this independent of @xmath452 .
@xmath459 is defined at the beginning of section 2 by @xmath460 this factor has to be multiplied to guarantee ( [ pert - difference ] ) .
this expansion is reliable in a particular weyl chamber of the coulomb branch , with sufficiently large @xmath453 for all simple roots .
the coefficient @xmath461 at given self - dual string numbers @xmath462 is computed from ` m - strings , ' which are m2-branes suspended between separated m5-branes .
@xmath463 is the elliptic genus of the 2d quiver gauge theory with @xmath464 gauge group .
see @xcite for the details .
@xmath461 is given by @xcite ( we follow the notations of @xcite ) @xmath465 where @xmath466 denotes the position of each box in the young diagram , and @xmath467 @xmath468 is the length of the @xmath19th row of @xmath112 , and @xmath469 is the length of the @xmath470th column of @xmath116 .
@xmath471 and @xmath472 are empty by definition .
we study the s - duality of @xmath161 in this setting . by using @xmath473
one can show that @xmath463 transforms as @xmath474 z_{(n_i)}(\tau , m,\epsilon_{1,2})\ .\ ] ] @xmath475 is the cartan matrix for @xmath456 , given by @xmath476 , @xmath477 and @xmath478 otherwise .
this transformation can be expressed as a modular anomaly equation , as follows .
first note that jacobi s theta functions can be expressed as @xmath479\ .\ ] ] @xmath217 dependence of @xmath463 can be understood as its dependence through @xmath201 , since @xmath480 factors cancel out in ( [ m - string - elliptic ] ) . as we did in section 2.1 for the prepotential
, the @xmath217 dependence of @xmath463 can again be decomposed into its dependence through @xmath481 , and the rest . to emphasize this , we write @xmath482 . in this setting
, the modular anomaly @xmath483 of ( [ theta - s - dual ] ) appears due to the dependence of @xmath484 on @xmath481 , which is not modular .
therefore , we can rephrase ( [ m - string - s - dual ] ) as @xmath485 let us consider @xmath486 , where we defined @xmath487 with the choice of simple roots @xmath488 . here
we view @xmath489 as a function of @xmath490 s , since it depends only on the differences of @xmath194 s . now with the replacement @xmath491 in the laplace transformation , one obtains @xmath492 this is essentially the ` holomorphic anomaly equation , ' eqn.(3.54 ) of @xcite . in our context
, we continue to study it as the modular anomaly equation .
we shall get better intuitions on the modular anomaly equation .
consider @xmath493 with @xmath494 , still without @xmath495 included .
since the @xmath194 independent part of @xmath459 is independent of @xmath496 , it does not affect ( [ modular - anomaly - m-1 ] ) .
the @xmath194 dependent part can be written as @xmath497v_i=\sum_{i=1}^n(n+1 - 2i)v_i\ .\end{aligned}\ ] ] since @xmath498 , @xmath499 acting on a function of @xmath500 s can be written as @xmath501 , @xmath502 the laplacian of @xmath194 acting on a function of @xmath490 is given by @xmath503 one also finds that @xmath504 for any number @xmath19 .
acting @xmath505 on @xmath506 , one obtains @xmath507\end{aligned}\ ] ] where we used @xmath508 . using ( [ modular - anomaly - m-1 ] ) ,
this can be rewritten as @xmath509\ .\ ] ] thus , one finds that the partition function @xmath510\tilde{z } = \exp\left[\frac{n^3-n}{288 } \frac{(m^2-\epsilon_+^2)^2}{\epsilon_1\epsilon_2}e_2 -\varepsilon_0\right]\hat{z}\ ] ] satisfies the modular anomaly equation @xmath511 also , from the m - string expansion form of @xmath489 , and the the form of the prefactors we multiplied to define @xmath512 , @xmath512 satisfies the following quasi - modularity condition , @xmath513 where @xmath362 .
( [ z - modular - anomaly ] ) and ( [ z - quasi - modular ] ) are the two main properties of @xmath512 . using ( [ z - modular - anomaly ] ) and
( [ z - quasi - modular ] ) , we would like to study the relation between @xmath514 and @xmath515 .
since @xmath516 is related to @xmath517 in a simple manner , answering this question will tell us how @xmath516 transforms under s - duality . then , since we already understand how the @xmath88 part @xmath458 transforms under s - duality @xcite , we shall in turn know the s - duality transformation of @xmath161 . using ( [ z - quasi - modular ] )
, we should understand how @xmath518 and @xmath519 are related to each other , at same values of @xmath520 but with a shift of @xmath345 by @xmath362 .
let us formally regard @xmath345 as time variable , and @xmath512 as a wavefunction on the space formed by @xmath86 .
then ( [ z - modular - anomaly ] ) takes the form of heat equation , or euclidean schrdinger equation for a free particle .
more precisely , taking @xmath217 to be purely imaginary for simplicity , @xmath363 is real and negative .
so @xmath521 plays the role of time , and we would like to understand the time evolution @xmath522 by @xmath523 following the heat equation . the heat equation comes with proper sign when @xmath524 . in this case , the time evolution is described by evolving @xmath525 by the gaussian heat kernel as @xmath526 where @xmath527\ ] ] is the heat kernel which approaches @xmath528 when @xmath529 . when @xmath530 , the ` time evolution ' from @xmath525 to @xmath522 is described by the time reversal of the heat equation .
therefore , one finds @xmath531 with @xmath532 $ ] for @xmath530 .
the s - duality of @xmath341 can be studied from ( [ z - s - dual - gaussian ] ) or ( [ z - s - dual - inverse ] ) by a saddle point approximation of the @xmath533 integral at @xmath168 . using both equations yield idential results .
one finds @xmath534\ ] ] in the limit @xmath168 , so that @xmath535 we shall show that @xmath341 defined this way is the same @xmath341 defined and computed in section 2.1 .
firstly , note that @xmath536 from the relations @xmath537^n\hat{z}$ ] and @xmath538 . inserting this in ( [ f - s - dual - m ] ) , one obtains @xmath539 this is completely the same as the relation between @xmath163 and @xmath341 that we found in section 2.1 , except the last term on the right hand side .
however , we know that the last term comes from using slightly different perturbative partition function in @xmath161 , ( [ pert - difference ] ) . also this term can be completely ignored for studying s - duality since it is separately invariant under s - duality .
therefore , after discarding this last term , we find that @xmath341 is the same as @xmath341 defined in section 2.1 .
now one can independently check that @xmath341 is s - dual .
firstly , since @xmath512 is quasi - modular , so is @xmath341 , i.e. @xmath540 secondly , inserting @xmath541 $ ] into ( [ z - modular - anomaly ] ) and keeping the leading terms in the limit @xmath168 , one finds @xmath542 , the same modular anomaly equation that we studied in section 2.1 @xcite .
the last two equations guarantee the s - duality of @xmath341 , completing an alternative proof of s - duality based on m - strings . while making an alternative derivation of the s - duality and its anomaly
, we did not assume the smallness of @xmath20 .
on the other hand , around ( [ pert - difference ] ) , we required @xmath543 for the positive roots @xmath47 , and that @xmath189 for positive roots be in the range @xmath190 $ ] , to justify the uses of @xmath193 and @xmath151 .
so at least in this range , @xmath544 is exact at finite @xmath20 .
one can scan the whole complex planes of @xmath194 and @xmath20 , considering the multiple values of @xmath545 function , to find the most general form of the s - duality anomaly when ( [ pert - difference ] ) is violated .
we shall not do this exercise here . before closing this subsection
, we comment on the nature of the s - duality transformation of @xmath512 or @xmath161 , and what it implies to the s - duality of @xmath31 .
this issue is also related to the s - duality of the 4d limit of @xmath31 , which was studied in the context of agt correspondence @xcite . taking the 4d limit @xmath219 with fixed @xmath546 , ( [ z - s - dual - gaussian ] )
asserts that @xmath547 is related to its s - dual by a gaussian s - duality kernel .
this is because the @xmath213 dependence @xmath548 is given by @xmath549\ , \ ] ] and the overall @xmath213 dependence is absorbed into the @xmath533 integration to be @xmath550 . expanding the exponent of the kernel , and inserting @xmath551 , one obtains @xmath552 \exp\left[\frac{\pi i a_da^\prime}{\varepsilon_1\varepsilon_2}\right]\ , \ ] ] where @xmath553 .
the two @xmath330 s can be absorbed into two @xmath554 s on the left and right hand sides of ( [ z - s - dual - gaussian ] ) .
then , the 4d limit of ( [ z - s - dual - gaussian ] ) states that s - dualization is fourier transformation .
our studies imply that the same result holds for @xmath512 in 6d .
in fact , the proper s - duality transformation of the 4d partition function is known _ not _ to be the fourier transformation . instead , based on the agt correspondence ,
the s - duality kernel is asserted to be a nontrivial function given by the partition function of the 3d @xmath555 $ ] theory on @xmath556 @xcite . on the other hand , it has been found that the s - duality of the omega deformed partition function is the fourier transformation at all perturbative orders in @xmath557 @xcite .
an explanation of this was given in @xcite , which finds that the choice of proper ` normalization factor ' independent of @xmath217 dressing the instanton partition funciton yields such a nonperturbative correction in @xmath216 .
this is the ` choice ' of @xmath558 , which was called @xmath559 and @xmath560 in @xcite .
incidently , what we find in 6d is analogous to the findings of @xcite .
namely , @xmath31 was defined in section 2 with @xmath107 which is manifestly invariant under weyl symmetry .
however , @xmath161 was defined with @xmath151 which is not invariant under weyl symmetry . in fact
, we checked that the ratio @xmath561 at small @xmath166 is nonperturbative in @xmath166 , which is qualitatively consistent with @xcite .
so along this line , it will be interesting to pursue the 6d extensions of @xcite .
we stress again that , all our findings in this subsection concerns the prepotential in the @xmath168 limit , for which the distinction of @xmath107 or @xmath151 is irrelevant .
we generalize some studies we made for @xmath14 type @xmath2 theories to @xmath15 and @xmath8 type theories . for @xmath15 type theories ,
@xmath341 can be compared with microscopic instanton calculus for the 5d @xmath7 @xmath21 theory , or the d0-d4-o4 matrix quantum mechanics @xcite . for @xmath8 types ,
we make a prediction of the s - duality and our knowledge of 5d perturbative prepotential .
all ade results will be partly tested in section 3.2 from 6d chiral anomalies . in the setting of section 2.1
, the leading coefficient @xmath386 of @xmath562 is obtained from @xmath563 , where @xmath138 is the rank of the gauge group .
the result is @xmath564 then using ( [ modular - anomaly - rec ] ) at @xmath395 , one finds @xmath397 to proceed , we classify the roots @xmath36 depending on their norm with @xmath47 .
the possibilities are @xmath565 it is again important to understand the set @xmath404 for ade , which we explain now . for @xmath566 ,
the @xmath567 roots in @xmath393 are given by @xmath568 , where @xmath400 and @xmath569 .
elements of @xmath404 are given for various @xmath47 by @xmath570 for @xmath347 , the number of roots is @xmath571 .
@xmath572 roots take the form of @xmath568 where @xmath399 and @xmath573 , from the @xmath574 subalgebra .
additional @xmath575 roots take the form of @xmath576 , where the total number of @xmath577 signs is even .
the structure of @xmath404 is given for various @xmath47 as follows .
firstly , when @xmath405 , then @xmath578 where @xmath579 means that all possible signs are allowed in the @xmath575 spinorial elements .
thus , one finds @xmath580 elements of @xmath404 in this case .
similarly , for @xmath581 , one finds @xmath582 where @xmath579 means the same .
so again , one finds @xmath583 . for @xmath584 , one can do a similar analysis .
finally , @xmath47 can be one of the @xmath575 spinorial elements , @xmath585 with @xmath586 and @xmath587 .
then , @xmath588 so @xmath589 . for @xmath590 , @xmath591 .
@xmath592 roots take the form of @xmath568 , @xmath593 , from @xmath594 subalgebra .
additional @xmath595 roots take the form of @xmath596 , with total number of @xmath577 signs being even .
finally , @xmath95 more roots are given by @xmath597 .
when @xmath598 , @xmath404 takes the same structure as that shown for @xmath347 .
for instance , for @xmath581 , one finds @xmath599 with @xmath600 . when @xmath601 , with @xmath602 , one finds @xmath603 with @xmath604 . finally , when @xmath605 , one finds @xmath606 with @xmath607 .
the case with @xmath608 is similar . for @xmath609 , @xmath610 .
@xmath611 roots take the form of @xmath568 , @xmath612 , from @xmath613 subalgebra .
additional @xmath614 roots take the form of @xmath615 with number of @xmath577 signs being even , forming the @xmath613 spinor representation . for @xmath581 ,
one finds @xmath616 , with @xmath617 .
other cases with roots of the form @xmath598 can be studied similarly . for @xmath618 with @xmath619 ,
one finds @xmath620 with @xmath621 . including the @xmath456 case studied in section 2.1 , one finds @xmath622 , where @xmath623 is the dual coxeter number .
see table [ ade ] .
another useful fact that can be checked with all @xmath404 we listed above is that , if @xmath624 , then @xmath625 is also a root .
one also finds @xmath626 , since @xmath627 .
so at given @xmath47 , one finds @xmath628 for any expression @xmath629 .
.data on simply laced lie algebras [ cols="^,^,^,^,^,^",options="header " , ] by following the analysis for the @xmath6 case , till ( [ root - sum ] ) , one finds @xmath630\ .\ ] ] now we use the identity ( [ psi - identity ] ) to rewrite the last term in the parenthesis as @xmath631\ .\ ] ] on the second term , we relabeled @xmath47 into @xmath632 in the first sum , and then took @xmath633 with @xmath624 as labeling the elements of @xmath634 .
the third term is simply the second term with renaming @xmath635 .
using the identity @xmath413 , ( [ perm - ade ] ) becomes @xmath636 .
thus , one obtains @xmath637=
-\frac{e_2(\tau)}{24}\left[\sum_{\alpha\in\delta } { \rm li}_{-1}(e^{\alpha(v)})+\frac{c_2|g|}{12}\right]\ , \ ] ] where at the last step we used the identity @xmath638 for simply - laced lie algebra .
@xmath394 contains @xmath639 , so makes a prediction on the instanton corrections . for @xmath640
, one can compare this against microscopic instanton calculus for the 5d @xmath21 theory @xcite .
we compared the two results at 1 instanton level for @xmath641 .
namely , ( [ ade - m4 ] ) implies @xmath642 where @xmath643 . on the other hand ,
the single instanton partition function @xmath644 for the @xmath7 theory can be obtained by starting from the witten index for the quantum mechanics describing an o4@xmath645 plane , @xmath95 d0-branes and @xmath646 d4-branes ( in the covering space ) .
the index is a complicated residue sum .
one should further subtract the contributions from d0-branes unbound to d4-o4 , which was explained in @xcite . following this procedure
, we checked that @xmath647 one can continue to generate higher order @xmath343 s , and also the microscopic instanton calculus for general @xmath15 at higher order in @xmath233 , and compare them . here
we simply conjecture @xmath648 for all @xmath649 , where @xmath138 is the rank of @xmath130 . for @xmath650 ,
we have tested it extensively in section 2.1 , after adding one free tensor multiplet to make it @xmath6 . for @xmath640 , we tested it till @xmath391 , @xmath651 order only at @xmath652 , but in principle one can do all the calculus of section 2.1 , following the methods of @xcite . for @xmath8 , this is just a prediction by assuming s - duality and 5d perturbative results . the last term proportional to @xmath653 will be further tested in section 3 , from the 6d chiral anomaly of @xmath654 r - symmetry .
in this section , we compute the asymptotic form of the prepotential at strong coupling , or high ` temperature ' @xmath655 .
this is the limit in which the compactification radius @xmath82 of the sixth circle becomes large , or equivalently in which d0-branes become light .
the key technique of computation will be the anomalous s - duality that we developed in section 2 .
our convention is that the strong coupling theory of our interest is the ` s - dualized ' theory .
so we take @xmath656 , and @xmath657 .
recall @xmath341 satisfies @xmath658 where @xmath256 , @xmath659 , @xmath279 .
we replaced @xmath341 by @xmath163 when it appears with @xmath194 derivatives , since @xmath660 is independent of @xmath194 . inserting @xmath661
, one finds that @xmath662 using ( [ abelian - s - dual ] ) and @xmath663 , one obtains @xmath664 inserting this in ( [ micro - dual ] ) , one obtains @xmath665 we shall study it in the limit @xmath235 with @xmath666 and @xmath332 fixed .
the limit @xmath235 on the right hand side has to be understood with care , since @xmath667 scales with @xmath217 .
also , we should study how @xmath194 scales with @xmath235 , at fixed @xmath332 .
had @xmath177 not scaled with @xmath217 , one would have naively expected that the instanton corrections in @xmath163 would have been suppressed at @xmath234 , so that we could replace @xmath163 on the right hand side by @xmath387 .
let us check when this is correct .
this expectation is correct if @xmath668 does not scale to be larger than @xmath669 . from ( [ zk - young ] ) , @xmath670 scales like @xmath671 at @xmath672 . for this factor to be smaller than @xmath669
, one should require @xmath673 .
let us take @xmath217 to be purely imaginary for convenience ( although most of our final results are valid for complex @xmath217 ) .
then , @xmath670 can be ignored if @xmath674 when @xmath675 reaches @xmath676 , we encounter a phase transition , beyond which one should make a new @xmath233 expansion on the right hand side .
the correct nature of this phase transition will be commented on later . to make the simplest calculus at @xmath656 , we take @xmath666 to satisfy ( [ mass - requirement ] ) .
let us also discuss how @xmath194 should scale at fixed @xmath332 .
we shall first assume that @xmath194 is finite at finite @xmath332 , and then show that it is consistent with ignoring @xmath677 .
if @xmath677 can be ignored , then the relation between @xmath194 and @xmath332 can be simplified as @xmath678 @xmath679 is given by @xmath680\ .\ ] ] since we assume that @xmath194 is finite , the first term not containing @xmath20 yields a subleading contribution , from the @xmath681 factor in ( [ v - relation ] ) . to be definite ,
we take @xmath682 so that @xmath683 . then , @xmath684 where we used @xmath685 with the branch cut at @xmath686 .
so one can approximate @xmath687\ , \ ] ] where @xmath688 can be ignored at @xmath683 .
we ignored all the terms that vanish after summing over @xmath47 , or are subleading in the @xmath681 limit . expanding the square on the right hand side ,
the term proportional to @xmath689 will vanish upon summing over @xmath47 .
the next term proportional to @xmath690 will be the nonzero leading term .
one obtains @xmath691 where we used @xmath692 here , @xmath85 is the @xmath76 projection to @xmath456 .
decomposing @xmath693 , one finds that @xmath694 at @xmath235 .
, one finds that @xmath695 diverges . in this case
, one has to approximate ( [ v - relation ] ) by assuming that @xmath194 can scale with @xmath217 . at @xmath696
, we checked for @xmath424 that @xmath695 scales like @xmath697 , which grows large but is much smaller than @xmath667 .
due to this fact , @xmath194 does not affect the asymptotic free energy , and our final result for @xmath698 below will be reliable even at @xmath696 . ] inserting this back to @xmath679 , one obtains @xmath699 also @xmath387 itself is given by @xmath700 where we used @xmath701 if the real part of @xmath176 is positive and large .
therefore , the asymptotic prepotential is given by @xmath702 in particular , one finds that the coulomb vev @xmath332 does not appear in the asymptotic limit .
this is natural since the coulomb vev is a dimensionful parameter , which should not be visible in the large momentum limit .
this is a result for @xmath703 . when @xmath704 , all the analysis above is same except the step of approximating @xmath705 . in this case , @xmath706 is replaced by @xmath707 . combining the two cases , one obtains @xmath708\ , \ ] ] where @xmath709 signs are for @xmath710 , respectively .
finally , when @xmath711 , @xmath705 provides subleading contribution in @xmath217 so that one finds @xmath712\ , \ ] ] where the superscript @xmath713 means vanishing imaginary part of @xmath666 . at this stage
, we note that @xmath698 at @xmath714 can be written as the following holomorphic function with a branch cut , @xmath715 this expression will be helpful later .
we first investigate @xmath698 for purely imaginary @xmath716 , at @xmath717 .
one finds @xmath718\ .\ ] ] the partition function undergoes a phase transition at @xmath719 , from certain perturbative particles being massless at @xmath720 .
one may wonder how @xmath698 behaves beyond @xmath721 . at @xmath721 ,
one finds from the s - dual picture that @xmath677 can not be ignored , since @xmath722 at @xmath723 .
this means that particles with nonzero instanton number become light at these points .
one can get some insights on these nonperturbative massless particles .
system , width=491 ] to see this , it is helpful to recall the type iib 5-brane web realization of the 5d @xmath21 system . more precisely , we realize the ` s - dualized ' setting at @xmath235 , using weakly coupled type iib string theory .
the brane web first consists of @xmath11 d5-branes and @xmath301 ns5-brane , all extended on @xmath724 directions , transverse to @xmath725 , and forming a web on the @xmath726-@xmath727 plane .
one makes a twisted compactification @xmath728 .
the d5-branes wrap @xmath726 direction , and form a web with the ns5-brane extended along @xmath727 , as shown in fig.[web ] .
the twisted compactification guarantees that the open strings with tension @xmath729 ( where @xmath730 ) suspended between d5-branes across the web have mass @xmath731 .
d1-branes wrapping @xmath726 ending on ns5-brane are identified as yang - mills instantons .
unit instanton s mass is given by @xmath732 , which should be identified with @xmath236 in our m5-brane setting .
so one obtains @xmath733 . on the other hand , @xmath217 is given at zero axion by @xmath734 , which should be identified in our m5-brane setting as @xmath735 .
so one finds @xmath736 .
these provide the relations between the parameters @xmath737 , @xmath738 and the m5-brane parameters @xmath739 .
the slope of the @xmath740 5-brane is @xmath741 , where @xmath742 and @xmath743 are the distances between the two ends of the segment on fig.[web ] .
we stated above that @xmath744 , so one finds @xmath745 . in this setting ,
the segment of @xmath740 5-brane shrinks at @xmath746 . here
, the perturbative hypermultiplet particle becomes massless , corresponding to the fundamental strings connecting d5-branes across the ns5-brane .
this causes the so - called flop phase transition .
the singular term proportional to @xmath747 in ( [ high - t - imaginary ] ) is caused by @xmath387 in the s - dual setting , from the particles becoming massless at @xmath746 .
so we conclude that the cusp @xmath748 is due to the flop transition . as one increases positive @xmath214
, the next transition happens when the @xmath740 brane segment goes around the circle in @xmath726 direction , as shown on the bottom - right side of fig.[web ] .
this happens at @xmath749 , @xmath750 so one finds that the transition happens at @xmath751 precisely when @xmath171 can not be ignored . across @xmath721 , i.e. @xmath752 , the @xmath11 d5-brane segment shrinks .
so across this value , another transition happens , with the d1-brane segment extended along the shrinking segment being massless . as one continues to change @xmath214 , transitions due to
non - perturbative massless particles will happen at @xmath753 with @xmath754 being integers . at @xmath755 , the nature of this transition
is hard to study .
this is because the massless particles are nontrivial bound states of d1-branes . also , studying the @xmath235 approximations around @xmath719 , not all massless particles were responsible for the cusp at @xmath719 .
so it will be important to know which types of massless particles contribute to the cusp of @xmath698 at @xmath753 .
however , if @xmath754 is a multiple of @xmath11 , one finds from the 5-brane web diagram that the transition is an @xmath5 transformation of the transition at @xmath719 , so that the same type of cusp will happen .
indeed this has to be the case , since @xmath756 ( or @xmath757 ) is the periodicity of the instanton partition function .
interestingly , if one takes the holomorphic extension ( [ high - t - li4 ] ) within @xmath758 to the whole region of @xmath666 , one gets a definite prediction on @xmath698 as a function of real @xmath176 , and also on the nature of phase transitions at all @xmath754 . plotting ( [ high - t - li4 ] ) for the entire real @xmath176
, one finds fig.[continuation ] .
@xmath759 for @xmath760 is given by simply translating the function in the range @xmath761 by @xmath762 .
this means that all the cusp structures are completely the same at all @xmath754 , at least in @xmath759
. it will be interesting to understand how the non - perturbative massless particles cause the same cusp in ( [ high - t - li4 ] ) .
also , in ( [ high - t - li4 ] ) or in fig.[continuation ] , @xmath698 has a shorter period @xmath763 .
it will be interesting to see if the reduced period has to do with multiple - wrapping of m5-branes on @xmath0 , analogous to the mutiple - winding fundamental strings @xcite .
now we study @xmath698 for purely real @xmath666 .
the asymptotic free energy is given by @xmath764\ .\ ] ] holding real @xmath666 fixed , and further taking the large @xmath11 limit , one finds that the free energy is proportional to @xmath4 .
namely , one finds that the single particle index @xmath765 in the limit @xmath168 , @xmath238 is given by @xmath766\ , \ ] ] where we dropped the @xmath767 subscripts .
this shows that the microscopic entropy ( with minus sign for fermions ) of light d0-branes bound to @xmath11 d4-branes exhibit large number of bound states proportional to @xmath4 .
the second term proportional to @xmath11 clearly comes from @xmath11 free tensor multiplets , as this comes from the s - dualization of @xmath211 .
one can understand that the first term proportional to @xmath768 is a remnant of the cancelation between bosonic and fermionic states in the index , since this term vanishes at @xmath339
. it will be interesting to guess what kind of index @xmath769 would exhibit the above behavior in the high temperature limit .
in particular , having the analytic expression ( [ high - t - li4 ] ) given in terms of @xmath770 functions , with chemical potentials multiplied by @xmath11 , it will be interesting to seek for an interpretation using multiple - wrapping of m5-branes , or instanton partons @xcite .
finally , we comment that one can obtain the asymptotic free energy at @xmath656 for all ade theories , starting from ( [ s - dual - ade ] ) and following the analysis of this section . to make a similar calculation
, one also needs to know the perturbative partition function , and the range of @xmath675 in which the instanton correction @xmath677 can be ignored on the right hand side .
the perturbative prepotential is straightforward for all ade . as for the instanton part
, we should know when @xmath771 is much smaller than @xmath301 at @xmath772 for large real part of @xmath667 .
the leading behavior of @xmath670 for large real @xmath20 can be easily inferred , by knowing the correct parameter scalings between the 5d @xmath21 theory and the pure @xmath773 theory .
namely , one finds @xmath774 where @xmath623 is the dual coxeter number of the gauge group @xmath130 .
this is because the pure 5d @xmath773 theory is obtained by taking the limit @xmath775 , @xmath772 , with @xmath776 held fixed .
this means that one can ignore the instanton part in the region @xmath777 . by following the analysis for the @xmath6 case , the asymptotic free energies of ade theories are given by @xmath778\ , \ ] ] where @xmath709 signs are for @xmath779 , and @xmath780\ ] ] for @xmath711 .
we provide a small consistency check of @xmath698 for the @xmath37 case . by this exercise
, one can also get better intuitions on the true nature of the approximations and phase transitions , which perhaps may be a bit obscure in our s - duality based approach . in the previous s - duality
based approach , we first took @xmath18 limit of the partition function , to focus on the prepotential only .
then using the s - duality , we extracted out the @xmath781 asymptotics of the prepotential , where @xmath782 , at finite @xmath20 and @xmath11 . we reconsider the same limits directly with the @xmath37 instanton partition function .
the instanton partition function is given by @xmath783\sim\exp\left[\frac{4}{\epsilon_1\epsilon_2 } \sum_{n=1}^\infty\frac{\sinh^2\frac{nm}{2}}{n^3}\frac{e^{-n\beta}}{1-e^{-n\beta}}\right]\ ] ] in the @xmath18 limit .
now we take the @xmath781 limit at fixed @xmath20 .
this is somewhat tricky at real @xmath20 , which we also take to be positive .
this is because the above formula is valid for @xmath784 when @xmath20 is real .
physically , this is because the partition function @xmath31 has poles at @xmath785 for all positive integers @xmath754 .
so with fixed real @xmath20 , one would hit many poles as one takes the @xmath781 limit . to deal with this situation more easily , we first continue @xmath20 to be purely imaginary , @xmath786 , and continue back later to complex @xmath20 . inserting @xmath786 and taking @xmath781 limit , one obtains @xmath787= \exp\left[\frac{1}{\epsilon_1\epsilon_2\beta } \sum_{n=1}^\infty\frac{1}{n^4}\left(e^{inx}+e^{-inx}-2\right)\right]\nonumber\\ & = & \exp\left[\frac{1}{\epsilon_1\epsilon_2\beta } \left({\rm li}_4(e^{ix})+{\rm li}_4(e^{-ix})-2{\rm li}_4(1)\right)\right]\ .\end{aligned}\ ] ] the final expression can be continued to complex @xmath176 .
here , we use the property @xmath788 where @xmath789 for @xmath790 .
@xmath183 are the bernoulli polynomials , given by @xmath791 in particular , one finds @xmath792 , so that @xmath793\ ] ] for @xmath794 .
this leads to the asymptotic formula @xmath795\ ] ] for @xmath796 , which is in complete agreement with the s - duality - based result , ( [ high - t - imaginary ] ) , upon inserting @xmath797 and @xmath156 .
when @xmath798 , we use a different identity of @xmath770 function to find a similar expression , with the sign of the @xmath799 term flipped .
this also shows that the continuation ( [ high - t - li4 ] ) beyond @xmath800 by the @xmath770 functions to complex @xmath176 is indeed correct . in this subsection
, we shall discuss the connection between the s - duality anomaly and the 6d chiral anomalies of global symmetries .
in particular , we shall independently compute some part of our asymptotic free energy @xmath698 based on chiral anomaly only .
however , let us start by giving a general comment , on why one should naturally expect s - duality anomaly of the partition function if the system has chiral anomaly . consider a partition function of even dimensional chiral theories on @xmath3 , like 2d theories on @xmath3 or our system on @xmath13
. for a global symmetry , one turns on a background gauge field @xmath801 .
in particular , let us turn on the flat connection of @xmath801 on @xmath3 .
we shall only be interested in abelian flat connections , characterized by the commuting holonomies along the two circles of @xmath3 .
large gauge transformations would have made both holonomies to be periodic , had there been no chiral anomalies . for simplicity ,
let us take a rectangular torus with two radii @xmath802 , @xmath803 , respectively .
then the large gauge transformations would have given the periods @xmath804 and @xmath805 .
with matter fields having integral charge @xmath233 of this global symmetry , the modes of these fields would have frequencies @xmath806 on @xmath3 , with integral @xmath807 , which is invariant under the periodic shifts of @xmath808 , @xmath809 .
this is a consequence of these gauge symmetries .
however , in quantum observables like the partition function on @xmath3 , these large gauge transformations may fail to be symmetries for theories with chiral anomalies .
this is because one has to regularize the path integral over these modes , by regarding one of the two directions as temporal circle @xcite . by this procedure ,
one of the two holonomies @xmath810 fail to be periodic in the partition functions .
this is precisely what happen for the 2d elliptic genera @xcite .
we expect that similar things will happen to 6d chiral theories on @xmath13 , but we can not make this expectation more precise here
. we shall simply assume the failure of double periodicity of background holonomies due to chiral anomalies , and then explain that it forces the partition function to have s - duality anomaly , as we found in section 2 by nonzero @xmath24 .
let us write the background holonomies as a complex number @xmath20 .
had a free energy @xmath811 on @xmath3 been exactly s - dual , then its exact s - duality @xmath812 means that @xmath20 has double period .
this is because if the right hand side has period in one direction , say @xmath813 , the left hand side forces @xmath814 , and thus @xmath815 , contradicting the obstruction of double periodicity from chiral anomaly .
this comment applies to our 6d partition functions .
so we naturally expect s - duality anomaly .
with these motivations in mind , rather than trying to elaborate on it , we shall make a concretely calculation which shows that a particular term in our asymptotic high temperature free energy dictated by @xmath24 can be computed using 6d chiral anomaly only .
let us first explain the anomalies of the 6d @xmath2 theory of @xmath14 type .
more precisely , we shall consider the anomaly of the interacting @xmath14 type theory times a decoupled free self - dual tensor multiplet theory .
this corresponds to the system of @xmath11 m5-branes including the decoupled center - of - mass multiplet .
the anomaly polynomial @xmath38-form is given by @xmath816 where @xmath817 is the anomaly of the single m5-brane theory , or one free @xmath2 tensor multiplet , @xmath818\ .\ ] ] the pontryagin classes are defined by @xmath819\ .\end{aligned}\ ] ] here , traces are acting on either @xmath820 matrices for @xmath821 tangent bundle @xmath822 , or @xmath823 matrices for @xmath654 normal bundle @xmath11 . taking their curvatures to be @xmath213 and @xmath16 , respectively , one finds @xmath824\nonumber\\ & & + \frac{n^3-n}{24}\left(-\frac{1}{4}{\rm tr}f^4+\frac{1}{8}({\rm tr}f^2)^2\right)\ .\end{aligned}\ ] ] we shall restrict @xmath16 to a cartan part . in particular , since we shall be taking the omega backgrounds to be small , the cartan for @xmath825 will have much smaller background field than @xmath278 , from @xmath826 .
so we shall only turn on the background field for the cartan in @xmath827 , corresponding to our @xmath21 mass @xmath20 .
@xmath16 is a @xmath823 matrix - valued 2-form , whose components are @xmath828 with @xmath829 .
the component corresponding to the cartan of @xmath278 is obtained by keeping @xmath830 only . with this restriction ,
one finds @xmath831 , @xmath832 . inserting these ,
the @xmath821 and @xmath833 anomalies are given by @xmath834\ .\ ] ] only the first term @xmath835 will be relevant for the computations below .
our goal is to compute some part of the asymptotic free energy at high temperature @xmath836 , using 6d chiral anomalies .
recall that we found @xmath837\ ] ] where @xmath579 stands for the @xmath838 term which exists when @xmath20 has imaginary component .
the @xmath838 term will not be of our interest in this subsection .
we obtained this expression at @xmath839 and @xmath836 , where @xmath840 is the same @xmath841 used before .
often , we used purely imaginary @xmath841 with @xmath842 , but we keep real @xmath843 in this subsection to see a clear relation to chiral anomalies .
for a reason to be explained below , we would like to study the asymptotic free energy when all the chemical potentials @xmath844 are purely imaginary . so inserting @xmath845 , @xmath846 ( with real @xmath844 ) in the places of @xmath844 in ( [ seff ] ) , one obtains @xmath847 $ ] . in this setting
, we focus on the imaginary part of the effective action , @xmath848\ , \ ] ] and compute it from 6d chiral anomalies .
especially , we shall compute part of @xmath849 from the 5d effective action approach for the 6d theory on small temporal circle .
6d chiral anomaly determines a special class of terms in the 5d effective action .
it turns out that , knowing the terms determined by anomaly , one can only compute the term proportional to @xmath391 .
so we shall pay attention to the first term @xmath850 we shall argue below that this term is completely dictated by 6d chiral anomaly , and then we re - compute this term using chiral anomaly only .
this will provide another strong test of our findings from the d0-d4 calculus . then , since one naturally expects that supersymmetrization of ( [ imaginary - normal ] ) is holomorphic in @xmath841 , one can reconstruct the term @xmath851 in ( [ seff ] ) .
we shall consider the 6d anomaly from the viewpoint of 5d effective action , obtained by compactification on a small circle of circumference @xmath852 , and discuss our asymptotic free energy @xmath853 on @xmath854 in this setting .
on @xmath3 , regarding one circle as the temporal circle , the partition function is an index of the form @xmath855\ .\ ] ] real @xmath844 is consistent with the conventions for the partition function presented at the beginning of section 2 . in this setting , the chemical potentials @xmath844 will twist the translation on the temporal circle in a way that the twisted time evolution is not unitary ( simply because the factors in the trace are not unitary transformations ) .
this would cause a complex deformation of the euclidean action by twisting with chemical potentials . for a technical reason
, it will be convenient to keep these twistings to preserve the reality of the action .
so we replace @xmath856 which will make real twists of the euclidean action .
this is equivalent to the insertions of @xmath857 around ( [ imaginary - normal ] ) .
the factor @xmath858 demands us to consider a 6d euclidean theory whose temporal coordinate @xmath859 satisfies periodicity @xmath860 .
this forms a circle of the @xmath3 .
another circle factor is labeled by @xmath176 , which we take to have periodicity @xmath756 .
defining @xmath861 , one obtains @xmath862 so @xmath841 is the complex structure of @xmath3 .
this torus is endowed with the metric @xmath863 and periods @xmath864 . including the chemical potential @xmath865 , the metric of @xmath13 is given by @xmath866 where @xmath867 are complex coordinates of @xmath868 with charges @xmath869=\delta_{ab}$ ] .
finally , the chemical potential @xmath20 is realized as the background gauge field @xmath870 for @xmath833 .
also , @xmath871 , where @xmath96 is a supercharge preserved by the index .
so @xmath31 is independent of @xmath100 .
following @xcite ( see also @xcite ) , we shall make a kk reduction on the small circle along @xmath859 , for small inverse - temperature @xmath852 . to this end , one rewrites the background in the form of @xmath872 where @xmath873 with @xmath874 is the 5d metric , @xmath875 is the dilaton , and @xmath876 is the gravi - photon field , where @xmath877 . the 6d background gauge field @xmath801 for @xmath833 is also rewritten in the form @xmath878 , where @xmath879 is the 5d background gauge field and @xmath880 is the 5d scalar .
so one finds @xmath881 and @xmath882 .
if the 6d theory compactified on a small circle has no 5d massless modes , one can express the thermal partition function in terms of a 5d local effective field theory of background fields , where the 5d derivative expansion corresponds to a @xmath36 series expansion .
as noted in @xcite , with massless modes in 5d , there could be nonlocal part of the effective action which is smooth in the @xmath781 limit . in our case ,
the non - local part comes from the 5d perturbative maximal sym .
there is additional difficulty in using the derivative expansion in our setting , since some of our background fields are proportional to @xmath883 , which may spoil the orderings provided by the derivative expansion .
so it appears tricky to directly employ the formalism of @xcite .
however , one can study the imaginary part ( [ imaginary ] ) of our asymptotic free energy using the 5d approach .
the imaginary part can be computed completely by knowing the 5d chern - simons like terms . to explain this , note first that we have been careful to set all our background fields to be real , e.g. by setting our chemical potentials to be imaginary . with real background fields turned on , suppose that we first reduce the 6d theory on a small circle to a general 5d _ lorentzian _ spacetime .
then the 5d effective action is real , since hermiticity is not broken in the lorentzian theory .
now we wick - rotate the ` time ' direction in this 5d setting .
since all background fields are real , the only possible step which may cause complex effective action is the wick rotation to euclidean 5d space . here , note that we are seeking for an effective action of the vectors @xmath884 ( spin connection ) , tensor @xmath873 , and scalars @xmath880 , @xmath298 . to compute the imaginary part , one can focus on the local terms .
this is because the nonlocal terms come from the determinant of 5d maximal sym whose fields are covariantized by real background fields , which is real . among the local terms obtained from scalar lagrangian density
, we should seek for terms containing the tensor @xmath885 to obtain imaginary contribution after wick rotation .
it should be contracted with antisymmetric tensors formed by the background fields .
there are many possibilities , arranged in derivative expansion .
for instance , there could complicated terms like @xmath886 , and so on .
although there are many terms , let us comment that there can be gauge invariant terms and gauge non - invariant terms in the imaginary action .
the latter class should exist because the 5d effective action should realize 6d chiral anomalies .
the coefficients of the terms in the latter class are thus completely determined by known 6d anomalies @xcite . among the gauge invariant terms
, there can be action coming from gauge invariant lagrangian density , like the term that we illustrated in the last paragraph .
finally , there may be chern - simons terms in which lagrangian densities are not gauge invariant but their integrals are .
so the imaginary action takes the following structure , @xmath887 where @xmath888 is the radius of the small sixth circle with circumference @xmath889 .
@xmath890 consists of the gauge invariant chern - simons terms .
@xmath891 is part of the gauge non - invariant chern - simons terms that comes from @xmath892 normal bundle anomaly in 6d , namely the first term @xmath893 of ( [ anomaly - pol - reduced ] ) .
anomaly matching fixes @xmath894 , as well as the relative coefficients as shown on the second line . , taking into account the covariant anomaly rather than the consistent anomaly .
this is sufficient for our calculus of the free energy . ]
the omitted terms @xmath579 in @xmath895 are other chern - simons terms containing @xmath896 , which we do not need here . the omitted terms in @xmath891 can all be computed from mixed anomalies and gravitational anomalies of ( [ anomaly - pol - reduced ] ) , which we do not work out here as we shall not need them . finally , @xmath897 is the action containing @xmath885 associated with gauge invariant lagrangian density , e.g. @xmath898 , @xmath899g(\phi , a_6)$ ] , and so on .
one point we emphasize is that @xmath897 can come in infinite series of derivative expansion , while @xmath895 and @xmath900 consist of finite number of terms and can be completely classified . the imaginary terms have rich possibilities . here
we consider the terms which are nonzero with our background , and also the leading terms in small @xmath166 , proportional to @xmath313 .
@xmath881 is constant in our background . also , @xmath901 is constant times the graviphoton .
plugging in these values , one obtains @xmath902 the parenthesis consists of the fields reduced from 6d metric ( [ 6d - metric ] ) . note that , after plugging in constant @xmath880 and @xmath901 , all terms should be formally gauge invariant in the remaining fields .
this is because the only possible gauge non - invariant terms @xmath900 , completely dictated by anomaly , also become gauge invariant like @xmath903 with constant @xmath880 .
now we note the fact that , in the 6d metric , all @xmath867 coordinates of @xmath46 are multiplied by @xmath865 .
so in the parenthesis of ( [ imaginary - reduced ] ) , the only @xmath904 s not associated with @xmath865 are derivatives .
so one makes a formal derivative expansion of this term , assigning the ` mass dimensions ' @xmath905=0 $ ] , @xmath906=0 $ ] , @xmath907=0 $ ] , @xmath908=1 $ ] .
the lowest order term comes in two derivatives , and is proportional to @xmath909 .
there can be no other gauge - invariant terms at this order .
this term indeed yields the desired @xmath313 scaling .
firstly , the integral @xmath910 can be scaled into @xmath911 times a measure depending on @xmath912 .
also , two derivatives in @xmath909 can also be scaled with @xmath913 , yielding another overall factor @xmath914 .
@xmath867 in the remaining integral appear in the combination @xmath912 , including the integration variable , so is independent of @xmath865 .
so this term yields the right scaling @xmath915 .
therefore , to compute ( [ imaginary - normal ] ) , we only need to consider those terms that reduce to @xmath916 upon plugging in our background .
this implies that one does not have to consider @xmath897 of ( [ 5d - derivative ] ) , since they are associated with local lagrangian density and can not provide terms like ( [ cs - leading ] ) .
so we only consider @xmath895 and @xmath900 of ( [ 5d - derivative ] ) . unlike the coefficients of @xmath900 , coefficients of @xmath895
can not be determined with our limited knowledge of the 6d theory .
so even after restricting our interest to the imaginary part ( [ imaginary ] ) of the effective action , we can not compute them all due to our ignorance on these coefficients . since the second term of @xmath890 is quadratic in @xmath879
, we can not compute the @xmath917 term of ( [ imaginary ] ) .
this is why we shall not need the mixed anomaly contributions in @xmath891 coming from the term @xmath918 in ( [ anomaly - pol - reduced ] ) , which will also yield a contribution at @xmath917 , since knowing them is incomplete to compute the whole @xmath917 contributions .
also , the @xmath919 term can not be computed since we do not know @xmath920 .
however , the chern - simons terms that are quartic in @xmath879 and @xmath880 are completely dictated by 6d anomalies , as shown on the second line of ( [ 5d - derivative ] ) .
note that quartic chern - simons term is allowed precisely because we allow gauge non - invariant chern - simons term , to match 6d anomalies which are fourth order in the fields .
thus , we can compute ( [ imaginary - normal ] ) from @xmath891 of ( [ 5d - derivative ] ) .
note also that , for imaginary chemical potentials , we have found earlier in this section that @xmath698 undergoes phase transitions due to massless particles .
this only changes @xmath919 or lower order terms , so that the @xmath391 order that we are going to compute is unaffected .
we also note in passing that , we can turn the logic around and use our d0-d4 results to constrain the 5d effective action .
namely , we know from our d0-d4 calculus the @xmath917 and @xmath919 coefficients of @xmath921 , and also the vanishing of the @xmath922 coefficient
. this knowledge can be used to constrain @xmath923 of ( [ 5d - derivative ] ) .
this information may be useful for studying other high temperature partition functions of the 6d @xmath2 theories .
coming back to the computation of ( [ imaginary - normal ] ) , we plug @xmath924 and @xmath881 into @xmath891 of ( [ 5d - derivative ] ) to obtain @xmath925 to compute this , one should evaluate the gravi - photon chern - simons term , @xmath926 where @xmath927 , with @xmath928 being the cartesian coordinates of @xmath46 . since @xmath929 , @xmath930 , ( [ adada ] ) becomes @xmath931 where @xmath932 , @xmath933 .
so one obtains @xmath934 where we plugged in @xmath888 .
this precisely agrees with ( [ imaginary - normal ] ) , based on d0-d4 calculus .
finally , let us comment that the same calculation can be done to test some part of ( [ ade - asymp ] ) for all ade theories
. for ade , ( [ ade - asymp ] ) yields the imaginary part @xmath935 simply by changing the coefficient @xmath936 from ( [ ade - asymp ] ) . on the other hand , the anomaly polynomial ( [ anomaly - u(n ) ] )
is replaced by the following polynomial @xmath937 for ade . again after restricting @xmath938 to @xmath833 , the term @xmath835 of ( [ anomaly - pol - reduced ] ) is replaced by @xmath939 .
so the calculations of this subsection can be done by replacing all @xmath4 by @xmath940 , completely reproducing ( [ seff - ade ] ) .
in this paper , we explored s - duality of the prepotential of the 6d @xmath2 theories compactified on @xmath3 , on the coulomb branch .
we found evidences of s - duality and its anomaly . using this result , we computed the asymptotic free energy of this system compactified on @xmath0 ( in the index version ) , when the omega background parameters @xmath166 and the chemical potential @xmath36 conjugate to the kk momentum are small .
the asymptotic free energy is proportional to @xmath4 in a suitable large @xmath11 limit , showing that the light kk fields exhibit the @xmath4 degrees of freedom .
after suitably complexifying the chemical potentials , we showed that the imaginary part of the free energy proportional to @xmath4 is completely reproduced from the 6d chiral anomaly of the @xmath654 r - symmetry .
most results are generalized to the ade class of @xmath2 theories . in the literatures ,
the @xmath4 scalings of various observables of 6d @xmath2 theory have been found , using various approaches .
thermal entropy of black m5-branes @xcite or various other quantities are computed from the gravity dual .
chiral anomalies are computed from the anomaly inflow mechanism @xcite .
the supersymmetric casimir energy on @xmath941 was computed from the superconformal index @xcite .
perhaps among these , the mysteries of 6d cft may be most directly addressed from the thermal partition function calculus of @xcite .
so it would be desirable to have a microscopic view of this phenomenon by directly counting states of the 6d cfts .
as far as we are aware of , such a direct account for @xmath4 scaling of states has not been available from a mircroscopic quantum calculus yet .
our studies show the @xmath4 scalings of the microscopically counted degrees of freedom .
more precisely , we compactified the 6d scft on @xmath0 , so @xmath4 degrees of freedom are absent at low energy .
however , at high temperature compared to the inverse - radius of the circle , we expect the 6d cft physics to be visible , hopefully in our @xmath16 .
one subtlety is that fermionic states are counted with minus sign in the index , so there may be cancelation between bosons and fermions . even after this possible cancelation , we find that the uncanceled free energy still exhibits @xmath4 scaling , which proves that the 6d cft has @xmath4 degrees of freedom .
we have provided an alternative study of the asymptotic free energy based on 6d chiral anomalies , which completely agrees with our d0-d4 calculus .
our studies based on d0-d4 system also shows that the light d0-brane particles are responsible for the uv enhancement of degrees of freedom . since d0-branes are the key objects which construct m - theory at strong coupling limit of the type iia strings , it is natural to see that they are also responsible for the @xmath4 degrees of freedom of the 6d @xmath2 theory .
it will be interesting to better understand the the single particle index @xmath942 which yields this behavior .
in particular , conjectures on instanton partons @xcite may be addressed in more detail .
the coulomb branch partition function on @xmath13 was used as building blocks of interesting cft indices in the symmetric phase .
we comment that our asymptotic free energy proportional to @xmath4 does not appear in these symmetric phase indices .
let us explain this with the 6d superconformal index , and the dlcq index .
firstly , it has been proposed that the d0-d4 partition function , or more precisely this partition function multiplied by the 5d perturbative part , is a building block for the 6d superconformal indices @xcite on @xmath943 .
so one might wonder whether our finding @xmath944 ( with @xmath945 ) at high temperature has implications to the supercofonrmal index .
one can immedidately see that the answer is negative . for this discussion ,
the relevant formula is presented in @xcite , which uses the product of @xmath946 copies of coulomb branch partition functions on @xmath13 as the integrand .
the angular momentum chemical potentials of @xmath947 on @xmath941 are labeled by three numbers @xmath948 satisfying @xmath949 . in this
setting , the @xmath946 sets of omega deformation parameters are given by @xmath950 , @xmath951 , @xmath952 respectively . since the asymptotic formula for @xmath31 is obtained in the limit of small @xmath35 , one can study the superconformal index in the limit of small @xmath948 . in this limit ,
the most divergent part in @xmath166 is given by @xmath953\ .\ ] ] it is an identity that the sum in the square bracket vanishes , so that the leading asymptotic part proportional to @xmath4 vanishes on @xmath943 .
so our @xmath698 has no implication to the superconformal index .
however , study of the the subleading part @xmath954 will be interesting , along the lines of our section 2.2 .
we hope to come back to this problem in the near future .
secondly , the m5-brane theory compactified on a lightlike circle can be studied using the d0-d4 quantum mechanics @xcite .
its index at dlcq momentum @xmath10 can be computed by integrating the d0-d4 index in the coulomb branch suitably with the coulomb vev @xmath194 , as explained in @xcite .
so one finds ( again with @xmath955 ) @xmath956\ .\ ] ] here , unlike the partition function on @xmath13 , where we have notion of multi - particles so that @xmath957 itself is meaningful as the singe particle index , the dlcq index is defined with a confining harmonic potential on @xmath46 @xcite .
thus , the exponent can not be physically meaningful separately . also , the definition of @xmath958 is such that @xmath959 has to be real and bigger than other fugacities , as @xmath960 plays the role of main convergence parameter .
so one has to set @xmath530 .
this implies that @xmath958 does not exhibit exponential growth , but is rather highly suppressed at small @xmath36 , presumably due to boson / fermion cancelation .
omega deformed partition functions can also be used to study 6d @xmath962 superconformal field theories .
in fact , for many 6d @xmath962 systems , the index on @xmath13 is known in the ` self - dual string expansion , ' similar to the m - string expansion explained in our section 2.2 .
the coefficients like @xmath463 of section 2.2 are elliptic genera of 2d cfts for the 6d self - dual strings in the tensor branch .
those elliptic genera have been studied for various 6d @xmath962 theories @xcite .
the s - duality anomaly and the high temperature asymptotic free energies could be studied using the approaches explored in this paper .
this may be an interesting approach to explore the rich physics of 6d cfts and their compactifications to 5d/4d .
it would also be interesting to further study the s - duality of the full index of the @xmath2 theory , based on some ideas sketched in our section 2.2 .
following @xcite , we find it interesting to study the wilson/t hooft line defects uplifed to 6d surface operators .
s - dualities of other defect operators should also be interesting .
finally , one may ask if a suitable m2-brane partition function on @xmath963 can exhibit @xmath964 scaling , where @xmath965 is the omega deformation parameter .
although this scaling has been microscopically computed from the @xmath556 partition function , or the entanglement entropy , perhaps better physical intuitions can be obtained by directly accounting for where such degrees of freedom come from , like we did for 6d scfts on @xmath0 from d0-branes ( instanton solitons ) .
we thank prarit agarwal , joonho kim , kimyeong lee , jaemo park , jaewon song , shuichi yokoyama for helpful discussions , and especially hee - cheol kim for many inspiring discussions and comments .
we also thank joonho kim for helping us with the @xmath641 instanton calculus .
this work is supported in part by nrf grant 2015r1a2a2a01003124 ( sk , jn ) , and hyundai motor chung mong - koo foundation ( jn ) . c. montonen and d. i. olive , phys .
* 72b * , 117 ( 1977 ) . h. osborn , phys . lett . * 83b * , 321 ( 1979 ) . a. sen , phys .
b * 329 * , 217 ( 1994 ) [ hep - th/9402032 ] .
e. witten , nucl .
b * 443 * , 85 ( 1995 ) [ hep - th/9503124 ] .
e. witten , in * los angeles 1995 , future perspectives in string theory * 501 - 523 [ hep - th/9507121 ] .
h. c. kim , s. kim , e. koh , k. lee and s. lee , jhep * 1112 * , 031 ( 2011 ) [ arxiv:1110.2175 [ hep - th ] ] .
b. haghighat , a. iqbal , c. kozcaz , g. lockhart and c. vafa , commun .
phys . * 334 * , no . 2 , 779 ( 2015 ) [ arxiv:1305.6322 [ hep - th ] ] . y. hwang , j. kim and s. kim , jhep * 1612 * , 148 ( 2016 ) [ arxiv:1607.08557 [ hep - th ] ] .
l. di pietro and z. komargodski , jhep * 1412 * , 031 ( 2014 ) [ arxiv:1407.6061 [ hep - th ] ] .
r. flume and r. poghossian , int .
j. mod .
a * 18 * , 2541 ( 2003 ) [ hep - th/0208176 ] .
u. bruzzo , f. fucito , j. f. morales and a. tanzini , jhep * 0305 * , 054 ( 2003 ) [ hep - th/0211108 ] . c. hwang , j. kim , s. kim and j. park , jhep * 1507 * , 063 ( 2015 ) addendum : [ jhep * 1604 * , 094 ( 2016 ) ] [ arxiv:1406.6793 [ hep - th ] ] . m. bullimore , h. c. kim and p. koroteev , jhep * 1505 * , 095 ( 2015 ) [ arxiv:1412.6081 [ hep - th ] ] .
m. bullimore and h. c. kim , jhep * 1505 * , 048 ( 2015 ) [ arxiv:1412.3872 [ hep - th ] ] .
b. haghighat , c. kozcaz , g. lockhart and c. vafa , phys .
d * 89 * , no . 4 , 046003 ( 2014 ) [ arxiv:1310.1185 [ hep - th ] ] .
j. kim , s. kim and k. lee , jhep * 1602 * , 170 ( 2016 ) [ arxiv:1503.07277 [ hep - th ] ] .
m. billo , m. frau , l. gallot , a. lerda and i. pesando , jhep * 1311 * , 123 ( 2013 ) [ arxiv:1307.6648 [ hep - th ] ] . l. f. alday , d. gaiotto and y. tachikawa , lett . math .
phys . * 91 * , 167 ( 2010 ) [ arxiv:0906.3219 [ hep - th ] ] .
n. drukker , d. gaiotto and j. gomis , jhep * 1106 * , 025 ( 2011 ) [ arxiv:1003.1112 [ hep - th ] ] .
k. hosomichi , s. lee and j. park , jhep * 1012 * , 079 ( 2010 ) [ arxiv:1009.0340 [ hep - th ] ] .
d. galakhov , a. mironov and a. morozov , jhep * 1406 * , 050 ( 2014 ) [ arxiv:1311.7069 [ hep - th ] ] .
n. nemkov , jhep * 1510 * , 039 ( 2015 ) [ arxiv:1504.04360 [ hep - th ] ] .
o. aharony , m. berkooz , s. kachru , n. seiberg and e. silverstein , adv .
* 1 * , 148 ( 1998 ) [ hep - th/9707079 ] . o. aharony , m. berkooz and n. seiberg , adv .
* 2 * , 119 ( 1998 ) [ hep - th/9712117 ] .
j. kim , s. kim , k. lee , j. park and c. vafa , arxiv:1411.2324 [ hep - th ] .
b. haghighat , a. klemm , g. lockhart and c. vafa , fortsch .
phys . *
63 * , 294 ( 2015 ) [ arxiv:1412.3152 [ hep - th ] ] .
a. gadde , b. haghighat , j. kim , s. kim , g. lockhart and c. vafa , arxiv:1504.04614 [ hep - th ] .
j. kim , s. kim and k. lee , arxiv:1510.03128 [ hep - th ] .
h. c. kim , s. kim and j. park , arxiv:1608.03919 [ hep - th ] .
m. del zotto and g. lockhart , arxiv:1609.00310 [ hep - th ] .
|
we study m5-branes compactified on @xmath0 from the d0-d4 witten index in the coulomb phase .
we first show that the prepotential of this index is s - dual , up to a simple anomalous part .
this is an extension of the well - known s - duality of the 4d @xmath1 theory to the 6d @xmath2 theory on finite @xmath3 . using this anomalous s - duality
, we find that the asymptotic free energy scales like @xmath4 when various temperature - like parameters are large .
this shows that the number of 5d kaluza - klein fields for light d0-brane bound states is proportional to @xmath4 .
we also compute some part of the asymptotic free energy from 6d chiral anomalies , which precisely agrees with our d0-d4 calculus .
addtoresetequationsection + * asymptotic m5-brane entropy from s - duality * seok kim and june nahmgoong _ department of physics and astronomy & center for theoretical physics + seoul national university , 1 gwanak - ro , gwanak - gu , seoul 08826 , korea . _
+ e - mails : [email protected] , [email protected]
|
slow , velocity - controlled molecular beams have become an attractive tool in the study of molecular dynamics , the measurement of fundamental constants and in conducting high resolution spectroscopy @xcite .
many advances have been made to produce slow beams such as using a high - pressure seeded expansion @xcite , pre - cooling from a buffer gas source @xcite , or by passing the gas through a cryo - cooled valve nozzle @xcite or a room - temperature nozzle spinning in a counter - rotating configuration @xcite .
molecules in seeded supersonic beams can also be decelerated by undergoing inelastic collisions in a crossed beam set up to remove kinetic energy @xcite , or by using the interaction of the electric or magnetic dipole moments of the molecules with external optical dipole , electric , or magnetic fields @xcite .
many of these recent advances in the production of cold atomic and molecular beams would benefit from using the shortest gas pulse that is compatible with maintaining a high number density , narrow velocity distribution and cold rotational distribution .
very short pulsed beams produce fewer gas molecules in the vacuum chamber , thereby lowering the gas load on the pumps for the experimental apparatus , and reducing background collisions with the beam ; the low velocity spread and cold internal temperature of the molecules in the beam are thus preserved @xcite .
short - pulse molecular beam sources are primarily pulsed valves that use modified , fast - acting , resonantly - driven valve components .
for example , the jordan pulsed supersonic valve provides cold molecular beams of duration during choked flow and during low flow @xcite .
this valve operates on the magnetic beam repulsion principle by passing opposing currents through two parallel - beam conductors , one of which is a thin metal strip that blocks the gas outlet .
the currents produce a magnetic force which moves the strip and thus opens the valve @xcite .
another fast - acting pulsed valve was recently constructed by janssen and colleagues ; it uses a cantilever piezo to produce molecular beam pulses of duration at high repetition rates of @xcite .
the even - lavie pulsed valve uses light , miniaturized components to enable operations at high backing pressures of up to , resulting in a high intensity molecular beam with a duration of and with very cold rotational temperatures for entrained molecules of around @xcite .
most recently , the nijmegen pulsed valve was constructed , by pulsing a current through an aluminum strip located within a magnetic field .
the lorentz force causes the strip to break its seal to the nozzle in a periodic fashion , resulting in molecular beams with durations as low as and a rotational temperature of around @xcite .
these beam sources are typically constructed from many small individual components .
the high surface area to volume ratio ratio of these components makes them susceptible to corrosion from redox reactions with chemically reactive seed molecules , such as halogen species @xcite .
thus , the use of a valve that contains a few robust components , while still realizing a full expansion of the molecular beam , is advantageous to experiments using reactive seed molecules .
the commercially available general valve series 9 valve system has a relatively simple design , and the components are robust against corrosion . the main body is constructed with corrosion - resistant stainless steel , and a solenoid is used to drive a teflon cylinder with a conical tip , which breaks the seal to the valve nozzle , producing a fully - expanded molecular beam @xcite .
the teflon cylinder can be easily replaced , and the buildup of halogen - induced oxidation on the metal components can be easily removed and sanded off without changing the shape of the valve geometry .
thus , the valve is fully compatible with halogen - seeded molecular beams , and residue from the oxidation of the halogen molecules on the valve can be easily washed off with deionized water or acetone .
our current work uses a series 9 valve to generate a molecular beam of br@xmath0 seeded in ar . in these experiments ,
near - threshold photolysis of br@xmath0 produces cold br atoms in a so - called ` photostop ' experiment , and the atoms are contained in a permanent magnetic trap on the molecular beam axis for up to @xcite .
a disadvantage of using this type of valve is that the beam pulse duration is typically on the order of a hundreds of microseconds to a few milliseconds , so the high density of carrier - gas ( ar ) atoms in the tail of the pulsed beam ( typically 10@xmath3 @xmath4 at the center of the trap ) is the dominant cause of trap loss of br atoms @xcite . in order to accumulate density of br atoms in the trap from successive pulses of the beam ,
we need to minimize the trap loss in the time interval between the successive photolysis pulses .
thus , we need to reduce the pulse - duration of the molecular beam , both to minimize collisions with the tail of the beam , and also to reduce the background pressure in the chamber .
an alternative method of producing short gas pulses , in conjunction with a valve that initially generates long pulses , is to mechanically block the majority of the original molecular beam pulse , and only open the beam path in a narrow time window to let a thin slice of the pulse width through to the experimental chamber .
_ constructed a molecular beam chopper using an electromagnet to resonantly drive a magnetic field across a contact pin .
this oscillation resulted in the blocking and unblocking of the molecular beam as it passed by the contact pin , with the varying density of the beam closely matching that of a sinusoidal function @xcite .
a more common setup is the use of a spinning chopper wheel @xcite .
the molecular beam is directed perpendicularly to a spinning disk , into which slots are cut .
the rotation of the disk is synchronized to the gas pulse so that only a portion is allowed through the slot , and the rest of the beam is blocked .
et al . _ also constructed a chopper set up by rotating a copper disk at to produce molecular beam pulse durations of around , in order to measure the velocity distributions and translational temperatures of the molecular beam from a series 9 valve @xcite .
however , two disadvantages of using a chopper wheel to shorten molecular beams have prevented this method from dominating over modified pulsed valves .
first , the relatively low mechanical speed of rotation used for chopper wheels to date limits the shortest gas pulse attainable .
increasing the disk radius of the chopper increases the tangential velocity , and hence reduces the pulse width for a given slot dimension .
but a larger disk also increases the mechanical and heat load on the system , and the radius is typically limited by the geometry of the vacuum chamber and the ability to balance the disk onto the motor at high revolution speeds .
second , the large portion of the molecular beam that does not pass through the chopper slit may contribute to the background pressure in the vacuum chamber . in some cases
this gas buildup has been mitigated by placing the chopper wheel along the path of a set of skimmed differentially pumped stages @xcite . in this paper , we demonstrate the use of a new high - speed chopper wheel construct to produce a gas pulse .
we furthermore show that this technique can operate with a simplified differential pumping stage .
we have applied this technique to shorten beams seeded with corrosive no and br@xmath0 molecules , for which other short - gas - pulse sources may have limited durability .
a schematic of the experimental setup is shown in figure [ fig : expsetup ] .
the pulsed valve ( general valve , series 9 ) generates a molecular beam , which passes through two skimmers separated by the chopper wheel and passes through skimmer b into the differentially - pumped detection chamber .
the chopper , its motor , and the pulsed valve are housed within the stainless steel ` top hat ' , located at the end of the source chamber .
the source chamber is pumped by a diffusion pump , while the detection chamber is pumped by a turbo pump .
the chopper wheel , shown in figs .
[ fig : expsetup ] and [ fig : expsetup2 ] , is a diameter aluminum ( 7075 aircraft alloy ) disk with two -wide slits cut at 180@xmath5 into the radius of the disk .
the disk is thinned at the edge to less than , and directly coupled to the shaft of a diameter brushless dc motor ( maxon motor model ec25 ) and balanced at up to in ambient conditions .
as shown in fig .
[ fig : expsetup ] the chopper wheel is placed between skimmer - a and with the slitted edge aligned with the center of the molecular beam .
the chopper is driven by a home - built controller which generates the ac wave forms used to drive the brushless motor at the desired speed , ranging from to .
the motor is cooled by contact with the wall of the vacuum chamber , and remains below at speeds up to .
this brushless dc motor is highly suitable for this application and its characteristics make possible the high speed of the chopper .
it is compatible with a vacuum environment due to it having fewer components than other types of motor ; it has a high torque output per power input due to the low friction around its bearings ; it has low electromagnetic interference ; it has an ability to operate without requiring the use of additional cooling agents such as airflow or water @xcite ; finally , the motor uses ball bearings which are vacuum compatible and minimizes the buildup of heat around the rotor responsible for limiting the speed of the motor .
skimmer - a ( orifice diameter ) is rigidly mounted to the general valve , as shown in figure [ fig : expsetup ] .
this rigid mounting mechanically aligns to the valve orifice at a distance of from the skimmer tip .
the mount is surrounded by a shroud made of plastic polyethylene terephthalate ( pet ) foil , which forms an effective differential pumping stage .
this setup enables only the central part ( diameter ) of the molecular beam , passing through the first skimmer , to reach the region of the chopper wheel .
the rest of the beam traveling off - axis is directed by the pet foil to the diffusion pump , which has a pumping speed sufficient to allow the background gas to be pumped away before further buildup from the next gas pulse .
any gas escaping around the edge of the shroud has a long enough diffusion time that it does not interfere with the fast on - axis beam .
the molecular beam pulse shortened by the chopper wheel subsequently passes through a skimmer ( b ) , which blocks further background gas scattered by the chopper wheel from reaching the detection chamber . in the detection chamber
, the beam passes through a series of electrostatic plates , labeled g ( ground ) , e ( extractor ) , m ( magnet ) , and r ( repeller ) , designed for time - of - flight detection of ions , produced by resonance - enhanced multiphoton ionization ( rempi ) at a position close to plate m. two ndfeb magnets are mounted on plate m , and can be used to confine low - velocity paramagnetic species in the photostop experiment ( but are not used as such in the experiments reported here ; see @xcite for further details ) .
two laser beams pass between the two magnets parallel to plate m in counter - propagating directions .
laser 1 , a -pulsed dye laser beam ( pulse length ) is used to photodissociate molecules in the beam ( br@xmath0/br experiments only ) .
laser 2 , a frequency doubled dye laser , is the probe laser and is used for rempi of either the no molecules or the br fragments .
the chopper generates the clock signal used to trigger the pulsed valve and the probe laser(s ) .
the chopper itself serves as the base clock , thereby eliminating the need to synchronize the chopper phase to an external clock . as the edge of one of the two chopper slits passes between the photodiode ( shown in figure [ fig : expsetup2 ] ) and a led light , a trigger pulse is passed back to the chopper controller through a pair of vacuum feedthroughs ( allectra gmbh ) .
this trigger pulse train is corrected to a -wide pulse , independent of rotation frequency , and a gate allows only the first pulse in every through to the output .
this output is therefore synchronized to the chopper and nominally at a frequency demanded by the pulsed laser system .
the actual frequency is allowed to vary between and to maintain the synchronization , with negligible impact on the laser power .
this pulse triggers the input of a delay generator ( quantum composer 9500 plus series ) , which in turn , triggers the pulsed valve ( after delay a ) and the probe lasers after a further delay b , as shown in figure [ fig : timingpulse ] .
the terminal velocity of a molecular beam , @xmath6 , is given by @xcite @xmath7 where @xmath8 is the mean heat capacity ratio of the mixture , calculated to be @xmath9 for a weighted average of 8% br in ar and @xmath10 for 8% no in ar ; @xmath11 is the boltzmann constant , @xmath12 is the initial temperature of the gas in the reservoir ( @xmath13 ) , and @xmath14 is the weighted average mass of the mixture @xcite . the standard deviation of velocities around @xmath6 in the moving frame is related to the translational temperature of ar in the moving frame , @xmath15 . @xmath16 the packet of molecules that passes through the chopper has a length @xmath17 ( immediately after the chopper wheel ) that is inversely proportional to the rotational frequency of the wheel and is given by @xmath18 where @xmath19 is the width of the slit tangential to the edge of the chopper ( equal to ) @xmath20 is the radius of the chopper wheel , and @xmath21 is the rotational frequency of the chopper in revolutions per second ( note that in the results reported below , the rotation frequency of the chopper wheel is quoted in revolutions per minute ( rpm ) ) .
the moving - frame temperature causes longitudinal expansion of the packet during this flight time , with the spatial spreading given by @xmath22 where @xmath23 is the distance between the chopper and intersection with the detection laser , and @xmath24 is the mean time taken for the packet to travel that distance . using equations ( [ eq : w0 ] ) and ( [ eq : ws ] ) , the length of the detected packet at the laser detection volume , @xmath25 , is obtained from the convolution of the length of the initially chopped packet with that due to its longitudinal velocity distribution @xmath26 .
a gaussian packet and velocity distribution is assumed to give @xmath27 at low rotational frequency of the chopper , the spatial packet size of the molecular beam is determined by the first term in equation [ eq : wbeam ] , while at high frequency it is determined by the spread of velocities around the moving frame , @xmath26 , as indicated by the second term .
the chopper setup was first tested with a molecular beam composed of 8% no seeded in @xmath28 ar backing gas .
no was chosen due to its strong , well - resolved , rotational lines in the rempi spectrum ( enabling ease of calibrating the rotational temperature of the molecular beam from the intensities ) and the convenient wavelength of the probe laser in our setup .
the photodissociation laser was not used for these experiments .
the no is detected by ( 1 + 1 ) rempi via the @xmath29 transition at a wavelength around @xmath30 @xcite .
the fundamental wavelength was generated from a pulsed dye laser ( sirah continuum , @xmath31 dye ) and doubled with a bbo crystal .
the no@xmath32 cations produced by rempi close to plate m are guided towards the mcp detector ( absent from figure [ fig : expsetup ] ) , by the field applied across plates r , m and e , using a wiley - mclaren configuration @xcite . the rotational temperature of no in the molecular beam was determined by recording the rempi signal of no while scanning the detection wavelength @xmath33 to @xmath34 .
the rotational line intensities were fitted to both a boltzmann plot , and also to the full simulation of the spectrum using the simulation program pgopher @xcite . using both methods ,
the distance between the chopper and the pulsed valve was set such that a lowest measured value of the rotational temperature of the no in this molecular beam was obtained at .
the low rotational temperature indicates that there was little or no buildup of background gas behind the chopper wheel to interfere with the molecular beam passing through it .
the no experiment was used to test the feasibility of the effective differential pumping stage using the pet foil .
figure [ fig : plasticshield ] shows the signal of no with the probe laser fixed at a wavelength of @xmath30 ( corresponding to the position of the overlapping q@xmath35 ( 0.5 ) r@xmath36 lines ) and taken with a chopper speed of , with and without this plastic shroud .
the delay between the synchronized clock signal from the chopper and the probe laser is scanned to record the temporal profile of the beam passing through . in both cases ,
a clear sharp signal of no indicates that a molecular beam pulse of the same duration passes through the chopper wheel , as shown at a delay of around in figure [ fig : plasticshield ] .
however , in the case without using the pet foil , a broad intense signal of no is observed at longer time delays for which we would expect the chopper to have shut off the path of the molecular beam .
the broad signal results from gas that does not pass through the first skimmer but diffuses to the detection region .
the broad signal is eliminated when using the pet foil ; the foil acts as a differential pumping stage , and therefore allows observation of a clean sharp peak in the time of flight . for a range of revolution
speeds , the temporal profile of the no rempi signal was recorded with delay * a * set to a constant optimized value of to allow the largest gas density of the original molecular beam packet to reach the chopper slit .
figure [ fig : golayfit ] shows an example of such a profile at .
a savitzky - golay method was used to smooth the raw data ( black stars ) @xcite .
this method applies a 2nd order polynomial least - square fit to the data set , as shown in figure [ fig : golayfit ] , where the and curves in the fit above the @xmath37 axis indicate fine and rough fits , respectively , according to the number of data points used in the smoothing algorithm .
next , the first - derivative is taken from these curves , and the savitzky - golay method is applied again , as shown by the curves located below the zero @xmath37 axis in figure [ fig : golayfit ] .
the beam width is defined as the difference between the delays on the representing the peak and zenith of this final , smoothed , first derivative . the beam width for the profile in figure [ fig : golayfit ]
is determined to be and , from the fine and rough fits , respectively .
a series of these beam widths was determined for no over a range of revolution speeds of the chopper wheel , from to , and the values are plotted as the orange points in figure [ fig : golay ] .
, described by equation [ eq : wbeam ] is represented by the solid - orange and dashed - purple curves for no and br@xmath0 respectively .
] each data point is an average of three beam - profile measurements , each fitted using a rough fit ( dashed - blue in figure [ fig : golayfit ] ) to provide an upper limit to the calculated beam widths .
the error bars show the variation to one standard deviation of this fluctuation over repeated runs .
thus the value of the lowest averaged beam width of for a revolution speed of is greater than some of the individually measured beam widths , such as that shown in figure [ fig : golayfit ] .
the theoretical model value of @xmath25 , defined by equation [ eq : wbeam ] , was used to calculate the expected temporal width of the pulse and is plotted for an no seeded beam ( solid orange line ) in the figure . in both the experimental measurements and theoretical model , beam widths at revolution speeds beyond converge toward a single value on the order of , limited by the longitudinal velocity distribution , @xmath26 , of the molecular beam . to show that the results from the chopper were consistent with other seed molecules , particularly with corrosive halogen species - a key objective of our work - measurements were also taken with a molecular beam of 8% br@xmath0 seeded in ar at a backing pressure of . to detect the temporal profile of the br@xmath0 molecules arriving at the laser detection volume , a nd : yag pumped pulsed dye - laser saturates the transition , leading to prompt dissociation to form br fragments - for each dissociated br@xmath0 , a pair of ground state br(@xmath38 ) and spin - orbit excited br(@xmath39 ) fragments are produced .
the ground state br(@xmath38 ) atoms are ionized via ( @xmath40 ) rempi at @xmath30 by the probe laser @xcite and directed to the mcp detector . the dissociation laser is fired prior to the rempi probe laser , to ensure that the br atoms are detected at the instant they are dissociated in the probe laser volume . note that the primary reason for using this two - step detection of br@xmath0 was that our laser system was set up for br detection in the photostop experiments @xcite .
as shown in figure [ fig : molbeamcompare ] , without the use of a chopper ( black line ) , the duration of the molecular beam has a full - width half - max ( dashed lines ) of , whereas with a chopper running at ( blue ) around 75% of its maximum speed the duration of the beam is reduced to .
seeded in 3 bar ar without the presence of a chopper ( black ) and with a chopper running at ( blue ) .
the dotted lines show boundary of the full - width half - maximum ( fwhm ) molecular beam profile without the chopper . ] in figure [ fig : molbeamcompare ] , the molecular beam profile has been smoothed using a savitzky - golay algorithm on the experimental data points . note
that although the intensities of each signal are shown as having the same peak value , the relative normalization is only approximate ; the removal or re - installation of the chopper requires substantial changes to the setup and therefore it is difficult to ensure that there is no residual dependence of the signal intensity on the power and the alignment of the dissociation and probe lasers in the two sets of measurements . in practice
such measurements are separated by days .
nevertheless it is clear that there is substantial shortening of the molecular beam pulse . in figure
[ fig : golay ] the measured and calculated pulse widths for the bromine beam are plotted in purple , and these are seen to fit on the same line as the data for no , verifying the simple analytical model .
the chopper wheel was designed to ensure that the arrival of the next slot on the chopper wheel will not overlap with the front or tail of the molecular beam . at the highest revolution speed of
, the period between the slits located on opposite sides of the chopper wheel is .
thus , even if the very front of the molecular beam were to pass through the chopper slit , the tail of the beam would have been blocked by the chopper wheel before the next slit intersects the molecular beam axis .
it is interesting to note in the context of the br photostop experiments , that for the highest chopper wheel rotational speeds , there is a strong correlation between the longitudinal velocity of a br@xmath0 molecule and its arrival time at the laser probe volume - faster molecules in the distribution arrive earlier .
hence the nanosecond - pulse dissociation laser picks out a subset of the molecular beam pulse with a narrow velocity distribution - much narrower than for the overall pulse .
thus , in the photostop experiment the wavelength of the dissociation laser needs to be specified more precisely in order to match the molecular frame photofragment velocity to the lab - frame molecular beam velocity .
this means that a higher proportion of the molecules that are dissociated will be photostopped , which is beneficial for the trapping experiment .
in summary , a novel home - built molecular - beam chopper system has been used to shorten the duration of a molecular beam by a factor of ten to .
our design demonstrates that a second differential pumping stage is not needed by mounting a second skimmer ( a ) to the pulsed valve , attached with a plastic shroud to guide the skimmed molecules toward the diffusion pump . with just a few robust components ,
this method is particular useful for experiments using corrosive species such as halogens as the seed molecules in molecular beams .
the use of a brushless dc motor enables the chopper to run at a range of speeds from to , limited only by the cooling of the motor .
this range of speeds produces a range of molecular beam widths ranging from to .
the beam widths are repeatable and fall within the same error bars , regardless of whether no or br@xmath0 is used as the seed molecule .
measurements of the beam widths as a function of chopper wheel revolution speed closely match our theoretical model , @xmath25 , plotted in figure [ fig : golay ] .
the ability to use such a simple , yet durable system , offers an opportunity to expand the breadth of chemical systems for which short - pulse molecular beams can be used .
further work can be done to decrease the pulse duration below , by a combination of increasing the revolution speed of the chopper wheel , which requires additional cooling of the motor rotor to maintain its temperature below , and by cooling the valve to decrease the spread of velocities around the moving frame of the carrier gas .
tps acknowledges the financial support of the epsrc under grants ep / g00224x and ep / i029109 and the wiener anspach foundation .
jl is grateful for support from the nsf fellowships program and cjr from the ramsay memorial trust .
54ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1002/anie.200805503 [ * * , ( ) ] link:\doibase 10.1038/nphys1031 [ * * , ( ) ] link:\doibase 10.1039/b820625b [ * * , ( ) ] link:\doibase 10.1140/epjd / e2004 - 00160 - 9 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.143004 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.012702 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.173201 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2717178 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1505662 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.72.3166 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.033409 [ * * , ( ) ] link:\doibase 10.1021/jp993560x [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4727883 [ * * , ( ) ] link:\doibase 10.1126/science.1090679 [ * * , ( ) ] , link:\doibase 10.1038/nphys339 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.83.1558 [ * * , ( ) ] link:\doibase 10.1103/physreva.60.3882 [ * * , ( ) ] link:\doibase 10.1103/physreva.75.031402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.093003 [ * * , ( ) ] @noop _ _ , vol .
( , ) http://scitation.aip.org/content/aip/journal/rsi/49/5/10.1063/1.1135470 " , doi = " http://dx.doi.org/10.1063/1.1135470 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1136813 [ * * , ( ) ] @noop `` , '' ( ) , @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) https://books.google.co.uk/books?id=x712zmzhkvmc[__ ] , referex engineering ( , ) @noop _ _ , ed .
( , ) @noop _ _ ( , ) @noop _ _ , ed .
( , ) \doibase http://dx.doi.org/10.1063/1.1145444 [ * * , ( ) ] link:\doibase 10.1039/c0cp02472d [ * * , ( ) ] @noop * * , ( ) @noop * * , ( )
@noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop `` , '' ( ) , @noop `` , '' ( ) , @noop `` , '' ( ) , @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.1715212 [ * * , ( ) ] @noop `` , '' ( ) , @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( )
|
a chopper wheel construct is used to shorten the duration of a molecular beam to .
molecular beams seeded with no or with br@xmath0 and an initial pulse width of @xmath1 @xmath2s were passed through a spinning chopper wheel , which was driven by a brushless dc _ in vacuo _ motor at a range of speeds , from to .
the resulting duration of the molecular - beam pulses measured at the laser detection volume ranged from to , and was the same for both no and br@xmath0 .
the duration is consistent with a simple analytical model , and the minimum pulse width measured is limited by the spreading of the beam between the chopper and the detection point as a consequence of the longitudinal velocity distribution of the beam .
the setup adopted here effectively eliminates buildup of background gas without the use of a differential pumping stage , and a clean narrow pulse is obtained with low rotational temperature .
|
deep multiwavelength surveys are progressively sheding light on the history of galaxy evolution since very high redshifts . from new surveys in the ultraviolet regime ( e.g. martin et al . 2005 ) to the latest ir , sub - millimetre and radio campaigns ( e.g. werner et al . 2004 ; dunlop 2005 ; condon et al .
2003 ) , the advent of numerous datasets is enhancing our understanding of when galaxies were formed and how they evolved through cosmic time .
uv / optical observations can unveil the sources of stellar emission but are limited to detect unobscured systems , missing an important fraction of the galaxies with on - going star formation .
ir and longer - wavelength surveys provide an unbiased way of studying star - forming systems , but are insensitive to galaxies with little or no dust content . near - ir
observations , on the contrary , appear as one of the most suitable methods for making a complete census of galaxy populations , as they are sensitive to the stellar emission from both young and old galaxies , and are also relatively unaffected by the presence of dust .
the rest - frame @xmath15 band is the most efficient tracer of the stellar mass content of the universe .
recent studies of near - ir - selected galaxies allowed for a rapid progress in setting different constraints on the history of galaxy formation ( cimatti et al .
2002 ; daddi et al .
2003 ; caputi et al .
2004 ; caputi et al . 2005a ; among others ) and , in particular , on the evolution of the stellar mass content of the universe with redshift ( dickinson et al . 2003
; fontana et al . 2004 ; glazebrook et al . 2004 ; caputi et al . 2005a ; drory et al . 2005 ) .
it is now recognised that , although more than a half of the stellar mass has been formed at redshifts @xmath16 $ \sim$}}~}1.5 $ ] , near - ir - selected galaxies are already present before at least redshifts @xmath17 and contain a significant fraction of the stellar mass observed at redshift @xmath5 .
however , one of the main problems of existing near - ir surveys is that , when sufficiently deep , they are usually restricted to small areas of the sky , making the results prone to large error bars due to cosmic variance . also , the lack of statistics prevents refined studies of galaxy evolution , where trends for different near - ir sub - populations can be individually investigated .
the goods project ( giavalisco et al . 2004 ) is a public program of multiwavelength observations which is now close to completion .
the data sets are kindly released to the astronomical community in a fully reduced mode . as part of the goods project , @xmath18 of the cdfs
have been observed with the advanced camera for surveys ( acs ) on board the hubble space telescope ( hst ) in four passbands : @xmath3 , @xmath19 , @xmath20 and @xmath21 .
in addition , the goods project includes deep near - ir imaging taken with the infrared spectrometer and array camera ( isaac ) on the ` antu ' very large telescope ( antu - vlt ) . a first release of the european southern observatory ( eso ) imaging survey ( goods / eis v0.5 release ) provided data in the @xmath22 , @xmath23 and @xmath24 bands for @xmath25 of the goods / cdfs field . in this sub - region , roche , dunlop & almaini ( 2003 ) selected a deep sample of extremely red galaxies ( ergs ) with @xmath26 ( vega ) , whose redshift distribution and luminosity evolution were studied by caputi et al .
a detailed comparison of the properties of ergs with other @xmath2-selected galaxies was presented by caputi et al .
( 2005a ) . over the last year
, however , new datasets have been made public within the goods project , increasing the coverage in wavelength and area of the goods / cdfs .
the second goods / eis ( v1.0 ) release extends the deep @xmath22 and @xmath24 band imaging to @xmath1 of the goods / cdfs . also ,
the infrared array camera ( irac ) on board the spitzer space telescope imaged this field at @xmath27 , @xmath28 , @xmath29 and @xmath30 .
the availability of these data allows one to map the rest - frame @xmath2-band light up to redshifts @xmath17 . in this work , we take advantage of the further wavelength and area coverage of the goods / cdfs , in conjunction with the information provided by the latest released spectroscopic and combo17 data , to set tighter constraints on the properties and evolution of @xmath2-selected galaxies . the layout of this paper is as follows .
first , in section [ sec_sample ] , we present the sample selection and give details on the multiwavelength photometry . in section [ sec_redsh ] , we explain the technique we applied to construct an optimised redshift catalogue for our galaxy sample and show the resulting redshift distribution . in section [ sec_lf ]
, we present the rest - frame @xmath2-band luminosity function ( lf ) for our sample of @xmath2-selected galaxies . in section [ sec_mass ] , we discuss several aspects of the evolution of different mass systems . in section [ sec_nearir ] , we explore additional hints from the study of the purely near - ir colours of our galaxies . finally , in section [ sec_conc ] we present some concluding remarks .
we adopt throughout a cosmology with @xmath31 , @xmath32 and @xmath33 . a salpeter ( 1955 ) imf over stellar masses @xmath34 is assumed , unless where otherwise stated .
we used the public deep isaac @xmath2-band images in the goods / eis v1.0 release to select a catalogue of @xmath0 ( vega ) sources in 131 arcmin@xmath35 of the goods / cdfs .
this new release is composed of 21 maps which cover more than twice the area of the goods / eis v0.5 release .
in addition , further improvements have been applied to the v1.0 imaging , such as a revised data reduction technique and a uniform photometric calibration of the maps across the entire field .
we performed the source extraction using the public code sextractor ( bertin & arnouts , 1996 ) on the @xmath2-band images .
we cut the resulting catalogue at a magnitude @xmath36 , a limit half a magnitude brighter than the cut we used to select our @xmath2 sample in caputi et al .
( 2005a ) .
several reasons justify our new magnitude limit .
first , a few of the additional maps in the v1.0 release have lower exposure times than the average exposure times of the maps composing the v0.5 release .
second , this new limit constitutes an adequate balance between having access to galaxies up to very high redshifts and requiring for our catalogue to have a high level of completeness .
the completeness levels of our v1.0 sample are very similar to those estimated for the v0.5 sample by caputi et al .
( 2005a ) : 100% , 90% and 86% for @xmath37 , @xmath38 and @xmath39 , respectively .
finally , our new sample has been selected having in mind the possibility of a complete quantitative study of galaxy morphology , only possible for high signal - to - noise objects .
the results of the latter will be presented in a companion paper ( mclure et al .
2005 , in preparation ) .
our total v1.0 @xmath0 sample consists of 3184 sources .
we measured multiwavelength photometry for our @xmath2-selected galaxies in order to model their spectral energy distribution ( sed ) , as we shall explain in section [ sec_redsh ] .
we ran sextractor in ` double - image mode ' to perform aperture photometry on the goods / eis v1.0 @xmath22-band images , centred at the position of the @xmath2-band extracted sources .
we measured magnitudes in an aperture size of 2.83-arcsec diameter for all the objects , to coincide with one of the aperture sizes used to construct the public goods / cdfs acs catalogues .
the latter have been constructed running sextractor on the @xmath21-band images , and making ` double - image mode ' runs to perform photometry in the @xmath3 , @xmath19 and @xmath20 bands .
we looked for counterparts of our @xmath0 ( vega ) sources in the acs catalogues and adopted the corresponding 2.83-arcsec diameter aperture magnitudes in each of the four acs passbands .
a 2.83-arcsec diameter aperture is sufficient to collect all the flux of the vast majority ( @xmath4090% ) of our galaxies in the different optical / near - ir passbands . even in the @xmath2 band ,
virtually all of our galaxies have 2.83-arcsec - diameter aperture magnitudes in agreement with the corresponding total @xmath2 magnitudes within the error bars .
the ( @xmath41 ) versus ( @xmath42 ) colours for the sources in the @xmath0 sample .
cross - like symbols correspond to those objects with sextractor stellarity parameter [email protected] . ]
deep spitzer / irac imaging for the goods / cdfs is now also publicly available , covering the wavelength range @xmath44 to @xmath45 . at the time of writing ,
only the first epoch of irac goods / cdfs data had been released .
this first epoch includes @xmath44 and @xmath46 images for @xmath47 of the area of the goods / cdfs and @xmath4 and @xmath45 images for another @xmath47 of this area , in such a way that only @xmath48 of the field have data in the four irac passbands .
we ran sextractor on the @xmath44 and @xmath4 images and looked for counterparts of the @xmath2-band sources in a matching radius of 1.2 arcsec .
we restricted the photometry to only the first two irac channels , as the templates we used for the galaxy sed fitting lose accuracy or do not cover wavelengths beyond @xmath49 .
we measured magnitudes in 2.83-arcsec diameter apertures .
however , the point - spread function ( psf ) of the irac images is much larger than the characteristic psf of the isaac and acs images : the mean full - width half maximum ( fwhm ) of the irac @xmath44 psf is 1.66 arcsec ( fazio et al .
2004 ) , in comparison with 0.4 arcsec and 0.1 arcsec for the @xmath2 and @xmath21 bands in the isaac and acs instruments , respectively .
thus , a 2.83-arcsec diameter aperture is insufficient to encompass the magnitudes measured in the irac bands with the magnitudes obtained in equivalent apertures on the acs and isaac images .
we computed aperture corrections for the magnitudes measured on the irac images studying the curve of growth of isolated stars in the field .
we determined aperture corrections of 0.50 and 0.55 magnitudes for the 2.83-arcsec - diameter aperture magnitudes in the @xmath44 and the @xmath4 bands , respectively . to clean our catalogues for stars and active galaxies , we applied criteria somewhat more strict than those implemented by caputi et al .
( 2005a ) .
we cleaned our catalogues for stars using the sextractor stellarity parameter ` class_star ' , as measured on the acs @xmath21 images .
to separate active galaxies , we looked for counterparts of our @xmath0 sources in the public x - ray catalogues available for the cdfs ( szokoly et al .
2004 ) .
222/3184 sources of our initial catalogue had [email protected] .
figure [ colstar ] shows the location of these 222 sources in a colour - color diagram , in comparison to the remaining sources in the catalogue .
we see that the majority of the objects with [email protected] occupy a clearly separated region in this plot , typically displaying colours @xmath50 $ \sim$}}~}1.1 $ ] ( vega )
. however , we find 6 outliers in the colour - colour diagram , which we studied on a case - by - case basis .
4 out of these 6 outliers are x - ray classified as active galactic nuclei ( agn ) or quasi - stellar objects ( qso ) .
the remaining 2 are not detected in the x - rays , but their seds can not be satisfactorily fitted by hyperz ( bolzonella , miralles & pell , 2000 ) with any of the templates in the gissel98 library ( bruzual & charlot 1993 ) , so they are likely to be also stars or agn / qso . as a consequence , we excluded the 222 sources with [email protected] from our sample . on the other hand , we found that 142/3184 of our @xmath2 sources were detected in the x - rays .
however , in principle , the fact that a galaxy is classified as active in the x - rays does not necessarily imply that its optical - to - near - ir sed is dominated by an active nucleus . to determine which x - ray - detected @xmath2 sources had to be excluded from our catalogue , we modelled their seds using the templates in the gissel98 library of hyperz .
the gissel98 library contains a wide range of synthetic seds based on stellar spectra but no power - law sed template .
we determined that the seds of 57/142 x - ray - detected @xmath2 sources could not be satisfactorily fitted by any template in the gissel98 library , so we excluded these 57 objects from our catalogue , as it is likely that their optical / near - ir light is at least partially contaminated by an active nucleus component .
5 additional x - ray - detected @xmath0 sources did not have either a satisfactory fit of their seds , but they already had been excluded on the basis of the stellarity parameter criterion . in summary , our final @xmath0 source catalogue consisted of 2905 galaxies .
other no - x - ray - detected active galaxies could still be present in our final galaxy catalog .
another characteristic allowing to identify active galaxies is the power - law shape of their seds from the uv / optical through the ir regimes .
thus , to investigate the presence of additional contaminating sources within our sample , we cross - correlated our galaxies with the catalogue of @xmath51-selected ir power - law sources in the cdfs constructed by alonso - herrero et al .
( 2005 ) . within our sample
, we found only 8 no - x - ray - detected @xmath0 sources included in the alonso - herrero et al .
we conclude , then , that the fraction of remaining plausibly - contaminating sources within our sample is basically negligible .
a ) photometric versus spectroscopic redshifts for @xmath0 galaxies in the goods / cdfs .
b ) distribution of relative errors .
the median of this distribution is @xmath52 and the rms is @xmath53 . ] the redshift distribution of @xmath0 normal galaxies in the goods / cdfs . ]
we used the public code hyperz ( bolzonella , miralles & pell , 2000 ) with the gissel98 template library ( bruzual & charlot 1993 ) to model the sed and obtain a photometric redshift ( @xmath54 ) for each galaxy in our sample , using photometry in up to eight broad - bands ( @xmath55 and irac @xmath44 and @xmath4 bands ) .
only models with solar metallicity have been considered .
dust - corrections have been taken into account convolving the template sed with the calzetti et al .
( 2000 ) reddening law . as it was shown in caputi et al .
( 2005a ) , the code hyperz has a tendency to favour high - redshift estimates when degenerate solutions in redshift space exist .
to control the hyperz output , we applied the following set of constraints on the obtained photometric redshifts : * we used the @xmath56 relation for radio galaxies ( see figure [ hubble ] ) to determine the maximum possible redshift for each object .
radio galaxies delimit the behaviour of the most massive galaxies formed at high redshifts ( e.g. willott et al . 2003 ) and normal galaxies are not usually found below this limit ( caputi et al .
2004 ; caputi et al .
* the estimated redshift was constrained to the best - fit solution with a maximum value @xmath57 for all those galaxies with a counterpart in the public shallow u - band catalogues existing for the cdfs - band images exist for the cdfs ( arnouts et al .
2001 ) , they are shallower and poorer in resolution than the acs and isaac images used here and , thus , we decided not to incorporate these data for the input catalogue of the photometric redshift algorithms . ] , as higher redshift sources are unlikely to be bright at such short wavelengths .
similarly , the estimated redshift was constrained to a maximum value @xmath58 for those sources detected in the @xmath3 band .
* we used the public code bpz ( bentez 2000 ) to obtain a second , independent , set of redshift estimates for all the sources in our catalog . in the cases of sources with hyperz photometric redshifts @xmath59
not confirmed by bpz , we adopted the lower estimates given by the bpz code .
* we used public combo17 photometric redshifts ( wolf et al . 2004 ) to replace the estimates of those sources with magnitude @xmath60 ( vega ) and @xmath61 ( the best regime for combo17 redshifts ) .
observed @xmath62 magnitudes versus redshifts @xmath63 .
the solid line corresponds to the empirical @xmath56 relation for radio galaxies as obtained by willott et al .
( 2003 ) , which approximately coincides with the passive evolution of a @xmath64 starburst formed at redshift @xmath65 .
the dotted line is a nominal extrapolation of the same law . ]
figure [ zqual]a ) shows the resulting estimated redshifts versus spectroscopic redshifts for all those @xmath0 sources included in different publicly available spectroscopy samples in the goods / cdfs ( e.g. cimatti et al .
2002b ; le fvre et al .
2004 ; vanzella et al .
we observe a very good agreement between estimated and real redshifts in most cases .
figure [ zqual]b ) shows the corresponding distribution of relative errors : the median is @xmath52 and the standard deviation is @xmath53 .
however , in spite of the good accuracy obtained for the photometric redshifts , the incorporation of existing spectroscopic redshifts would be of benefit for the determination of our galaxy redshift distribution .
consequently , in order to optimise the quality of our final redshift catalog , we used the available spectroscopic redshifts for the cdfs to replace the photometric estimates of our galaxies whenever possible .
the final redshift catalogue for our 2905 @xmath0 galaxies is composed of 686/2905 ( 23.6% ) spectroscopic redshifts , 629/2905 ( 21.7% ) combo17 redshifts and 1590/2905 ( 54.7% ) hyperz / bpz redshifts .
figure [ zhisto ] shows the resulting redshift distribution of our sample of 2905 @xmath0 normal galaxies in the goods / cdfs .
81% of these galaxies lie at redshifts @xmath66 , while the remaining 19% is found to be at @xmath67 .
the two peaks of the distribution in the redshift bins @xmath68 $ ] and @xmath69 $ ] are produced by known large - scale structure in the cdfs ( e.g. le fvre et al . 2004 ) .
the resulting hubble diagram @xmath2 versus @xmath63 is shown in figure [ hubble ] .
[ cols="^,^,^,^,^,^,^",options="header " , ] caputi et al . ( 2005a ) computed and analysed the rest - frame @xmath2-band luminosity function ( lf ) from redshifts @xmath70 to @xmath6 . in this work ,
we take advantage of the larger area covered by the @xmath2-band goods / eis v1.0 release to set tighter constraints on the galaxy rest - frame @xmath2-band lf , and to model its evolution up to @xmath6 .
the incorporation of irac @xmath44 and @xmath4 data enables us to compute this lf with no need of magnitude extrapolations up to redshifts @xmath71 and @xmath72 , respectively , and with smaller extrapolations at higher redshifts .
anyway , we note that the k - corrections in the @xmath2-band are less dramatic than those in other shorter wavelength bands ( poggianti et al .
1997 ) .
the distribution of the differences between the rest - frame absolute @xmath2-band magnitudes @xmath73 obtained from the closest available rest - frame band and the ones obtained extrapolating from the observed @xmath2 magnitudes in all cases .
the distribution is centered at @xmath73 ( closest ) @xmath74 and the rms is @xmath75 . ]
we obtained the rest - frame absolute magnitude @xmath73 of each galaxy directly from the hyperz output .
for those objects with spectroscopic , combo17 or bpz redshifts , we adopted the best sed fitted by hyperz constrained to the predetermined redshift .
for the remaining objects , the best - sed fit and redshift estimation were simultaneously obtained .
hyperz computes the rest - frame absolute magnitude in the selected filter from the observed passband which is the closest to that filter in the rest - frame . in our case ,
the rest - frame @xmath73 magnitudes are obtained either from the @xmath2 , @xmath44 or @xmath4 apparent magnitudes , depending on the redshift of the source and the availability of irac data . with this ` closest band ' method , magnitude extrapolations are minimised .
it is interesting to compare the rest - frame absolute @xmath2-band magnitudes @xmath73 obtained with this procedure , and those obtained when no irac data is taken into account for the sed fitting , i.e when the rest - frame @xmath2-band magnitude extrapolations are directly made from the observed @xmath2-band in all cases ( as , for instance , in caputi et al .
2005a , where irac data were not available ) . figure [ mksmks ] shows the distribution of the difference between the absolute magnitudes @xmath73 obtained with the two different methods .
the statistics has been made over the magnitude range relevant for the computation of our lf ( see below ) , i.e. over all the sources with @xmath76 , as obtained with the ` closest - band ' method used here .
we see that , although the resulting distribution has a long tail , the standard deviation @xmath77 is small : the center of the distribution is located at @xmath73 ( closest ) @xmath74 and the rms is @xmath75 . thus , we conclude that , in most cases , the rest - frame absolute magnitudes @xmath73 obtained extrapolating the observed @xmath2-band magnitudes are consistent with those derived from the longer - wavelength data used here .
we computed the rest - frame @xmath2-band lf using two different techniques : the @xmath78 method and a maximum likelihood ( ml ) analysis ( marshall et al . 1983 ) . in the @xmath78 method ,
the lf is directly computed from the data , with no parameter dependence or model assumption .
the ml analysis , on the other hand , is a parametric technique which requires an _ a priori _ functional form for the lf and its evolution with redshift .
we parametrised the rest - frame @xmath2-band lf with a schechter function ( schechter 1976 ) : @xmath79 d\left(\frac{l}{l^\ast(z)}\right ) , \nonumber\\ \frac{l}{l^\ast(z ) } & = & 10^{-0.4(m - m^\ast(z ) ) } , \end{aligned}\ ] ] and we assumed that the density and luminosity evolved with redshift as @xmath80 , \\
m^\ast(z ) & = & m^\ast(z=0 ) - \left(\frac{z}{z_m}\right)^{k_m}.\end{aligned}\ ] ] the parameters @xmath81 and @xmath82 have been fixed to the local value : @xmath83 and @xmath84 ( kochanek et al .
the obtained best - fit values for the remaining ( free ) parameters are given in table [ tab_sch ] .
in figure [ kslf ] , we present the resulting rest - frame @xmath2-band lf at redshifts @xmath85 and @xmath86 . the empty symbols indicate the lf obtained with the @xmath78 method .
the values shown are strictly above the luminosity completeness limits of each redshift bin .
the number of sources in the luminosity - completeness range are 822 , 335 , 145 and 83 for the redshift bins @xmath87 , [ 1.25;1.75 ] , [ 1.75;2.25]$ ] and @xmath88 $ ] , respectively . to correct for the small incompleteness of our sample to @xmath36 , we weighted each source with a multiplicative factor depending on its apparent magnitude : 1.00 , 1.11 and 1.16 for sources with @xmath37 , @xmath38 and @xmath39 , respectively ( see section [ sec_sample ] ) .
these correction factors have a minor impact on the resulting lf ( cf .
caputi et al .
the error bars correspond to the maximum of poissonian errors and the errors due to cosmic variance .
we set a fixed value of 15% of the number counts for the cosmic variance error bar at @xmath89 $ \sim$}}~}1 $ ] ( cf .
somerville et al .
the solid lines show the evolution of the lf with redshift , as obtained through the ml analysis . in the four panels of figure [ kslf ]
, we have added for comparison the local k - band lf measured from two micron all sky survey ( 2mass ) data by kochanek et al .
( 2001 ; dotted lines ) .
the lfs obtained by the two different methods are in good agreement in the different redshift bins . from the upper - left panel in figure [ kslf ]
, we can confirm once more that the bright end ( @xmath90 $ \sim$}}~}-25 $ ] ) of the rest - frame @xmath2-band lf significantly increases from redshifts @xmath5 to @xmath11 .
this trend has already been observed in previous works ( drory et al .
2003 ; feulner et al .
2003 ; pozzetti et al .
2003 ; caputi et al . 2004
; caputi et al .
2005a ) . from our ml analysis
we find that , although the total density of objects @xmath91 decreases by a factor @xmath92 between redshifts @xmath5 and @xmath11 , the characteristic magnitude @xmath93 brightens in @xmath94 magnitudes between those redshifts .
the overall result is a rise in the density of objects with @xmath90 $ \sim$}}~}-25 $ ] .
recently , dahlen et al . (
2005 ) found that the characteristic magnitude @xmath93 of the rest - frame @xmath22-band lf suffered a dimming from redshifts @xmath95 to @xmath96 .
this result was apparently in contradiction with the brightening found by feulner et al .
( 2003 ) for the same lf in a similar redshift range .
however , as dahlen et al .
( 2005 ) explain ,
the one - to - one comparison of the different parameters characterising the schechter function is misleading , as @xmath93 , @xmath81 and @xmath91 are coupled , and different behaviours in the evolution of @xmath93 can still produce consistent lfs .
interestingly , on the other hand , dahlen et al .
( 2005 ) infer that , given the dimming they find for @xmath93 in the rest - frame @xmath22-band lf , the rest - frame @xmath2-band lf should present a similar behaviour .
they suggested that the brightening observed by different authors in the rest - frame @xmath2-band lf might be due to inaccuracies in the k - corrections used to extrapolate from the observed to the rest - frame @xmath2-band at @xmath72 , rather than being a real effect .
the incorporation of irac @xmath44 and @xmath4 data in the present work allows us to determine that this is not the case .
once more , one should conclude that a one - to - one comparison of the best - fit parameters obtained for the schechter function ( and even more if they are inferred ) is misleading .
the comparison of the densities of objects in each luminosity and redshift bin found by different surveys is the correct way of concluding on the resulting lfs . as it was shown by caputi et al .
( 2004 ; 2005a ) , after the rise of the bright end of the rest - frame @xmath2-band lf from redshifts @xmath5 to @xmath97 , the very bright end ( @xmath90 $ \sim$}}~}-26 $ ] ) of this lf does not show any sign of decline up to at least redshift @xmath98 .
our ml analysis suggests that the density of the @xmath90 $ \sim$}}~}-26 $ ] objects might increase up to @xmath99 and then stay constant up to @xmath98 within the error bars ( cf .
figure [ kslf ] ) . again
, this can be explained by the combined effect of the number and luminosity evolution we obtain through the ml analysis : the total density of objects @xmath91 decreases by a factor @xmath100 between @xmath5 to @xmath98 , but the characteristic magnitude @xmath101 brightens by more than one magnitude between these redshifts .
the combined density and luminosity evolution we determined through the ml analysis also implies a particularly strong decrement in the number of intermediate and low luminosity ( @xmath102 ) objects with increasing redshift .
we remark , however , that we do not directly see these intermediate and low luminosity objects at high redshifts , given the limits of our survey .
we only extrapolate this behaviour from the lf obtained through the ml analysis , which assumes the validity of a schechter function for the description of the lf at all redshifts .
thus , two alternative conclusions can be extracted from this result : 1 ) a significant fraction of low and intermediate luminosity objects are created between @xmath98 and @xmath5 , or 2 ) at least as many low luminosity objects are present at high as at low redshifts , and our assumptions for the computation of the lf with the ml analysis are not valid to describe the faint - end of the lf at high redshifts .
the direct access to low luminosity objects at high redshifts through deeper surveys is necessary to conclusively prove the behaviour of the faint - end of the lf . in hierarchical models ,
the creation of new objects is the consequence of the production of galaxy mergers and subsequent triggering of star formation .
very recently , conselice ( 2005 ) found that the fraction of mergers in faint and low mass galaxies peaked at lower redshifts than for higher luminosity and more massive galaxies .
this result appears to be consistent with the presence of luminous galaxies at high redshifts @xmath103 , as they would be produced from mergers at these or even higher redshifts .
the decline in the number of intermediate luminosity objects at high redshifts that we extrapolate from the ml analysis would also be consistent with the fact that a significant fraction of them form through mergers at later times .
however , the differential redshift peak observed for mergers of different mass and luminosity still can not decide on the formation epoch of all the lowest luminosity objects , and the two hypotheses discussed above could still be possible . at redshifts
@xmath16 $ \sim$}}~}1 $ ] , major mergers become progressively less important ( e.g. patton et al .
2002 ) , so it is expected that the evolution of the @xmath2-band lf is mainly governed by the dimming of the already present old stellar populations .
the significant decrease of the bright end ( @xmath104 ) of the lf from redshifts @xmath97 to @xmath5 is probably indicating that the most massive galaxies present at @xmath105 must have finished the construction of the bulk of their stars by this redshift , and mainly fade out all the way through redshift @xmath5 ( cf .
section [ sec_masscol ] ) .
we used the rest - frame @xmath2-band luminosity of each galaxy to compute its stellar mass .
we obtained the rest - frame absolute magnitude @xmath73 ( and corresponding luminosity ) from the hyperz best - fit sed on an individual basis , as explained in section [ sec_lf ] .
we used the public code galaxev ( bruzual & charlot 2003 ) to construct grids of mass - to - light ( @xmath106 ) ratios depending on galaxy age , assuming a salpeter imf over stellar masses @xmath34 .
the advantage of using luminosities in the rest - fame @xmath2-band is that other parameters , such as the dust - correction or the galaxy star - formation history , have a minor importance in the determination of the @xmath106 ratios . as in caputi
( 2005a ) , we modelled the evolution of these ratios using two galaxy templates , with no dust , corresponding to different exponentially declining star - formation histories with characteristic times @xmath107 and @xmath108 .
we determined the @xmath106 of each galaxy using the model most similar to the hyperz best - fit template in each case , and interpolating between the values which corresponded to the closest ages in the grids .
our estimated stellar masses are typically accurate within a factor @xmath109 $ \sim$}}~}2 $ ] ( where the larger uncertainties correspond to plausible degeneracies in parameter space , especially in those cases in which the galaxy redshift is simultaneously determined with the best - sed fit ) . on the contrary
, the dependence on star - formation history and dust corrections would be much stronger if optical rather than near - ir luminosities were used to derive the stellar masses .
as we discussed in section [ sec_lf ] , the rest - frame @xmath2-band luminosity derived as here and the luminosity that would be derived from the extrapolation of the corresponding observed @xmath2-band magnitude are consistent within a rms @xmath75 .
this implies that the derived stellar masses will be consistent within @xmath110 , i.e. within a factor @xmath111 .
thus , we conclude that the use of the observed b through @xmath2-band data , with no incorporation of irac data , would have introduced an additional uncertainty of @xmath112% on the derived stellar masses .
this is a relatively minor contribution to the error budget , which is mainly dominated by plausible degeneracies in parameter space and especially by the adoption of a fixed imf .
figure [ imkmass ] shows the observed ( @xmath113 ) colours versus redshift for galaxies with stellar masses within different mass ranges . in all the panels ,
circles represent those objects with precise ( @xmath113 ) colours , while triangles indicate lower limits to the colours of objects with magnitude @xmath114 ( vega ; equivalent to @xmath109 $ \sim$}}~}2\sigma$ ] detections in the @xmath20 band ) , computed as ( @xmath115 ) .
the horizontal dashed line shows the typical colour cut defining extremely red galaxies ( ergs ; @xmath116 , vega ) , while the vertical dotted lines delimit the region of completeness for the lowest masses in each mass bin .
the mass completeness limits have been estimated using the median of the k - corrections and the maximal @xmath106 ratios at each redshift ( corresponding to a galaxy formed in a single burst at redshift @xmath117 ) . at redshifts
@xmath89 $ \sim$}}~}1 $ ] , typical erg colours indicate the presence of old stellar populations and/or young stellar populations heavily enshrouded by dust . on the contrary
, @xmath118 colours at @xmath119 imply the presence of unextincted on - going star - formation . at
@xmath16 $ \sim$}}~}1 $ ] , the observed @xmath120 colour cut is merely formal , as only systems with very large dust extinctions would show colours above that limit .
inspection of figure [ imkmass ] shows that the @xmath121 colours of galaxies at a given redshift depend on the stellar mass .
the fraction of red - to - blue galaxies systematically increases with assembled stellar mass , even considering each mass bin only up to the completeness redshift limit . in particular , the vast majority of galaxies with @xmath8 between redshifts @xmath11 and @xmath122 are ergs . some of these galaxies might be purely passive systems ( cf .
e.g. daddi et al .
2005 ) , while in other cases the extremely red colours would only be explained by the superposition of old stellar components and dust ( see caputi et al .
2004 ; papovich et al . 2005 ; and cf .
section [ sec_nearir ] ) .
the extremely red colours characterising these very massive galaxies reveal important constraints on their star - formation histories .
in contrast to what is observed in other less massive systems , our results indicate that unobscured time - extended star - formation at @xmath89 $ \sim$}}~}1 $ ] is rare in galaxies with assembled stellar mass @xmath8 .
the main build - up of these systems must be produced in short bursts and/or with considerable amounts of dust .
the observed @xmath121 colours versus redshift for galaxies with estimated stellar mass @xmath123 .
the symbols are the same as in fig .
[ imkmass ] .
the solid and dotted - dashed lines show the modelled evolution of the @xmath121 colours for galaxies formed in an instantaneous burst at redshift @xmath124 and @xmath125 , respectively , both with no dust and passive evolution thereafter . ]
figure [ imkhighm ] shows an amplified version of the lower - right panel of figure [ imkmass ] . for reference , we added in this figure the modelled evolution of the @xmath121 colours for galaxies formed in an instantaneous burst at redshift @xmath124 and @xmath125 ( solid and dotted - dashed lines , respectively ) , both with no dust and passive evolution thereafter . from figure [ imkhighm ]
we can see that , at @xmath119 , galaxies within the erg regime with assembled stellar masses @xmath123 display a wide range of @xmath121 colours .
this suggests that they must undergo a variety of star - formation histories , either different formation redshifts or at least different formation redshifts for the youngest stellar population present in them . at redshifts
@xmath16 $ \sim$}}~}1 $ ] , on the contrary , the galaxies with stellar mass @xmath123 present in our sample show quite homogeneous @xmath121 colours .
this is again in contrast to what is observed in less massive systems .
the homogeneous colours displayed by these most massive galaxies suggest that their optical / near - ir light must be dominated by old stellar populations .
however , this does not necessarily mean that all of them are following a strictly passive evolution , but that any possible on - going star formation would only produce minor changes to their properties ( e.g. stellar mass ) .
in fact , a substantial fraction ( @xmath126 $ \sim$}}~}30\%$ ] ) of these galaxies are detected in the mid - ir , i.e. experience on - going star - formation ( caputi et al .
2005b , 2005c ) .
the typical ir - derived star - formation rates of these mid - ir detected galaxies at @xmath127 are @xmath128 ( caputi et al .
if we consider that the lifetime of a starburst episode is at most @xmath129 , we infer that these galaxies should be forming @xmath130 per starburst episode at those redshifts , i.e. an additional minor amount to their already existing stellar mass
. some of these very massive systems at @xmath16 $ \sim$}}~}1 $ ] could be candidates for cluster e / s0 galaxies , as early - type cluster galaxies show very little dispersion in their colours from @xmath5 ( e.g. bower et al .
1992 ) to @xmath72 ( e.g. ellis et al . 1997 ; stanford et al . 1998 ; blakeslee et al .
indeed , the existence of large - scale structure is suggested by the colour - redshift diagram , as a considerable number of our @xmath8 galaxies appear to lie at discrete redshift values . on the contrary , only a morphology study would be able to conclude on the plausible early or late - type classification of these systems ( mclure et al .
2005 , in preparation ) . the comoving number density of galaxies with stellar mass @xmath131 ( circles ) and @xmath132 ( triangles ) ( salpeter 1955 imf , restricted to @xmath133 ) .
the filled triangle at @xmath134 denotes the re - computed number density of @xmath132 galaxies after correcting for the effects of large - scale structure in the corresponding redshift bin .
the star - like symbols indicate the density values for the same mass thresholds at @xmath5 , obtained integrating the local stellar mass function ( cole et al .
2001 ; bell et al .
the dashed lines show the evolution of the number density of haloes with total mass @xmath135 and @xmath136 ( upper and lower curves , respectively ) , as obtained from @xmath137-cdm models of structure formation ( with a code kindly provided by will percival ) .
the dotted line corresponds to the predicted number density of galaxies with mass @xmath138 obtained with sph simulations ( nagamine et al .
2005 ) . ]
the comoving number density of galaxies with stellar mass @xmath131 ( chabrier 2003 imf ) .
the symbols are the same as in figure [ cnumd ] . ] as we did in caputi et al .
( 2005a ) , we analyse the evolution of the number density of different - mass galaxies over cosmic time . in the present work , however , due to the analysis of a considerably larger area , we are able to select a statistically significant sample of very massive galaxies ( @xmath132 ) , which enables us to explore the very - high mass end of the stellar - mass evolution
. figure [ cnumd ] shows the comoving number density of galaxies with assembled stellar mass @xmath138 ( circles ) and @xmath132 ( empty triangles ) .
the asterisks denote the corresponding local values , as obtained from the integration of the local stellar mass function ( cole et al .
2001 ; bell et al .
as in the computation of the galaxy lf ( cf .
section [ sec_lf ] ) , the error bars correspond to the maximum of poissonian errors and the errors due to cosmic variance . all the results presented in this section incorporate @xmath139 corrections , but deliberately not sample incompleteness corrections ( which anyway are very small ) .
we decided to do so to explicitly show that the contrast betweeen some of the observational and theoretical results discussed below is independent of such corrections .
the evolution of the densities of galaxies with stellar mass @xmath138 shows that a minimum of 15% of the local value is in place before redshift @xmath12 ( this value rises to @xmath13% considering the number densities cut at stellar masses @xmath140 , i.e. the regime of strict mass completeness to @xmath12 ) .
this results is consistent with caputi et al .
( 2005a ) .
however , in contrast to this previous work , we observe here that the number density of @xmath138 systems does increase from redshifts @xmath141 to @xmath142 , although this increment is modest ( @xmath13% of the local value ) .
more than 50% of the local @xmath138 galaxies appear to have formed at redshifts @xmath16 $ \sim$}}~}1.5 $ ] .
the dashed lines in figure [ cnumd ] trace the evolution of dark - matter haloes with total mass @xmath135 and @xmath136 ( upper and lower curves , respectively ) , as obtained from the extended press - schechter formalism ( bond et al .
1991 ) , with a code kindly provided to us by will percival .
the mass thresholds for the halo distributions have been deliberately selected to coincide with the corresponding number densities of galaxies at @xmath5 , assuming a single halo occupation number .
the compared observed and predicted density evolutions confirm the trend found in caputi et al .
( 2005a ) : at redshifts @xmath143 $ \sim$}}~}z { \raisebox{-0.13cm}{~\shortstack{$<$ \\[-0.07 cm ] $ \sim$}}~}3 $ ] , the density of haloes with total mass @xmath135 is systematically larger than the density of galaxies with stellar mass @xmath138 .
caputi et al .
( 2005a ) also showed that the number density of galaxies and haloes coincide again at higher redshifts @xmath144 , although this can not be re - assessed from the present sample as the @xmath36 cut implies completeness only for stellar masses @xmath132 up to @xmath145 .
on the other hand , we investigate the very high mass end ( @xmath132 ) of galaxy number density evolution and also compare with the corresponding theoretical predictions .
inspection of figure [ cnumd ] shows that up to @xmath13% of the local massive galaxies with assembled stellar mass @xmath132 is in place before @xmath122 , while the construction of the majority of these very massive systems occurs between redshifts @xmath12 and @xmath146 . within our sample
, we find that @xmath147% of the local galaxies with mass @xmath132 is already in place by redshift @xmath148 .
this is contrast to what is observed for less massive systems , whose evolution is only halfway complete by that redshift .
the comparison with theoretical models shows that , if we assume the same conversion between baryonic and dark - matter halo mass as inferred at @xmath5 , there appear to be too few dark matter haloes of adequate mass in place at high redshift to host the observed number of high mass galaxies .
we observe that the deficit is particularly important at @xmath89 $ \sim$}}~}3 $ ] .
similar incompatibilites between the presence of very massive systems and the predictions of theoretical models at high redshifts have already been pointed out in previous studies ( e.g. fontana et al .
2003 ; genzel et al . 2003 ;
pozzetti et al . 2003 ; cimatti et al . 2004 ;
fontana et al .
2004 ; mccarthy et al . 2004
; saracco et al .
our results show that this incompatibility might certainly exist for the most massive galaxies .
it could be argued , however , that our excess of very massive galaxies might be biased by cosmic variance effects ( more than what is accounted for within the error bars ) . in effect , as we noted in section [ sec_masscol ] , the existence of large - scale structure within our sample can be deduced from the colour - redshift diagrams . within our sample ,
we find 16 galaxies with @xmath132 at @xmath149 , yielding a number density @xmath150 .
9/16 of these galaxies have photometric redshifts concentrated in the narrow interval @xmath151 , and there is also a spectroscopically confirmed qso at redshift @xmath152 within our surveyed field ( not included in the final galaxy sample analysed in this work ) , strongly suggesting the presence of large - scale structure for very massive galaxies at these redshifts . to obtain a lower limit on the number density of galaxies with stellar mass @xmath132 at @xmath153
, we re - computed this density excluding the 9 galaxies at @xmath151 .
we obtained @xmath154 , i.e. @xmath155 ( filled triangle in figure [ cnumd ] ) , while the density of the corresponding haloes at that redshift is only @xmath156 .
thus , it seems that even taking into account the effects of large - scale structure , we can not completely reconcile the theory predictions with the observations at high redshifts .
the total stellar mass densities obtained from the present study ( large circles and up - triangle , with error bars , for precise values and lower limit , respectively ) , compared to values previously obtained by other authors : dickinson et al .
2003 ( down - triangles ) ; fontana et al .
2003 ( asterisks ) ; rudnick et al .
2003 ( diamonds ) ; glazebrook et al .
2004 ( right - triangles ) ; fontana et al .
2004 ( squares and double squares for the observed and extrapolated values in the k20 survey , respectively ) ; caputi et al .
2005a ( small circles and double circle for total and @xmath157 mass densities , respectively ) ; drory et al .
2005 ( crosses ) . the circle with a star at @xmath158 show the local value of the total stellar mass density obtained from the integration of the local galaxy stellar mass function ( cole et al . 2001 ,
bell et al . 2003 ) .
the solid line is the evolution of the stellar mass density as obtained from the integration of the star - formation rates derived from sloan digital sky survey ( sdss ) data ( heavens et al .
dickinson et al .
values correspond to the average of the values obtained with the 1-component and 2-component models with solar metallicity ( see table 3 in dickinson et al .
rudnick et al .
values are computed only on galaxies with rest - frame v - band luminosity @xmath159 and , thus , miss a significant fraction of the mass .
glazebrook et al .
values only correspond to galaxies with estimated mass @xmath160 and are at least partially incomplete above redshift @xmath161 .
all the values in this figure assume a salpeter imf over stellar masses @xmath133 and a cosmology with @xmath162 , @xmath32 and @xmath33 . ] recently , nagamine et al .
( 2005 ) showed that the properties of high redshift red and massive galaxies could be well reproduced through the use of hydrodynamical simulations with star - formation rate histories that peak at higher redshifts than those used in semianalytic models .
it is then interesting to compare also their predictions with our observed evolution of number densities of massive galaxies .
the dotted line in figure [ cnumd ] shows the predicted number density of galaxies with stellar mass @xmath138 , as obtained with the nagamine et al . smoothed particle hydrodynamics ( sph ) simulations .
we observe that these simulations yield densities below the observed values , and the discrepancies are larger at low than at high redshifts .
however , in the nagamine et al .
models , galaxies masses are directly obtained from the simulations and are imf independent .
thus , for a proper comparison with these model results , we need to test whether the adoption of a different imf to compute our stellar masses could improve the agreement between model and data .
figure [ chabrier ] shows the evolution of the comoving number density of galaxies with stellar mass @xmath138 considering a chabrier ( 2003 ) imf .
both our observed densities and the local value have been transformed in accordance to this imf . the change from a salpeter ( 1955 ) to a chabrier ( 2003 )
imf produces an overall re - scaling of all our stellar masses by @xmath163 , in such a way that a cut @xmath138 for masses in the chabrier imf corresponds to a cut @xmath164 $ \sim$}}~}2 \times 10^{11 } m_\odot$ ] in the salpeter imf .
we repeat the same exercise as in figure [ cnumd ] and compare the observed values with the densities of dark matter haloes with a mass threshold deliberately selected to coincide with the density of galaxies at @xmath5 . as in figure
[ cnumd ] , we observe that , at very high redshifts @xmath89 $ \sim$}}~}3 $ ] , the predicted number of haloes appears to be insufficient to encompass the observed densities of @xmath138 ( chabrier ) galaxies . on the contrary , the sph predicted number densities of @xmath138 galaxies by nagamine et al .
( 2005 ) appear to be in quite good agreement with our observed values from redshifts @xmath72 through @xmath165 , although they produce some excess of massive local galaxies .
we note that a similar behaviour would be deduced from the densities of dark matter haloes if we did not constrain the comparison to those haloes with densities coinciding with the local densities of massive galaxies .
this shows that what we find here is a direct discrepancy with the predictions of hierarchical models , and does not depend on the specific recipes adopted either by semianalytic models or sph simulations . in summary ,
hierarchical models do not seem to be able to reproduce the observed densities of the most massive galaxies from the local universe to very high redshifts .
alternatively , it could also be argued that the use of a single imf is not suitable to convert from luminosities to stellar masses in the whole range of redshifts @xmath5 to @xmath122 ( see e.g. kroupa & weidner 2005 , for a discussion ) .
unluckily , the validity of a single imf is an issue extremely hard to test at high redshifts , and it is not possible to overcome the imf uncertainty when doing the comparisons between observations and theory . finally , we computed the evolution of the total stellar mass density with redshift , as derived from our sample of @xmath0 galaxies .
figure [ compstm ] shows our obtained stellar mass densities ( large circles and up - triangle , with error bars ) , compared to other values previously published in the literature . from the present sample we find that the total stellar mass densities at @xmath166 and @xmath167 are @xmath168 and @xmath169 , respectively , i.e. @xmath170 and @xmath171 of the local value .
these values very well agree with the previous determinations made by caputi et al .
( 2005a ) and other works . at higher redshifts ,
the completeness limits of our sample only allows us to obtain lower limits on the total stellar mass densities . the up - triangle in figure [ compstm ] represents a lower limit to the total stellar mass density at redshift @xmath98 .
as our galaxy sample is only complete for masses @xmath172 up to @xmath141 , we compare this lower limit to the stellar mass density contained in galaxies with stellar mass @xmath173 obtained by caputi et al .
( 2005a ) ( double circle in figure [ compstm ] ) , and observe a very good agreement .
the total stellar mass densities at redshifts @xmath89 $ \sim$}}~}2 $ ] have been better constrained by previous deeper surveys .
the estimated stellar mass densities at @xmath103 and @xmath144 are ( 20 - 24)% and ( 13 - 18)% of the local values , respectively ( caputi et al . 2005a ; drory et al .
the observed versus predicted @xmath174)$ ] and @xmath175)$ ] colours ( upper and lower panels , respectively ) for the @xmath0 ( vega ) galaxies in the restricted area of the goods / cdfs studied by caputi et al .
( 2005a ) .
the dashed lines indicate the region of consistency , taking into account the error bars on the observed colours . ]
caputi et al . ( 2005a ) predicted the pure near - ir @xmath2/irac colours of @xmath2-selected galaxies based on the best - fit sed of each object .
these predictions were made for the galaxies selected in a sub - region of 50.4 arcmin@xmath35 of the area analysed in this work . in the light of the currently available irac data , we can assess the validity of these predictions .
figure [ colobspr ] shows the observed versus predicted @xmath174)$ ] and @xmath175)$ ] colours for the @xmath2-band galaxies in the restricted goods / cdfs area ( upper and lower panels , respectively ) . for clarity
, we only show those sources with small error bars in their observed irac magnitudes , i.e. @xmath176 and @xmath177 . as in all cases
the @xmath2 magnitudes have @xmath178 $ \sim$}}~}0.2 $ ] , the measured colours are typically precise within @xmath179 0.4 mag ( dashed lines in both panels of figure [ colobspr ] ) . from this figure
we determine that 89% and 84% of the predicted @xmath174)$ ] and @xmath175)$ ] colours , respectively , are in agreement with the observed colours , within the error bars .
we can confirm , then , the majority of the near - ir colours predicted in caputi et al .
( 2005a ) . the observed @xmath174)$ ] and @xmath175)$ ] colours ( upper and lower panels , respectively ) versus redshift for the @xmath0 ( vega ) galaxies in the goods / cdfs .
] figure [ colnearir ] shows the observed @xmath174)$ ] and @xmath175)$ ] colours versus redshift ( upper and lower panels , respectively ) for the @xmath0 ( vega ) galaxies in the larger sample studied in this work . @xmath44 and @xmath4 data are available for alternative @xmath47 of the sources in our sample .
once more , for clarity , we only show those sources with magnitude errors @xmath176 and @xmath177 . the variation of these observed colours with redshift confirms the trends predicted by caputi et al .
( 2005a ) within the error bars .
inspection of figure [ colnearir ] reveals , however , the existence of a population of low - redshift ( @xmath16 $ \sim$}}~}0.5 $ ] ) galaxies with colours @xmath174)>1.0 $ ] and @xmath175)>1.5 $ ] , which are quite redder than those predicted for low - redshift galaxies in the previous work . the colour excess of these galaxies can not be reconciled with the predictions by simply taking into account the photometry error bars .
the analysis of these sources shows that 31 out of the 32 objects satisfying @xmath16 $ \sim$}}~}0.5 $ ] and @xmath174)>1.0 $ ] and all of the 17 objects satisfying @xmath16 $ \sim$}}~}0.5 $ ] and @xmath175)>1.5 $ ] are very low mass systems , with assembled stellar masses @xmath180 .
the majority of them are best - fit by young - galaxy templates .
the excess of near - ir colour could be explained by the presence of an active nucleus , whose irac magnitudes could not be properly predicted using the gissel98 stellar templates . on the other hand ,
an alternative and perhaps more likely explanation is that these very red @xmath2/irac colours are produced by the presence of prominent emission lines in the irac bands .
the presence of an aromatic feature at @xmath181 is characteristic of star - forming galaxies .
these aromatic features are not incorporated in the gissel98 templates , explaining why such very red colours could not have been predicted from the sed models .
if this is the case , the colour excess would be the signature of intense star - formation in these young objects
. only near - ir spectrocopy would be able to confirm the nature of these sources .
figure [ ksm45comp ] shows the observed @xmath175)$ ] colours versus redshift in different stellar mass bins , analogously to figure [ imkmass ] .
we observe that the @xmath175)$ ] colours increase up to redshift @xmath182 , and the colour - redshift relation becomes tighter with increasing stellar mass .
in contrast , the colour dispersion becomes quite large above that redshift , where the observed near - ir bands start to map the rest - frame optical bands .
this colour dispersion is in part due to larger photometry uncertainties , but it is probably also an indication of a wide variety of star - formation histories and dust extinction in these galaxies .
figure [ ksm45model ] shows an amplified version of the high - mass panel in figure [ ksm45comp ] .
we added for comparison the modelled evolution of the observed @xmath175)$ ] colours for galaxies formed in a single burst at redshifts @xmath125 and @xmath124 , both with no dust and passive evolution thereafter . some interesting conclusions can be extracted from the analysis of this figure .
firstly , it is quite evident that the observed near - ir colours of many massive galaxies at high redshifts ( @xmath89 $ \sim$}}~}1.5 $ ] ) can not be explained without considerable amounts of dust ( cf . also caputi et al .
2004 ) . independently of the formation redshift ,
the colours of a passively - evolving single burst are not sufficient to reproduce the red observer - frame near - ir colours of massive galaxies .
we conclude , then , that purely passive galaxies should be a minority at high redshifts .
secondly , the colours observed at lower redshifts ( @xmath16 $ \sim$}}~}1.5 $ ] ) indicate that the near - ir light of @xmath8 galaxies must be dominated by evolved stellar populations . a similar conclusion has been obtained from the analysis of the @xmath121 colours of these systems .
we note that the tight colour - redshift relation observed for high - mass galaxies at @xmath16 $ \sim$}}~}1.5 - 2.0 $ ] does not imply a common formation redshift or star - formation history .
for instance , as we mentioned in section [ sec_masscol ] , some of these galaxies are bright mid - ir sources , while others are not , suggesting a wide variety of on - going star - formation rates ( caputi et al .
2005b , c ) . however , and in contrast to what is progressively observed in less massive systems , all the plausible new star - formation in very massive galaxies at low redshifts could have a minor impact on their overall properties and , in particular , on their sed colours .
the evolution of the observed @xmath183)$ ] colours with redshift for galaxies with estimated stellar mass @xmath8 .
the solid and dotted - dashed lines show the modelled evolution of the @xmath183)$ ] colours for galaxies formed in an instantaneous burst at redshift @xmath124 and @xmath125 , respectively , both with no dust and passive evolution thereafter . ]
with the goods project close to completion , in this work we have been able to set tighter constraints on the evolution of near - ir - selected galaxies in the goods / cdfs field .
we took advantage of the latest available very deep and homogeneous multiwavelength photometry , in conjunction with the information provided by complementary spectroscopic programs and the combo17 campaign , to construct an optimised redshift catalogue for the @xmath0 galaxies in the goods - south field .
based on this catalogue , we performed a detailed analysis of galaxy luminosity and mass evolution up to very high redshifts @xmath184 .
the incorporation of the spitzer / irac data at @xmath185 and @xmath186 allowed us to compute rest - frame @xmath2-band luminosities and stellar masses with considerably smaller extrapolations than those applied in caputi et al .
( 2005a ) . in spite of this important difference
, we note that most of the main conclusions obtained by caputi et al .
( 2005a ) have been confirmed by this work .
the addition of irac data did not produce any overall change in the derived galaxy luminosities and stellar masses .
other authors also have concluded that the use of irac data produced a minor impact on the derived galaxy stellar masses , when other shorter optical / near - ir multiwavelength data are also available ( e.g. shapley et al . 2005 ) . from the modelling of the evolution of the @xmath2-band luminosities
, we determined that the characteristic magnitude @xmath93 of the rest - frame @xmath2-band lf significantly brightens from redshifts @xmath5 to @xmath98 .
this is simultaneously produced with a decrease in the total density of objects , so we conclude that the @xmath2-band lf has both a substantial luminosity and density evolution with redshift .
the characteristics of this evolution reveal some important aspects of the sequence of galaxy formation .
for instance , that most of the brightest galaxies are assembled at high redshifts ( @xmath89 $ \sim$}}~}1 $ ] ) , while the creation of a substantial number of intermediate - to - low luminosity objects might continue down to low redshifts ( although the latter will need to be confirmed by future deeper surveys , which have direct access to low - luminosity objects at high redshifts ) . we have also studied the very high stellar - mass end ( @xmath132 ) of galaxy evolution .
we found that most of these very massive systems in our sample have been formed in the @xmath187-period of time between redshifts @xmath141 and @xmath148 , although a significant fraction ( up to @xmath13% ) appear to have been in place before redshift @xmath122 .
very recently , from the cross - correlation of our @xmath0 galaxy sample with the spitzer/@xmath51 galaxies in the same field , caputi et al .
( 2005b ) found that the most - intense star - forming galaxies at redshifts @xmath188 are mainly massive ( @xmath189 ) galaxies .
this is also the main epoch of flourish of submillimetre galaxies ( chapman et al .
2005 ) , which presumably are progenitors of the most massive local galaxies ( e.g. stevens et al .
thus , this period appears to be critical to understand the formation and evolution of the bulk of the most massive systems present in the local universe .
theoretical models of galaxy formation in their current status are not completely able to account for all of these results .
several different clues have been proposed in recent works to try to improve the agreement of the models with the data at high redshifts , including e.g. the effects of qso and supernova feedback ( granato et al .
2004 ; cirasuolo et al .
2005 ) , the adoption of an alternative imf ( baugh et al .
2005 ) or the use of hydrodynamical models ( nagamine et al .
however , the discrepancies are not produced by the specific recipes used by theoretical models to account for star formation , but seem to lie at the more fundamental level of hierarchical model predictions .
hierarchical models do not seem to be able to reproduce the whole of observational properties from the local universe to high redshifts .
a likely alternative to this problem is the invalidity of a universal imf when deriving stellar masses from the observed luminosities .
a definitive comprehension of the main epoch of very massive galaxy formation , as well as the determination of the initial distribution of stars in these systems , still constitutes a major challenge from both the observational and theoretical points of view .
this paper is based on observations made with the advanced camera for surveys on board the hubble space telescope operated by nasa / esa ; with the infrared spectrometer and array camera on the ` antu ' very large telescope operated by the european southern observatory in cerro paranal , chile ; and the infrared array camera on board the spitzer space telescope , operated by the jet propulsion laboratory , california institute of technology , under a contract with nasa ; and form part of the publicly available goods datasets .
we thank the goods teams for providing reduced data products .
we are very grateful to kentaro nagamine for kindly providing us with his sph simulation predictions , and to will percival for providing us the code to compute the number densities of dark matter haloes .
we thank the anonymous referee for his / her useful comments and suggestions , which improved the discussion of results in this work .
kic acknowledges cnes and cnrs funding .
rjm acknowledges the support of the royal society .
alonso - herrero a. et al . , 2005 , apj , submitted
|
we have selected and analysed the properties of a sample of 2905 @xmath0 galaxies in @xmath1 of the great observatories origins deep survey ( goods ) chandra deep field south ( cdfs ) , to obtain further constraints on the evolution of @xmath2-selected galaxies with respect to the results already obtained in previous studies .
we made use of the public deep multiwavelength imaging from the optical @xmath3 through the infrared ( ir ) @xmath4 bands , in conjunction with available spectroscopic and combo17 data in the cdfs , to construct an optimised redshift catalogue for our galaxy sample .
we computed the @xmath2-band lf and determined that its characteristic magnitude has a substantial brightening and a decreasing total density from @xmath5 to @xmath6 .
we also analysed the colours and number density evolution of galaxies with different stellar masses . within our sample , and in contrast to what is observed for less massive systems , the vast majority ( @xmath7% ) of the most massive ( @xmath8 ) local galaxies appear to be in place before redshift @xmath9 . around @xmath10% of the total
assemble between redshifts @xmath11 and @xmath12 and most of them display extremely red colours , suggesting that plausible star formation in these very massive systems should mainly proceed in obscured , short - timescale bursts . the remaining fraction ( up to @xmath13% )
could be in place at even higher redshifts @xmath14 , pushing the first epoch of formation of massive galaxies beyond the limits of current near - ir surveys .
[ firstpage ] galaxies : evolution galaxies : formation galaxies : high - redshift galaxies : luminosity function , mass function
|
the opto - mechanical oscillator ( omo ) generates spectrally pure radio frequency ( rf ) signals @xcite due to ponderomotive interaction between photons and phonons .
phase noise and linewidth are the main characteristics detremining the performance of an oscillator and its practical usefulness .
it was shown that the omo linewidth can be small enough to be ultimately described by a schawlow - townes - like formula @xcite . a leeson model @xcite of the phase noise far from the carrierfor a radiation pressure driven omo was presented in @xcite . in this paper
we derive a generalized formula for the omo phase noise that takes into account the noise of the light pumping the resonator . using the formula ,
we analyze allan deviation of the oscillator frequency as well as its linewidth . the predicted performance of the omo
is then compared with the performance of existing electronic oscillators .
opto - mechanical oscillation can be described as a strongly nondegenerate parametric process in which a pump photon is transformed to a photon of lower frequency ( stokes photon ) and a phonon @xcite .
this occurs when the pump power exceeds a certain threshold determined by the loss of the system , and the coupling efficiency between the light and the mechanical modes . in the case of microcavity - based omo
, the process is allowed if the resonator has at least two optical modes with frequency equal to the frequency of a mechanical mode .
this condition satisfies energy conservation .
the momentum conservation ( phase matching of the parametric interaction ) is met by the requirement of spatial overlap between the optical and the mechanical modes that ensure the convolution integral of the electric fields and the mechanical displacement , as well as mechanical strain , is nonzero @xcite .
generally , the radiation by a laser is assumed to be phase insensitive .
it means that the phase of the emitted light in a laser does not depend on the phase of the pump , and that there is no phase matching conditions ( preferred directionality ) in the system .
an omo , even one based on brillouin lasing @xcite , has different properties .
the mechanical quality factor is high enough to endow the generated phonons with some well defined phase .
the phase of the stokes light emitted by the omo depends on the phonon phase , so the amplification of the stokes wave becomes phase sensitive and the phase fluctuations of the optical pump leak to the phase fluctuations of the stokes light . in what follows
we develop a model for an omo and find the phase noise of the generated signal .
it was shown that the linewidth of an omo increases if generation of an anti - stokes optical sideband is allowed by the system architecture @xcite . in this paper
we analyze an ideal case where only two high quality ( q- ) factor optical modes and a single mechanical mode interact .
the optical pump is resonant with higher frequency optical mode and the stokes optical sideband is generated in the lower frequency mode .
the bandwidth of the optical modes is assumed to be much smaller than the mechanical frequency .
this kind of interaction has been realized in oscillators based on stimulated brillouin scattering @xcite and surface acoustic waves @xcite .
we write three coupled langevin equations for the modes and solve them analytically to evaluate the phase noise of the omo . to obtain a consistent result ,
we take into account the phase noise of the laser used for pumping the omo .
we show that one limitation of an omo is that the frequency stability of the mechanical mode is determined in the same way as the electronic quartz oscillators .
the fluctuations of the thermal bath as well as the thermodynamic noise and drift of the mechanical mode limit the frequency stability .
the pumping light is used as the power source in an omo , similar to the electric power source in electronic oscillators and , therefore , it is unlikely that the omo will outperform its electronic analogs , unless the frequency bandwidth of the optical modes is significantly smaller than the frequency bandwidth of the mechanical mode .
an advantage of an omo , on the other hand , is in its potential long term stabilization via stability transfer from the optical frequency domain .
the paper is organized as follows .
the equations are presented in section ii , and their solutions are described in section iii .
the omo phase noise and linewidth are analyzed and compared to similar parameters of existing quartz oscillators in section iv .
the triply - resonant opto - mechanical interaction is described by equations @xmath0 where @xmath1 , @xmath2 , and @xmath3 are the slowly - varying amplitudes the pump ( optical ) , the stokes ( optical mode red shifted with respect to the pump ) , and the signal ( mechanical ) fields ; @xmath4 , @xmath5 , and @xmath6 are the linear resonant terms of optical and mechanical modes respectively @xmath7 @xmath8 and @xmath9 are the intrinsic decay rates of the optical and mechanical modes , @xmath10 and @xmath11 are optical loading ( coupling ) rates ( the loading of the optical modes can be different because modes belong to different families ) ; @xmath12 is the opto - mechanical coupling constant , @xmath13 @xmath14 is the correction coefficient showing that radiation pressure results not only in a change in the size of the resonator , but also in its index of refraction through strain , @xmath15 is an effective mass of the mechanical mode , @xmath16 is an effective spatial parameter of the mode . the terms @xmath17 and @xmath18 represent langevin forces with two uncorrelated parts arising from the internal and coupling loss of the modes @xmath19 @xmath20 where @xmath21 stands for ensemble averaging , and @xmath22 is the power of the external optical pump of the mode @xmath23 . the langevin force describing the thermal fluctuations of the mechanical system
is defined in the similar way @xmath24 where @xmath25^{-1}$ ] is the averaged number of thermal phonons leaking from the thermal bath of temperature @xmath26 to the mechanical mode , @xmath27 is the boltzmann constant .
we present the slowly - varying complex amplitudes as @xmath28 and derive two sets of equations for the amplitude and phase parts of the complex amplitudes from eqs .
( [ q2]-[q3 ] ) @xmath29 and @xmath30 where @xmath31 equations ( [ am1]-[ph3 ] ) completely describe the triply - resonant opto - mechanical interaction .
using the set of equations for the phase ( [ ph1]-[ph3 ] ) we get the equation for phase difference @xmath32 @xmath33 we introduce @xmath34 and rewrite eq .
( [ phi1 ] ) as @xmath35 next , we introduce an expectation time - independent value for the phase difference , @xmath36 , and fluctuational time - dependent part , @xmath37 , so that @xmath38 , and get @xmath39 + \frac{f_{ai}}{|a| } - \frac{f_{bi}}{|b|}-\frac{f_{ci}}{|c| } , \label{phi}\end{aligned}\ ] ] where @xmath40 $ ] means a deviation from the expectation value . using eqs .
( [ am1]-[ph3 ] ) we find several relationships for the expectation values of the oscillator parameters in the steady state @xmath41 the expectation value of the amplitude of the field in the pumped mode increases below the oscillation threshold with increase of the pump power , in accordance with @xmath42 , and then stays constant , in accordance with eq .
( [ s2 ] ) . for the sake of simplicity ,
we assume that the stokes sideband has much lower power compared with the pump , @xmath43 , and that the system is triply - resonant ( the expectation values of the frequencies are resonant with corresponding modes ) .
we note that @xmath44 so that for the case of relatively weak stokes sideband and all - resonant tuning @xmath45 using the assumptions and ( [ simplificat ] ) we find @xmath46 \approx ( \delta \phi_{f_a}-\delta \phi_a ) \frac{|\langle f_a \rangle|}{|a|},\end{aligned}\ ] ] where we took into account that , in accordance with eq .
( [ psi ] ) and eq .
( [ psiav ] ) , @xmath47 for the resonant tuning .
we obtain a set of linear equations consisting of eqs .
( [ ph1 ] ) and ( [ phi ] ) @xmath48 the condition of small stokes sideband can be written in the form @xmath49 substituting eq .
( [ force1 ] ) into ( [ dotphi ] ) and ( [ dotphia ] ) we derive @xmath50 the third equation for phase deviations can be derived from eqs .
( [ ph2 ] ) and ( [ ph3 ] ) : @xmath51 we solve the set of linear differential equations ( [ delphia1]-[delphia3 ] ) using fourier transform , e.g. @xmath52 where @xmath53 and @xmath54 are the fourier components of the noise , @xmath55 readout of the omo signal is accomplished by tracking thradio frequency beat note produced by the optical pump and the generated sideband on a fast photodiode , so the phase and frequency of the measured oscillation signal are determined by the argument of the product of optical amplitudes ( @xmath56 ) . the phase noise of the signal is given by difference @xmath57 . on the other hand
, the mechanical frequency can be read using electronics means , e.g. a capacitive displacement sensor .
neglecting by the electronics back action , we can estimate the phase noise of the signal evaluating @xmath58 .
we find expressions for fourier amplitudes of these parameters @xmath59 equations ( [ deltadelta ] ) and ( [ deltac ] ) can be used to find the phase noise of the omo . for example , single - sideband phase noise @xmath60 of the mechanical oscillation is defined as , @xmath61 using definition @xmath62 we find @xmath63 } \frac{\gamma+\gamma_{cb}}{2|b|^2}+ \\ \nonumber & & \frac{\omega^2 + ( \gamma+\gamma_{cb})^2}{\omega^2[\omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 ] } \frac{\gamma_{m}}{|c|^2}\left ( \bar { n}_{th}+\frac 1 2 \right ) , \end{aligned}\ ] ] where @xmath64 stands for fourier frequency dependent input phase noise of the pump laser .
this value is usually much larger than the contribution from the quantum white noise of the laser . using similar reasoning
we obtain @xmath65 } \frac{\gamma+\gamma_{cb}}{2|b|^2}+ \\ \nonumber & & \frac{(\gamma+\gamma_{cb})^2}{\omega^2[\omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 ] } \frac{\gamma_{m}}{|c|^2}\left ( \bar { n}_{th}+\frac 1 2 \right ) .\end{aligned}\ ] ] it is also useful to write an expression for the phase noise of the optical stokes mode , determining the stability and spectral purity of the brillouin laser @xcite @xmath66 } \frac{\gamma+\gamma_{cb}}{2|b|^2}+ \\ \nonumber & & \frac{(\gamma+\gamma_{cb})^2}{\omega^2[\omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 ] } \frac{\gamma_{m}}{|c|^2}\left ( \bar { n}_{th}+\frac 1 2 \right ) .\end{aligned}\ ] ] formulas ( [ lc ] ) , ( [ lab ] ) and ( [ lab1 ] ) can be simplified further using the ratio between the photon number in the stokes mode and the phonon number in the mechanical mode ( [ s1 ] ) , and the expression connecting the number of pump and stokes photons and output power of the pump and stokes light @xmath67 } \frac{\gamma_{cb}(\gamma+\gamma_{cb})}{\omega^2 } \frac{\hbar \omega_0}{p_{bout } } , \\ & & \label{pn2 } { \cal l}_{c}= \\ \nonumber & & \frac{\gamma_m^2}{\omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 } \frac{\gamma_{ca}^2}{\omega^2 + ( \gamma+\gamma_{ca})^2}{\cal l}_{in}+ \\ \nonumber & & \frac{\gamma_m^2\left \ { \bar
{ n}_{th}+1 + \omega^2/[2(\gamma+\gamma_{cb})^2 ] \right \}}{[\omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 ] } \frac{2 \gamma_{cb}(\gamma+\gamma_{cb})}{\omega^2 } \frac{\hbar \omega_0}{p_{bout } } , \\ & & \label{pn3 } { \cal l}_{b}= \\ \nonumber & & \frac{(\gamma+\gamma_{cb})^2}{\omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 } \frac{\gamma_{ca}^2}{\omega^2 + ( \gamma+\gamma_{ca})^2}{\cal l}_{in } + \\ \nonumber & & \frac{\omega^2 + 2\gamma_m^2(\bar { n}_{th}+1 ) } { [ \omega^2 + ( \gamma_m+\gamma+\gamma_{cb})^2 ] } \frac{\gamma_{cb}(\gamma+\gamma_{cb})}{\omega^2 } \frac{\hbar \omega_0}{p_{bout}}. \\\end{aligned}\ ] ] equation ( [ pn1 ] ) gives a complete description of the phase noise characteristic of the radio frequency photonic oscillator based on demodulation of the light output of the omo on a fast photodiode .
equation ( [ pn2 ] ) shows the limiting phase noise of the oscillating mechanical mode that could observed , if the mechanical oscillation is measured using an external devise that does not disturb the system .
equation ( [ pn3 ] ) describes the phase noise of the sbs laser .
these expressions can be used to find the linewidth and allan deviation of the corresponding omo signals . by definition ,
the allan variance of the frequency of the oscillator is given by @xmath68 let us consider the case where the omo signal is retrieved by demodulation of the light leaving the resonator on a photodiode .
the single sided power density of the phase noise is related to the linewidth of the oscillator @xmath69 as @xmath70 .
we find for the cases of relatively low and relatively high - q of the mechanical mode @xmath71 therefore , in the case of low - q mechanical mode , the linewidth of the radio frequency beat note generated by the omo is determined by the linewidth of the pumping light ; while in the case of high - q mechanical mode , the linewidth is described by schawlow - townes - like formula @xcite and the phase noise of the pumping light is suppressed .
let us estimate the phase noise of the oscillator and compare it to the phase noise of an electronic oscillator .
we assume that @xmath72 , @xmath73 rad@xmath74s , @xmath75 rad@xmath74s , @xmath76 rad@xmath74s , @xmath77 rad@xmath74s , @xmath78 rad@xmath74s , @xmath79 khz , @xmath80w , and @xmath81 mw .
the light escaping the cavity is demodulated on a photodiode to produce the radio frequency signal .
the photodiode introduces thermal noise and white shot noise @xmath82 in addition to the phase noise ( [ pn1 ] ) coming from the opto - mechanical oscillator . here
@xmath83 is the resistance of the photodiode , @xmath84 is the responsivity of the photodiode , @xmath85 is quantum efficiency of the photodiode , @xmath86 is the electron charge , and @xmath87 is the total optical power reaching the photodiode .
we assume that @xmath88 .
the thermal noise depends on the ambient temperature @xmath26 and the power of the radio frequency signal leaving the photodiode @xmath89 . to find its value we assume that the resistance at the output of the photodiode is @xmath90 ohm , the responsivity of the photodiode is @xmath91 a@xmath74w , and the temperature is @xmath92 k. for these parameters , the expected power of the radio frequency signal escaping the photodiode is @xmath93w . adding @xmath94 and @xmath95 we plot the spectrum of the single sideband phase noise ( line ( 1 ) in fig .
( [ fig1 ] ) ) .
it is dominated by the phase noise of the pump laser .
if a very narrow linewidth laser is used instead , the phase noise of the omo would be given by the convolution of curves ( 4 ) and ( 3 ) ( phase diffusion of the oscillator as well as shot noise ) .
in the same figure we also show the phase noise of a commercially available 100 mhz oven controlled quartz oscillator ( line ( 5 ) in fig .
( [ fig1 ] ) ) . apparently , the electronic oscillator has much lower phase noise than the omo , even though thermal drifts resulting in the flicker noise are not taken into account in our analysis . )
. curve ( 4 ) shows phase diffusion of the opto - mechanical oscillator .
curve ( 5 ) stands for a commercially available 100 mhz ovenized quartz oscillator.,width=321 ] let us evaluate the sources of frequency drift of the opto - mechanical oscillator .
we note that , according to ( [ omm ] ) and energy conservation law , @xmath96 , @xmath97 the omo frequency drifts if the frequencies of the optical , @xmath98 , and mechanical , @xmath99 , modes drift .
it is possible to lock the frequency of the pump light to the frequency of the corresponding optical mode , @xmath100 .
then the omo frequency becomes dependent on the parameters of the microcavity only : @xmath101 as well as @xmath99 .
this is similar to the case of a quartz oscillator , except the omo frequency depends on both the eigenfrequency of the mechanical and optical modes .
the dependence of the oscillation on the drift of the mechanical mode can be suppressed if the bandwidth of the optical mode is much smaller than the bandwidth of the mechanical mode .
a clear advantage of the omo over an electronic oscillator is that it can be readily stabilized by locking the pump frequency to the frequency of the corresponding cavity mode , @xmath100 , and then locking the temperature of the cavity to a thermally insensitive optical reference line , for example , an atomic transition .
such a stabilization of the temperature will stabilize the long term drift of the oscillation @xcite .
we have theoretically studied the phase noise of a triply - resonant opto - mechanical oscillator based on a nonlinear optical microcavity .
the oscillator generates stokes optical photons and mechanical phonons out of photons of a coherent pumping light .
we have shown that the spectral purity of the opto - mechanical signal is primarily limited by the phase noise of the pump laser .
the overall short term performance of the oscillator is expected to be worse than the performance of a conventional electronic oscillator .
an important advantage of the opto - mechanical oscillator is the possibility to optically stabilize it . with a proper electronic locking scheme it is possible to transfer frequency stability from the optical domain to radio frequency domain .
if stabilized to an atomic transition , an opto - mechanical oscillator can be made to have better long term stability compared to an oven controlled quartz oscillator .
|
we theoretically study the frequency stability of an opto - mechanical radio frequency oscillator based on resonant interaction of two optical and one mechanical modes of the same optical microcavity .
a generalized expression for the phase noise of the oscillator is derived using langevin formalism and compared to the phase noise of existing electronic oscillators .
|
the lawrence krammer representation @xmath0 , n(n-1)/2)$ ] was first introduced by lawrence @xcite and proved to be faithful by bigelow @xcite and krammer @xcite . as
a braid invariant @xmath1 is strong enough to distinguish all braids .
the characteristic polynomial @xmath2 , where @xmath3 denotes the @xmath4 dimensional identity matrix , of the lawrence
krammer matrix @xmath5 of a braid @xmath6 appears to be rather good as a conjugacy invariant of braids .
the author observed that @xmath7 does not detect the orientation reversal of strings of braids , hence is not a complete conjugacy invariant , and that the polynomial @xmath8 has a symmetry @xmath9 just like the alexander polynomial of links ( see corollary [ coro : incomplete ] and [ coro : rev ] ) .
actually this paper arised in search of an explanation for these phenomena .
let @xmath10 be an oriented disk with @xmath11 holes in the complex plane .
the boundary @xmath12 consists of @xmath11 puncture boundary components and the outer boundary component .
let @xmath13 denote the space @xmath14 of all unordered pairs of points in @xmath10 .
let @xmath15 be an open regular neighborhood of the subset @xmath16 .
let @xmath17 .
then @xmath18 is a 4-manifold with boundary .
the fundamental group @xmath19 is isomorphic to the subgroup @xmath20 of @xmath21 consisting of @xmath22-braids whose first @xmath11 strands go straight down without winding and only the last two strands freely wind around .
let @xmath23 be the homomorphism to the abelian group of monomials defined by @xmath24 where @xmath25 is the exponent sum of @xmath26 in artin generators @xmath27 and @xmath28 is the linking number of the last two strands with the first @xmath11 strands .
note that @xmath29 for any @xmath11-braid @xmath30 .
let @xmath31 be the covering space of @xmath18 whose fundamental group is the kernel of @xmath32 .
let @xmath33 denote the ring @xmath34 $ ] of two variable laurent polynomials .
the homology group @xmath35 admits a @xmath33module structure .
it is a free @xmath33module with rank @xmath4 ( see ( * ? ? ?
* section 4 ) ) .
the lawrence
krammer representation refers to the braid action on @xmath35 as @xmath33module automorphisms . by the the blanchfield duality theorem @xcite ( * ? ? ?
* appendix e ) , we have a non - degenerate sequilinear pairing @xmath36 given by @xmath37 where @xmath38 denotes the ordinary intersection number and @xmath39 denotes the torsion free part of a @xmath33module @xmath40 . by identifying @xmath41 with @xmath42
, we obtain a pairing @xmath43 .
since the braid action on @xmath35 is induced from self - homeomorphisms of @xmath31 , clearly it should preserve the pairing @xmath44 , i.e. , the lawrence
krammer representation is unitary .
this could be a non - constructive and terse proof of the following theorem .
[ thm : uni ] there exists a non - singular @xmath45 matrix @xmath46 over @xmath47 $ ] such that for each lawrence krammer matrix @xmath48 of an arbitrary @xmath11-braid @xmath6 and its conjugate transpose @xmath49 , the equality @xmath50 holds .
the symmetry property of the characteristic polynomial @xmath8 naturally follows . in order to find the explicit matrix @xmath46 representing the pairing , we need the concrete description of a base set of @xmath35 and its dual base in @xmath51 . in
@xcite the _ forks _ were used to express relative 2-cycles in @xmath52 .
bigelow showed that given each relative 2-cycle @xmath53 defined by a fork , @xmath54 is an image of a 2-cycle from a closed surface immersed in @xmath31 . using the set @xmath55 of relative 2-cycles corresponding to _ standard forks _ as a base
, he presented the explicit matrices for the lawrence
krammer representation .
let @xmath56 be the 2-cycle in @xmath35 which maps to @xmath57 .
we can define a @xmath45 matrix @xmath58 whose entries are given by @xmath59 for @xmath60 and @xmath61 .
then for the lawrence krammer matrix @xmath62 of an @xmath11-braid @xmath63 the equality @xmath64 follows from @xmath65 .
the author did not succeed in obtaining a concrete description of @xmath56 as a surface or in calculating the the matrix @xmath58 .
the proof of theorem [ thm : uni ] in this paper involves cumbersome matrix calculations on _ biforks _ defined in section [ sec : bifork ] .
the matrix @xmath46 of theorem [ thm : uni ] is presented in lemma [ lem : multab ] as a multiplication table of an algebra .
our proof has an advantage over the non - constructive one in that the same method also works for the proof of the following theorem , which relates the orientation reversal of a braid to the matrix transposition . as a corollory of this theorem we have @xmath66 , where @xmath67 is the anti - automorphism given by @xmath68 .
[ thm : rev ] there exists a non - singular @xmath45 matrix @xmath69 over @xmath47 $ ] such that for each @xmath11-braid @xmath6 the equality @xmath70 holds . in
@xcite it was shown that the lawrence
krammer representation is isomorphic to a summand of the representation on the birman
wenzl ( bmw ) algebra @xcite .
this implies that the artin generators @xmath27 satisfy certain cubic relations in the representation .
[ prop : cubic ] the lawrence
krammer representation satisfies the cubic relations @xmath71 for @xmath72 .
the eigenvalue @xmath73 is associated to a 1dimensional eigenspace . therefore @xmath74 maps to a projector on to the 1dimensional subspace .
this fact led the author to define the following skein relation of _ geometric biforks_. @xmath75 although all the proofs in this paper were solely done by matrix calculations independent of the geometric biforks , the definition of biforks and the main idea of the proofs came from the geometric biforks . in @xcite it was shown that as a braid invariant the burau representation @xmath76 ) $ ] is dominated by the finite type invariants .
this is due to the fact that @xmath77 , i.e , at @xmath78 the burau matrix is a finite type invariant of order 0 . likewise
the lawrence krammer matrix becomes a permutation matrix at @xmath79 and @xmath80 . in section [ sec : fini ] we show that the matrix @xmath81 is of finite type of order @xmath82 . therefore the lawrence
krammer representation is dominated by the finite type invariants . with the faithfulness of the representation together , we obtain the following well - known fact . the finite type invariants separate braids .
the author is interested in the question whether the finite type conjugacy invariants of braids are strong enough to distinguish all the conjugacy classes of braids .
the characteristic polynomial @xmath8 of the lawrence krammer matrix is rather strong as a conjugacy invariant , and is dominated by the finite type invariants . after observing that @xmath8
does not detect the orientation reversal of a braid , we raise the following question .
is there a finite type conjugacy invariant @xmath83 of braids such that @xmath84 for some braid @xmath63 ?
let @xmath85 $ ] be the ring of two variable laurent polynomials in @xmath86 and @xmath87 .
the lawrence
krammer module @xmath88 is a @xmath4 dimensional free @xmath33module generated by the set of standard forks @xmath89 .
the right action of @xmath90 on @xmath88 defined by ( [ eq : kram ] ) gives the faithful representation @xmath91 @xcite .
we follow the convention of @xcite on the sign of variable @xmath86 .
@xmath92 let @xmath93 be the bijection of index sets defined by @xmath94 .
this is the lexicographic order on the double indices .
we identify @xmath95 with the @xmath96-th standard base element @xmath97 , @xmath98 , which is the row vector whose @xmath99-th entry is @xmath100 and uniquely non - zero .
@xmath101 denotes the @xmath96-th row of the matrix @xmath102 .
we need only to show the relation holds for @xmath103 since @xmath27 s are all conjugate to each other .
firstly it can be routinely verified for @xmath104 that @xmath105 from this we calculate the images of @xmath106 and @xmath107 by @xmath108 as follows .
@xmath109 @xmath110 we also have @xmath111 the previous calculations exhibit that @xmath112 projects @xmath88 to the submodule generated by @xmath113 .
the equality @xmath114 follows from @xmath115 . from the proof of proposition [ prop : cubic ]
we can see that @xmath116 is an idempotent .
we denote this matrix by @xmath117 .
@xmath118 the first column of @xmath119 is the only non - zero column of @xmath120 .
we denote by @xmath121 the column vector given by the first column of @xmath120 so that the expression @xmath120 makes sense also as a matrix multiplication .
let @xmath122 be the set of matrices representing endomorphisms of @xmath88 .
@xmath122 assumes a @xmath90-bimodule structure by multiplications from the right side and from the left .
let @xmath123 be the @xmath90-bisubmodule of @xmath122 generated by the matrix @xmath120 , i.e. , @xmath124 for @xmath60 , let @xmath125 denote the permutation braid @xmath126 @xcite .
@xmath125 induces the permutation @xmath127 on @xmath128 which maps @xmath129 to @xmath100 , @xmath130 to @xmath131 , and whose all inversions involve either @xmath129 or @xmath130 .
let @xmath132 .
we call @xmath133 a _ standard dual fork_. we call @xmath134 a _
standard bifork_. we denote @xmath135 by @xmath136 in emphasis of the right and left braid action on @xmath123 .
the following lemma is the multiplication table for standard biforks .
[ lem : multab ] for @xmath60 and @xmath137 , the @xmath138 matrix @xmath139 is given as follows .
@xmath140 this multiplication table can be verified routinely from the definition of @xmath141 , @xmath142 , and the lawrence krammer representation .
one may first calculate the table on @xmath143 ( see ( [ eq : j4 ] ) ) in which every case of the double index correlations listed in the table occurs , and then use it for general @xmath11-braids .
we will not exhibit all the complicated calculations which could be too distracting . in section [ sec : geom ]
we introduce geometric bifork algebra and calculate the same multiplication table . for each picture - based calculation line in the proof of theorem [ thm : isom ]
the same calculation on biforks can be done in parallel by translating the isotopies to braid relations .
[ lem : gen ] the set of standard biforks @xmath144 generates the algebra @xmath123 as a @xmath33module .
we need to show that for all @xmath145 @xmath146 is a linear combination of standard biforks .
@xmath147 is a linear combination of standard forks from the definition of the lawrence krammer representation . from the following formula we can see that @xmath148 is also a linear combination of standard dual forks .
@xmath149 1 .
if @xmath150 , then @xmath151 if @xmath152 , then @xmath153 . if @xmath154 , then @xmath155 . for these two cases , the same calculation works as for the case @xmath150 .
if @xmath156 , then @xmath157 3 . if @xmath158 , then @xmath159 4 . if @xmath160 , then @xmath161 5 . if @xmath162 , then @xmath163 6 .
if @xmath164 , then @xmath165 the formula ( [ eq : invdual ] ) of @xmath166 for the left action of @xmath90 on @xmath123 , presented in the proof of lemma [ lem : gen ] , strikingly exhibits the duality between forks and dual forks , and between the right action and the left action of braids .
if we replace in ( [ eq : invdual ] ) @xmath167 with @xmath168 , the left multiplication with the right one , @xmath86 with @xmath169 , @xmath87 with @xmath170 , and @xmath171 with @xmath95 , then we obtain exactly the same formula ( [ eq : kram ] ) used in the definition of the lawrence krammer representation .
this observation interprets that if @xmath172 then @xmath173 where @xmath174 denotes the matrix @xmath175 .
[ lem : indep ] the set of standard biforks @xmath176 is linearly independent .
let @xmath177 be the @xmath45 matrix defined by @xmath178 , the multiplication table given in lemma [ lem : multab ] . at @xmath80 , @xmath46 is a diagonal matrix with non - zero diagonal entries , which implies @xmath179 even for generic @xmath87 .
suppose that @xmath180 for some @xmath45 matrix @xmath181 .
then for each @xmath182 @xmath183 which implies @xmath184 .
since @xmath46 is non - singular , @xmath185 for each @xmath186 .
lemma [ lem : indep ] shows in particular that the set of standard dual forks @xmath187 is linearly independent . given a matrix @xmath188 , @xmath49 denotes the conjugate transpose @xmath189 .
here we assumed @xmath86 and @xmath87 are evaluated at generic unit complex numbers so that @xmath169 and @xmath170 are complex conjugates of @xmath86 and @xmath87 .
let @xmath190 .
@xmath191 where @xmath192 and @xmath193 if @xmath194 . as shown above , @xmath46 is the same matrix given as the multiplication table in the proof of lemma [ lem : indep ] , where we showed that @xmath46 is non - singular .
it suffices to show the identity @xmath195 for the lawrence
krammer matrices @xmath196 of the artin generators @xmath168 of @xmath90 .
@xmath197 [ coro : incomplete ] let @xmath198 be the characteristic polynomial of @xmath5 .
then @xmath199 equals @xmath200 up to multiplication by units in @xmath47 $ ] . by theorem [ thm :
uni ] , @xmath201 as matrices over the quotient field @xmath202 .
@xmath203 let @xmath204 for @xmath60 , where @xmath205 denote the square root of the full twist that generates the center of @xmath90 . for each @xmath206 , we have @xmath207 .
[ lem : invkram ] with the base set @xmath208 , the lawrence
krammer representation is given by following formula .
@xmath209 from the definition ( [ eq : kram ] ) of the lawrence
krammer representation the formula for @xmath210 can be easily derived as follows .
@xmath211{.75\textwidth } \vskip-.5\baselineskip \begin{alignat}{1 } & x_{i , j } \qquad\qquad \quad \text{if $ m < i-1 $ or $ i < m < j-1 $ or $ j < m$ } \label{eq : inv1 } \\ & x_{i+1,j } \qquad\qquad \text{if $ m = i < j-1 $ } \label{eq : inv2 } \\ & x_{i , j+1 } \qquad\qquad \text{if $ j = m$ } \label{eq : inv3 } \\ & -t^{-1 } q^{-2 } x_{i , j } \quad \ !
\text{if $ m = i = j-1 $ } \label{eq : inv4 } \\ & q^{-1 } x_{i-1,j } + ( 1-q^{-1 } ) x_{i , j } - q^{-1 } ( 1-q^{-1 } ) x_{i-1,i } \label{eq : inv5 } \\ & \qquad \qquad \qquad \quad \text{if $ m = i-1 $ } \nonumber \\ & q^{-1 } x_{i , j+1 } + ( 1-q^{-1 } ) x_{i , j } + t^{-1 } q^{-1 } ( 1-q^{-1 } ) x_{j , j+1 } \label{eq : inv6 } \\ & \qquad \qquad \qquad \quad \text{if $ i <
m = j-1 $ } \nonumber \end{alignat } \end{minipage } \right.\ ] ] the formula of this lemma is verified by the following routine calculations . 1 .
if either @xmath212 , @xmath213 or @xmath214 , then @xmath215 , @xmath216 , or @xmath217 respectively . therefore @xmath218 2 . if @xmath219 , then @xmath220 so that by ( [ eq : inv3 ] ) @xmath221 3
. if @xmath222 , then @xmath223 so that by ( [ eq : inv2 ] ) @xmath224 4 . if @xmath225 , then @xmath226 so that by ( [ eq : inv4 ] ) @xmath227 5 . if @xmath228 , then @xmath229 so that by ( [ eq : inv6 ] ) @xmath230 6 . if @xmath164 , then @xmath231 so that by ( [ eq : inv5 ] ) @xmath232 for a word @xmath233 , @xmath234 denotes the reverse word of @xmath233 .
for a braid @xmath235 written as a word in artin generators , we define @xmath236 .
in other words @xmath237 is the anti - isomorphism given by @xmath68 .
geometrically this equals reversing the orientations of the strings of a braid .
observe in the formula ( [ eq : dualkram ] ) of lemma [ lem : invkram ] that if one replaces @xmath167 , @xmath238 , @xmath169 and @xmath170 with @xmath168 , @xmath95 , @xmath86 and @xmath87 , then one obtains exactly the same formula ( [ eq : kram ] ) .
[ thm : base ] there exists an invertible @xmath45 matrix @xmath239 over @xmath47 $ ] such that for each @xmath11-braid @xmath6 , the equality @xmath240 holds .
let @xmath241 for @xmath60 and define the matrix @xmath239 by @xmath242 , the @xmath243-th entry of @xmath244 for @xmath186 .
it suffices to show the equality for @xmath245 .
the previous observation on the similarity between lemma [ lem : invkram ] and the definition of the lawrence
krammer representation interprets that @xmath246 . taking the @xmath243-th entry of each side , we have @xmath247 from the definition of @xmath238 , @xmath248 for some permutation matrix @xmath249
. hence @xmath239 is invertible .
let @xmath250 .
then we have @xmath251 since transposition and conjugation do not alter the characteristic polynomial of a matrix , we obtain the following corollary from theorem [ thm : rev ] . [ coro : rev ] @xmath5 and @xmath252 have the same characteristic polynomial .
in this section we exhibit how to compute the matrices @xmath46 and @xmath69 of the main theorems for low braid index .
let @xmath253 .
@xmath143 is generated by the two elements @xmath254 and @xmath255 .
note that @xmath256 and @xmath257 .
for the proof of theorem [ thm : uni ] for 4-braids , it is enough to verify the equalities @xmath258 .
the three matrices @xmath259 , @xmath260 and @xmath261 can be explicitly written as follows : @xmath262 @xmath263 where the @xmath264 matrices @xmath171 are given by @xmath265 and @xmath266 the lawrence
krammer matrices of artin generators @xmath267 of @xmath268 can written as follows : @xmath269 let @xmath270 be the @xmath271 matrix taken from the first column of @xmath272 .
let @xmath273 where @xmath274 and @xmath275 . the three columns of @xmath276
are explicitly written as follows : @xmath277 let @xmath278 in @xmath268 . then @xmath279 is related to @xmath63 by a flype move , which changes the conjugacy class while preserving the link type of a closed braid @xcite .
one may check the equality @xmath280 and that @xmath281 shares with @xmath5 the same characteristic polynomial .
in this section we review the squier s result that the burau representation is unitary .
the reduced burau representation @xmath282)$ ] is defined by these two @xmath283 matrices : @xmath284 where @xmath285 denotes the elementary matrix whose only non - zero entry is the @xmath286 entry with value @xmath100 , and @xmath287 .
[ thm : squier ] there exists a nonsingular @xmath283 matrix @xmath288 over @xmath289 $ ] such that for each @xmath63 in @xmath90 it follows that @xmath290 .
the reduced burau representation can be interpreted as the action on the homology group @xmath291)$ ] of the infinite cyclic cover @xmath292 induced by the braid homeomorphism .
it is natural to expect the reduced burau representation is unitary because homeomorphisms should preserve intersection forms .
we clarify this point in the following proof .
the original proof in @xcite was done by giving the matrix @xmath288 and directly evaluating the equality .
consider the pairing @xmath293 $ ] defined by @xmath294 where @xmath295 denotes the usual algebraic intersection number .
the pairing @xmath44 is sesquilinear .
the equalities @xmath296 and @xmath297 follow from the definition .
if an automorphism @xmath298 of @xmath299 is induced by a homeomorphism @xmath300 , then @xmath301 we embed @xmath10 in the complex plane @xmath302 so that the @xmath129-th hole is placed around the point @xmath303 .
let @xmath304 denote the standard generator represented by the closed curve winding only around the @xmath129-th hole once .
let @xmath305 be the cycle which is the lift of @xmath306 .
then @xmath307 is a base for the free @xmath289$]module @xmath308 .
let @xmath309 denote the lift of the relative cycle connecting the @xmath129-th puncture boundary to the @xmath310-st one by a straight segment , for @xmath311 .
note that @xmath312 maps to @xmath313 by the inclusion @xmath314 .
let @xmath315 so that @xmath316 and @xmath317 .
let @xmath318 .
then the equality @xmath319 follows as below : @xmath320 that @xmath321 can be shown easily by evaluating @xmath288 at @xmath322 . given a relative cycle @xmath323 , the action of @xmath324 on @xmath325 is determined by the intersection number @xmath326 as follows : @xmath327 as pointed in @xcite . from this
we obtain @xmath328 this equation implies that the @xmath129-th row of the matrix @xmath329 is the unique non - zero row with @xmath330 as the @xmath130-th entry .
therefore we calculate the intersection pairing @xmath288 by @xmath331 the analogous formula for the lawrence krammer representation obtained in the previous section
is : @xmath332 where @xmath333 denotes the band generator @xmath334 .
in this section , we define geometric biforks and a skein algebra @xmath335 generated by the geometric biforks .
let @xmath336 be a set of @xmath11 distinct points in a disk @xmath337 .
then @xmath338 $ ] is the set of @xmath339 distinct points on the top and bottom of the solid cylinder @xmath340 $ ] .
geometric @xmath11-braid _ is a disjoint union of @xmath11 strings in @xmath340 $ ] having no local maxima or minima with their end points fixed in @xmath341 .
geometric bifork _ is a disjoint union of @xmath11 strings in @xmath340 $ ] having exactly one local maximum and one local minimum with one additionally attached string , which we call a _
handle _ , connecting the maximum point to the minimum point without touching the other @xmath342 strings nor making a local extremum .
we distinguish the handle from the other strings by drawing it with a wavy line as in figure [ fig : bifork ] .
the two strings to which the handle is attached are called _ tines_. two geometric biforks related by an isotopy which does not create a new local extremum are considered to be the same .
we construct an arbitrary geometric bifork as follows .
connect @xmath343 to @xmath344 by a string with exactly one local minimum in @xmath345 $ ] and connect @xmath346 to @xmath347 by a string with exactly one local maximum .
then connect the local minimum to local maximum by a straight wavy line .
for @xmath348 , connect @xmath349 to @xmath350 by straight strings .
now we obtained a simple geometric bifork @xmath351 which looks like a generator of the birman murakami
wenzl algebra except that it has a handle attached .
we may attach arbitrary braids @xmath63 and @xmath26 to the top and the bottom of the geometric bifork to obtain a general geometric bifork @xmath352 where @xmath353 .
one can easily see that every geometric bifork can be written as @xmath352 .
we define a @xmath33algebra @xmath335 generated by geometric biforks with the following relations . @xmath354
the multiplication in @xmath335 is given by concatenation as in the braid groups .
the relations ( [ eq : f1],[eq : f2 ] ) come from the definition that @xmath355 and @xmath356 .
the relations ( [ eq : f3 ] ) mean that @xmath357 for @xmath358 . the first relation in ( [ eq : f4 ] )
reflects that @xmath359 and the second one is from @xmath360 .
let @xmath361 .
we call @xmath362 a _ standard geometric bifork_. figure [ fig : bifork ] shows a typical one .
-.5 given an arbitrary geometric bifork @xmath363 , we can express @xmath363 as a linear combination of other geometric biforks , which have less under - crossings of tines than @xmath363 , by using the relations ( [ eq : f1][eq : f4 ] ) .
iterating this procedure , we express @xmath363 as a linear combination of standard geometric biforks .
therefore the set of standard geometric biforks generates @xmath335 as a @xmath33module .
one may check that the formulas ( [ eq : kram ] ) and ( [ eq : invdual ] ) with @xmath364 and @xmath365 in place of @xmath141 and @xmath142 , also hold by applying the relations ( [ eq : f1][eq : f4 ] ) .
the previous observation on the relations ( [ eq : f1][eq : f4 ] ) implies that there exists a surjective @xmath33homomorphism @xmath366 , @xmath367 .
[ thm : isom ] the geometric bifork algebra @xmath335 is isomorphic to the matrix algebra @xmath123 of biforks . by lemma [ lem : indep ] ,
the algebra @xmath123 is a free @xmath33module with rank @xmath368 .
the fact that @xmath335 is generated by @xmath369 many elements implies that the surjective homomorphism @xmath366 is also injective . in order to see
@xmath370 is an algebra homomorphism , we need @xmath371 . in the following
we verify the equality @xmath372 . 1 . for @xmath373 @xmath374 2 . for @xmath375 @xmath376 3
. for @xmath377 or @xmath378 @xmath379 4 . for @xmath380 @xmath381 5 . for @xmath382 or @xmath383 @xmath384 6 . for @xmath385 @xmath386 7
. for @xmath387 @xmath388 8 . for @xmath389 @xmath390
let @xmath391 be the ideal of the integral group ring @xmath392 $ ] generated by @xmath393 .
let @xmath394 be an abelian group .
if a @xmath395module homomorphism @xmath396 \to a$ ] vanishes on @xmath397 , we call @xmath83 a _ finite type invariant of order @xmath398_. if @xmath399 for each @xmath353 , we call @xmath83 a _ conjugacy invariant_. at @xmath79 and @xmath80 , it is easy to check that @xmath403 , so that at @xmath79 and @xmath80 , @xmath1 is a finite type invariant of order @xmath404 . in the series expansion of @xmath405 at @xmath79 and @xmath80 ,
the lowest degree is at least @xmath100 . in other words
, @xmath406 has the lowest degree at least @xmath100 for each @xmath407 .
if @xmath408 , then the lowest degree in the series expansion of @xmath406 is at least @xmath409 since @xmath233 is a linear combination of products of @xmath410 elements of @xmath411 .
the theorem follows from @xmath412 for @xmath413 .
|
we show that the lawrence
krammer representation is unitary .
we explicitly present the non - singular matrix representing the sesquilinear pairing invariant under the action .
we show that reversing the orientation of a braid is equivalent to the transposition of its lawrence krammer matrix followed by a certain conjugation . as corollaries
it is shown that the characteristic polynomial of the lawrence krammer matrix is invariant under substitution of its variables with their inverses up to multiplication by units , and is not a complete conjugacy invariant for braids .
|
there is an extensive theoretical and empirical literature on estimating average causal effects of binary treatments in observational studies based on the assumption of unconfoundedness or ignorable treatment assignment . under this assumption differences in outcomes for units with different treatment levels , but the same values for pre - treatment variables , can be interpreted as estimates of causal effects .
much of the literature builds on the seminal paper by @xcite ( rr83 from here on ) which clarified the central role of the propensity score ( the conditional probability of receiving the treatment given the pre - treatment variables or covariates ) in analyses of causal effects in such settings , and which proposed a number of widely used estimators .
see @xcite for a textbook treatment .
although important for empirical practice , much less theoretical work has been done on the setting with more than two treatment levels ( exceptions include @xcite ; @xcite ; @xcite ) . because in settings with multi - level treatments there is no scalar function of the covariates that has all the properties that rr83 presents for the propensity score in the binary treatment case , it has been claimed that there is no natural analogue to matching and subclassification on the propensity score ( @xcite ) . in the main contribution of the current paper
we show that , contrary to these claims , the essence of the results in rr83 generalizes to the setting with multi - level treatments .
in particular , we develop methods for matching and subclassification on scalar functions of the covariates that are valid irrespective of the number of distinct levels for the treatment .
the key insight is that we do not construct sets of units with the balancing property that within these sets the treatment level is independent of the covariates . doing so would require adjusting for the vector of propensity scores with length equal to the number of treatment levels minus one .
instead we focus on estimating the average of the potential outcomes separately for each treatment level , which requires adjusting only for the probability of receiving that particular level of the treatment .
this insight allows us to extend some of the most widely used methods for the binary treatment case to the multi - level treatment case without giving up the dimension reducing property of the propensity score .
we provide some simulation evidence that demonstrates the relevance of concerns with the previously proposed estimators and the promise of the new methods .
following @xcite and rr83 we use the potential outcome set up , generalized to the case with more than two , unordered , levels for the treatment in @xcite , @xcite , @xcite , and @xcite .
the treatment is denoted by @xmath0 . in the standard binary treatment case @xmath1 , the two treatments are often labeled treatment and control . for each unit
@xmath2 there are @xmath3 potential outcomes , one for each treatment level , denoted by @xmath4 , for @xmath5 . implicitly in this notation is the assumption that there is no interference between units and no versions of each treatment level ( the stable - unit - treatment - value assumption , or sutva , @xcite ) . the observed outcome for unit
@xmath2 is the potential outcome corresponding to the treatment received : @xmath6 we also observe a vector - valued covariate or pre - treatment variable , denoted by @xmath7 .
these pre - treatment variables are known a priori not to be affected by the treatment , typically measured prior to the determination of the treatment level , and so there are no multiple versions of these covariates corresponding to the different levels of the treatment
. these pretreatment variables may include fixed attributes of the units , or measurements prior to the treatment assignment that predict the outcome , for example prior health status .
although we do not stress this in the notation , there is implicitly a temporal aspect to the study with three stages : first , the pre - treatment variables are measured , or at least they assume their values prior to the assignment of the treatment ; second , the treatment is assigned or selected ; third , the outcome assumes its value and is measured .
we assume the sequence @xmath8 @xmath9 with the potential outcomes is i.i.d .
, so that the sequence of realized values @xmath10 @xmath11 is also i.i.d . following the literature ( e.g. , rr83 ) ,
we focus on average treatment effects as the causal estimands .
this is less restrictive than it may appear at first because we can first take transformations of the outcomes and pre - treatment variables . for the comparison between treatment levels @xmath12 and @xmath13 ,
the average effect is @xmath14.\ ] ] the expectation is taken with respect to the same population ( called the target population , @xcite ) for different treatment - level pairs @xmath15 .
some researchers , when analyzing data with multi - level treatments , have used conventional methods for comparing two treatment levels at a time .
often such analyses use only information on units exposed to one of those two treatment levels , which would lead to estimates of @xmath16 $ ] .
if the subpopulation of units with treatment levels @xmath12 or @xmath13 is different in terms of potential outcome distributions , these estimands are generally different from the @xmath17 defined in ( [ 2.1 ] ) , because the latter do not condition on @xmath18 . as a result
such binary - comparison analyses make it difficult to compare @xmath16 $ ] and @xmath19 $ ] because they refer to different populations . in this paper
we mainly focus on the case where the different treatment levels are qualitatively distinct . in that cases
the interest is typically in average effects of the form @xmath17 .
in other cases , however , the treatment levels may measure the quantity of a dose .
in such cases the researcher may be interested in weighted combinations of average effects .
for example , one might be interested in @xmath20 , with the weights adding up to one , which would correspond to a weighted average of unit increases in the dose .
one advantage of such estimands is that their variance may be lower than that for particular @xmath17 . in this case there
may also be particular interest in @xmath21 , the effect of the maximum dose .
our results also apply to all such estimands .
our focus is on observational studies where assignment to treatment is not completely random . instead ,
following a large strand of the observational studies literature , we assume that assignment to treatment is unconfounded so that , within subpopulations that are homogenous in observed pre - treatment variables , assignment to treatment is as good as random .
this is strictly weaker than complete randomization by allowing for general associations between the treatment level and the pre - treatment variables . in this section
we discuss the generalization of the notion of the propensity score , introduced in the causality literature by rr83 for the binary treatment case , to our setting with multi - level treatments . in the binary treatment case rr83
defines the propensity score as the conditional probability of receiving the active treatment rather than the control treatment , @xmath22 .
here we generalize that to the multi - level treatment case , following @xcite : ( generalized propensity score ) the generalized propensity score is the conditional probability of receiving each treatment level : @xmath23 before formally discussing unconfoundedness assumptions , let us assume that there is overlap in the covariate distributions : ( overlap ) for all values of @xmath24 the probability of receiving any level of the treatment is positive : @xmath25 without this assumption there will be values of @xmath24 for which we can not estimate the average effect of some treatments relative to others without relying on extrapolation . in section 6
we discuss methods for constructing a subsample with better overlap for cases where this assumption is ( close to ) violated .
we start by generalizing the conventional rr83 version of the unconfoundedness assumption to the case with multi - level treatments .
we refer to this as strong unconfoundedness to distinguish it from the weaker condition of weak unconfoundedness .
( strong unconfoundedness ) the assignment mechanism is strongly unconfounded if @xmath26 [ strong ] here we use the @xmath27 notation introduced by @xcite to denote ( conditional ) independence .
the assumption of strong unconfoundedness has no testable implications . in a particular application
the assumption is a substantive one , and often a controversial one .
often it can be made more plausible by collecting detailed information at baseline on characteristics of the units that are related to treatment and outcome . as a result
the dimension of @xmath7 may be high .
one implication of strong unconfoundedness is the following extension of the propensity score result in rr83 to the multi - level treatment case : [ lemma1 ] ( rr83 ) suppose the assignment mechanism is strongly unconfounded . then @xmath28 because @xmath29 , it follows that @xmath30 is a linear combination of @xmath31 , and so we do not need to include @xmath32 in the conditioning set .
if there are two levels of the treatment , the result in the lemma reduces to the result in rr83 . as pointed out in @xcite and @xcite , the dimension reduction property of the lemma
depends on the number of distinct levels for the treatment , and therefore the result is less useful in settings with a substantial number of treatment levels .
the problem is that without additional assumptions there is in general no scalar function @xmath33 of the covariates such that @xmath34 , suggesting that the advantages of the propensity score approach do not carry over to the multi - level treatment case .
@xcite discuss additional assumptions under which functions @xmath35 exist with this property and whose dimension is lower than @xmath36 . in particular , @xcite assume that a scalar balancing function @xmath35 exists and propose a matching estimator based on @xmath35 , and @xcite propose a subclassification estimator under this assumption .
nevertheless , in general such functional form assumptions may be controversial .
we improve the dimension reduction property of the generalized propensity score by weakening the requirement of strong unconfoundedness condition to weak unconfoundedness .
define the @xmath3 indicator variables @xmath37 : @xmath38 in terms of these indicator variables strong unconfoundedness is equivalent to @xmath39 now we can formulate the weak unconfoundedness notion , introduced in @xcite .
( weak unconfoundedness ) the assignment mechanism is weakly unconfounded if for all @xmath40 , @xmath41 although formally it is obviously weaker , we do not wish to argue that weak unconfoundedness is substantively weaker than strong unconfoundedness .
in fact neither have testable implications , and there appear to be no interesting estimands that are identified under the stronger assumption but not under the weaker assumption . rather , the two key insights , and the motivation for distinguishing between the two unconfoundedness assumptions , are , one , that , as shown in lemma [ lemma2 ] below , weak unconfoundedness is preserved if we condition on a scalar function of the pretreatment variables , whereas preserving strong unconfoundedness requires conditioning on a set of @xmath36 functions of the pre - treatment variables , as shown in lemma [ lemma1 ] , and two , that weak unconfoundedness is sufficient for identifying average causal effects , as formalized in lemma [ lemma3 ] below .
[ lemma2 ] ( weak unconfoundedness ) suppose the assignment mechanism is weakly unconfounded .
then for all @xmath40 , @xmath42 [ lemma3 ] ( average causal effects under weak unconfoundedness ) suppose the assignment mechanism is weakly unconfounded .
then @xmath43 = \mathbb{e}\bigl [ \mathbb{e}[y_i^{\rm obs}\mid w_i = w',p(w'\mid x_i)]\bigr]- \mathbb{e}\bigl [ \mathbb{e}[y_i^{\rm obs}\mid w_i = w , p(w\mid x_i)]\bigr].\ ] ] lemma [ lemma3 ] is the key result . for its interpretation
it is useful to compare it to the standard result under strong unconfoundedness . under the strong unconfoundedness assumption
we create subpopulations where we can simultaneously compare units with all different levels of the treatment , leading to @xmath44\ ] ] @xmath45\bigr]- \mathbb{e}\bigl [ \mathbb{e}[y_i^{\rm obs}\mid w_i = w , p(1\mid x_i ) , \ldots , p(t-1\mid x_i)]\bigr]\ ] ] @xmath46\bigr]\ ] ] to allow for comparisons of all treatments these subpopulations were defined by common values for the full set of @xmath36 propensity scores @xmath47 . under weak unconfoundedness
we do not , and in fact can not , construct such subpopulations .
however , in order to estimate the average effect @xmath48 $ ] it is not necessary to do so .
instead , we construct , for each treatment level @xmath12 separately , subpopulations where we can estimate the average value of the potential outcomes , but only for that single treatment level . for treatment level @xmath12
these subpopulations are defined by the value of a single score , @xmath49 , leading to the equality @xmath50 = \mathbb{e}\bigl [ \mathbb{e}[y_i^{\rm obs}\mid w_i = w , p(w\mid x_i)]\bigr].\ ] ] that difference in focus allows us to reduce the dimension of the conditioning variable to a scalar , irrespective of the number of treatment levels .
in this section we discuss matching methods .
first we discuss conventional matching on the full set of pre - treatment variables .
this is not a new method , but it will be useful to contrast with the proposed methods . then we discuss how the generalized propensity score can be used to develop a new matching estimator that matches only on a scalar function of the pre - treatment variables .
@xcite demonstrates covariate matching in multi - level treatments . here
we focus on nearest neighbor matching .
other modifications include multiple nearest neighbors matching , kernel matching and so forth .
reviews of general matching methods can be found in @xcite .
define the covariate matching function @xmath51 as the index for the unit with treatment level @xmath12 that is closest to @xmath24 in terms of covariates ( ignoring ties ) : @xmath52 here we use @xmath53 to denote a generic metric .
in practice one would typically use the mahalabonis metric , where @xmath54 , with @xmath55 , and @xmath56 . note that the set of indices we search over includes all units , including unit @xmath2 itself , so that for all @xmath2 , @xmath57 .
given the covariate matching function @xmath58 the potential outcomes for unit @xmath2 are imputed as @xmath59 for @xmath60 .
now we estimate @xmath17 as @xmath61 note that to estimate @xmath17 we impute potential outcomes @xmath62 and @xmath63 even for units who did not receive either treatment level @xmath12 or @xmath13 .
this ensures comparability of average treatment effects for different pairs of treatments .
just as in the binary treatment setting , matching on all covariates is not an attractive procedure in the multi - level treatment setting if the number of covariates is substantial ( _ e.g. _ , @xcite ) . in the binary treatment case rr83
proposed matching on the propensity score to reduce the dimensionality of the matching problem . if @xmath64 is the rosenbaum - rubin propensity score , the matching function for the binary treatment case would be @xmath65 one could generalize that to the multi - level treatment case by matching on the full set of scores , leading to @xmath66 here we generalize this to the case with multi - level treatments in a way that allows for a scalar matching variable . in this case
matching is conceptually quite different from matching on covariates .
we separate the estimation of @xmath67-\mathbb e[y_i(w)]$ ] into the two terms .
first we focus on the problem of estimating @xmath68 $ ] .
define the generalized propensity score matching function as @xmath69 here the treatment level @xmath12 enters into the matching function not only by limiting the set of potential matches to the set of units with @xmath70 , but also in the function of the covariates that is being matched , @xmath71 . in covariate and conventional propensity score matching the treatment level only affects the set of potential matches .
given the generalized propensity score matching function we impute @xmath62 as @xmath72 the average effect is estimated as @xmath73 note that the difference @xmath74 in ( [ drie ] ) is _ not _ generally an estimate of an average causal effect , whereas in ( [ twee ] ) the difference @xmath75 _ is _ an estimate of the average causal effect @xmath76 $ ] . in the binary treatment case
this distinction between covariate and propensity score matching does not matter : in that case there are only two values for @xmath12 , @xmath77 , so that @xmath78 , and therefore matching on @xmath79 is the same as matching on both @xmath79 and @xmath80 , and the same as matching on @xmath80 .
in web appendix , we provide mathematical details for inference and show that under mild regularity conditions , the matching estimator based on the generalized propensity score or the estimated generalized propensity score is asymptotically normally distributed .
in the binary treatment literature , a common alternative to matching is subclassification or stratification on the propensity score , originally proposed by rr83 . to put our proposed methods for the multi - level treatment case in perspective ,
let us briefly summarize their approach for the binary treatment case in our current notation to show why it does not directly extend to the multivalued treatment case . divide the sample into a number of subclasses by the value of the propensity score @xmath64 . based on cochran ( 1968 ) who shows that this removes much of the bias
, researchers often use five subclasses . to be specific ,
let @xmath81 be the @xmath82th quintile of the empirical distribution of @xmath83 , for @xmath84 , and define @xmath85 and @xmath86 .
then we construct the five subclasses , based on the propensity score being between @xmath87 and @xmath81 . for subclass @xmath82
one can estimate the average causal effect of treatment @xmath88 versus treatment @xmath89 as @xmath90 where @xmath91 is the number of units in subclass @xmath82 with treatment level @xmath12 . the overall average treatment effect is then estimated by averaging over the subclasses : @xmath92 because all @xmath93 are close to equal , at most differ by @xmath88 , ( assuming there are no ties ) this is essentially a simple arithmetic mean of the @xmath94 estimates @xmath95
. now consider the multi - level treatment case .
we are interested in @xmath17 for some pair of treatment levels @xmath12 and @xmath13 . again , and this is a cornerstone of our approach , we write this as a difference of two expectations , @xmath96- \mathbb e[y_i(w)]$ ] and separately estimate the two terms @xmath97 $ ] and @xmath98 $ ] . to estimate the second term , @xmath99 $ ]
we construct subclasses or strata based on @xmath100 .
let @xmath101 be the quintiles of @xmath102 in the sample .
then the average value of @xmath62 in subclass @xmath82 is estimated as @xmath103 where @xmath91 is the number of units with @xmath104 and @xmath105 .
the overall average of @xmath62 is then estimated as @xmath106 = \sum\limits_{j=1}^5 \frac{n_{w}}{n}\cdot \hat{\mu}_{jw}.\ ] ] the key is that , in contrast to what is done in the binary treatment case , we do not construct subclasses defined by similar values for the @xmath36 propensity scores such that we can estimate causal effects within the subclasses .
instead we construct subclasses defined by similar values for a single propensity score at a particular treatment level so that we can estimate the average potential outcome for that treatment level within the subclasses , and we do so separately for each treatment level , with different subclasses for each treatment level . in the binary treatment case
this amounts to the same thing because the two propensity scores @xmath64 and @xmath107 are linearly related , but in the multi - level case these two approaches are different .
we focus on assessing balance in the covariate distributions in terms of the propensity score as well as directly in terms of the covariates , following the discussion in @xcite for the binary case . for each treatment level @xmath12
, we calculate the average values for each component of the covariate vectors and their corresponding sample variances : @xmath108 define also for each treatment level the average value of the covariates for units with a treatment level different from @xmath12 and the average variance : @xmath109 respectively . the first approach to assessing
the covariate balance is to inspect the normalized differences for each covariate and each treatment level : @xmath110 we can also assess balance by looking at the generalized propensity score . for each treatment level we can calculate the normalized difference for the generalized propensity score for that treatment level : @xmath111 where @xmath112 and @xmath113 .
finally , one may wish to plot a histogram of @xmath102 for the @xmath114 units for @xmath105 and a histogram of @xmath102 for the @xmath115 units with @xmath116 in the same figure . in many applications
there are regions of the covariate space with low values for the probability of receiving one of the treatments .
this is likely in the setting with a binary treatment , but even more likely to be an issue in settings with many treatment levels . of note , lack of overlap affects the credibility of all methods attempting to estimate all pairwise average causal effects from the common population
in that case we may wish to modify the estimands to average only over the part of the covariate space with all treatment probabilities away from zero .
the question is how to choose the set of covariates with overlap . for the binary treatment case
, @xcite proposed a method for improving overlap by trimming the sample .
specifically they suggest dropping units from the analysis with low and high values of the propensity score .
the threshold for dropping units is based on minimizing the variance of the estimated average treatment effect on the trimmed sample . by trimming the sample
, this method generally alters the estimand to the so - called feasible estimand , by changing the reference population . using the feasible estimand
is widely recommended in the literature , as long as we are careful to characterize the resulting quantity of interest .
here we generalize the @xcite approach to the multi - level treatment case .
we focus on average treatment effects for subsets of the covariate space . formally , define the conditional average treatment effect : @xmath117.\ ] ]
the semiparametric efficiency bound for @xmath118 is , building on the work by hahn ( 1998 ) and hirano , imbens , and ridder ( 2003 ) , under homoskedasticity and constant treatment effects , @xmath119.\ ] ] in the binary case @xcite proposed choosing @xmath120 to minimize @xmath121 , which leads to dropping units with @xmath122 and units with @xmath123 , with @xmath124 an estimable function of the marginal distribution of the propensity score .
for the multi - level treatment case we suggest focusing on the subset of the covariate space @xmath120 that minimizes @xmath125.\ ] ] under homoskedasticity and a constant treatment effect this will lead to a set @xmath126 of the form @xmath127 where the threshold @xmath128 satisfies @xmath129.\ ] ] to implement the trimming method in practice in the multi - level treatment case we replace the expectation by an average and then find the largest @xmath128 that satisfies the inequality .
in this section we assess the performance of the two new estimators in cases of multi - level treatments ( matching on the generalized propensity score , gpsm , and subclassification on the generalized propensity score , gpss ) in a monte carlo study relative to five previously proposed estimators , first the simple difference in average outcomes ( dif ) by treatment status , second pairwise propensity score matching ( ppsm ) that compares two treatment levels at a time using the binary propensity score matching on the units exposed to one of those two treatment levels , third the estimator based on matching on the set of @xmath36 propensity score set ( pssm ) , fourth the estimator based on weighting , and fifth , matching on all covariates ( cov ) . in the binary treatment and missing data case previous simulations
have found that weighting estimators can have high variability , _
@xcite and @xcite , especially if the probabilities are close to zero .
@xcite found that the weighting estimator was inferior to pairwise matching estimators in terms of root mean squared error .
this is even more likely to be a concern in settings with multiple treatment levels than in the binary treatment case because , with the probabilities for the @xmath3 treatment levels adding up to one , with @xmath3 large some probabilities are likely to be close to zero . because in the binary treatment case it has been found that matching on high - dimensional covariates is not practical for commonly found sample sizes ( _ e.g. , _ the theoretical results in @xcite ) , it is likely that in settings with many treatment levels matching on all scores is not effective either .
these results motivate us to compare the seven estimators in settings with a large number of treatment levels , and where some of the treatment levels have low probability for some covariate values . in the simulations we focus on two designs , one with three treatment levels and one with six treatment levels , and both with six covariates . in the first design with three treatment levels the covariates @xmath130 , and @xmath131 are multivariate normal with means zero , variances of @xmath132 and covariances of @xmath133 ; @xmath134;x_{5i}\sim\chi_{1}^{2}$ ] ; and @xmath135 , with @xmath136 .
the three treatment groups are formed using multinomial regression model @xmath137 where @xmath138 is the treatment indicator , i.e. @xmath139 , if the unit @xmath2 belongs to treatment @xmath12 , and @xmath140 , @xmath141 , @xmath142 , and @xmath143 .
the outcome design is @xmath144 with @xmath145 , @xmath146 , @xmath147 , and @xmath148 @xmath149 .
the sample sizes are @xmath150 , for @xmath151 .
we compare the seven estimators over @xmath152 datasets .
the generalized propensity scores are estimated using multinomial logistic regression model with all covariates entering the model linearly .
@xmath153 confidence intervals for point estimates were calculated using : ( a ) @xmath154 and @xmath155 percentiles from @xmath152 bootstrap samples for dif , gpss , and weighting ; ( b ) point estimate @xmath156 for @xcite variance estimator for cov and pssm ; and for @xcite variance estimator for ppsm and gpsm , which takes into account the uncertainty of the matching procedure and the uncertainty of estimating generalized propensity scores , as in web appendix .
.simulation results , design i. estimators : ( 1 ) dif , simple difference in outcomes for units with different treatment levels ; ( 2)ppsm : pairwise comparison using binary propensity score matching ; ( 3 ) pssm : matching on the propensity score set ; ( 4 ) w : weighting estimator ; ( 5 ) cov : matching on all covariates ; ( 6 ) gpsm : matching on the generalized propensity score ; ( 7 ) gpss : stratification on the generalized propensity score . variance estimators : ( 1 ) bootstrapping variance estimator for dif , gpss , and w ; ( 2 ) abadie and imbens ( 2006 ) variance estimator for cov matching and pssm ; ( 3 ) abadie and imbens ( 2012 ) variance estimator for ppsm and gpsm . [ cols= " <
, < , < , < , < , < , < , < , < , < " , ] [ table : tab4 ]
in this paper , we develop new methods for estimating causal treatment effects using observational data in settings with multiple ( more than two ) treatment levels .
existing methods require additional assumptions assuming the existence of a scalar balance score , so as to facilitate matching and subclassification .
we show that , contrary to claims in the literature , matching and subclassification methods using the propensity score generalize naturally to the multi - level treatment case .
we focus on matching and subclassification on the generalized propensity score using the notion of weak unconfoundedness , and show that adjusting for a scalar function of the covariates can remove all biases associated with differences in observed covariates . as with other propensity based analyses ,
this approach depends on correct specification of propensity score modeling , and does not resolve the potential for bias due to unmeasured confounding .
an initial simulation study and example demonstrated the potential benefits of the proposed approach at reducing bias and providing causal inference comparisons for multiple cohorts on a common population .
this study and the preparation of this paper were funded in full by eli lilly and company , indianapolis , in , usa . editing support was provided by casie polanco , inventivhealth clinical , indianapolis , in , usa . declaration of personal interests : cui , faries , and kadziola are all employees and minor stockholders of eli lilly and company .
supplementary web appendices , referenced in sections 4 and 7 , are available with this paper at the _ biometrics _ website on wiley online library .
r code is available at https://github.com/shuyang1987/multilevelmatching .
bennett , r. m. , bushmakin , a. g. , cappelleri , j. c. , zlateva , g. & sadosky , a. b. ( 2009 ) .
minimal clinically important difference in the fibromyalgia impact questionnairea . _ the journal of rheumatology , _ * 36.6 * , 130411 .
lechner , m. ( 2001 ) . identification and estimation of causal effects of multiple treatments under the conditional independence assumption . _ in econometric evaluations of active labor market policies in europe , m. lechner & f. pfeiereds .
heidelberg : physica _ pp .
lu , b. , zanutto , e. , hornik , r. & rosenbaum , p. r ( 2001 ) . matching with doses in an observational study of a media campaign against drug abuse .
_ journal of the american statistical association _ * 96.456 * 12451253 .
mccaffrey , d.f . , griffin , b.a .
, almirall , d . ,
slaughter , m.e .
, ramchand , r. & burgetteb , l.f.(2013 ) . a tutorial on propensity score estimation for multiple treatments using generalized boosted models .
_ * 32 * , 3388414 .
peng , x. , robinson , r. l. , mease , p. , kroenke , k. , williams , d. a. , chen , y. , faries , d. , wohlreich , m. , mccarberg , b. & hann , d.(2015 ) . long - term evaluation of opioid treatment in fibromyalgia . _ the clinical journal of pain , _ * 31.1 * , 713 .
robinson , r.l . ,
kroenke , k. , mease , p. , williams , d. a. , chen , y. , dsouza , d. , wohlreich , m. & mccarberg , b.(2012 ) .
burden of illness and treatment patterns for patients with fibromyalgia .
_ pain med .
_ * 13 * , 136676 .
zanutto , e. , lu , b. & hirnik , r. ( 2005 ) .
using propensity score subclassification for multiple treatment doses to evaluate a national antidrug media campaign . _ journal of educational and behavioral statistics , _ * 30.1 * , 5973 .
|
in this paper , we develop new methods for estimating average treatment effects in observational studies , in settings with more than two treatment levels , assuming unconfoundedness given pre - treatment variables .
we emphasize propensity score subclassification and matching methods which have been among the most popular methods in the binary treatment literature . whereas the literature has suggested that these particular propensity - based methods do not naturally extend to the multi - level treatment case , we show , using the concept of weak unconfoundedness and the notion of the generalized propensity score , that adjusting for a scalar function of the pre - treatment variables removes all biases associated with observed pre - treatment variables .
we apply the proposed methods to an analysis of the effect of treatments for fibromyalgia .
we also carry out a simulation study to assess the finite sample performance of the methods relative to previously proposed methods .
keywords : generalized propensity score ; matching ; multi - level treatments ; potential outcomes ; subclassification ; unconfoundedness .
|
the discovery of neutrino oscillations has marked a turning point in our understanding of nature and has brought neutrino physics to the center of attention of the particle , nuclear and astrophysics communities . here
i summarize the determination of neutrino mass and mixing parameters in neutrino oscillation studies following ref .
@xcite to which the reader is referred for details on data analysis and experimental references . for future neutrino oscillation projects
see ref .
the structure of the three - flavour lepton mixing matrix in various gauge theories of neutrino mass was given in @xcite .
current neutrino oscillation data are well described by its simplest unitary form , with is no sensitivity to cp violation .
the effect of dirac cp phases in oscillations and majorana phases in constitute the main challenge for the future .
the interpretation of the data requires good calculations of solar and atmospheric neutrino fluxes @xcite , neutrino cross sections and experimental response functions , as well the inclusion of matter effects @xcite in the sun and the earth .
the solar neutrino data includes the rates of the chlorine experiment ( @xmath2 snu ) , the results of the gallium experiments sage ( @xmath3 snu ) and gallex / gno ( @xmath4 snu ) , as well as the 1496day super - k data ( 44 bins : 8 energy bins , 6 of which are further divided into 7 zenith angle bins ) .
the sno data include the data from the salt phase in the form of the neutral current ( nc ) , charged current ( cc ) and elastic scattering ( es ) fluxes , the 2002 spectral day / night data ( 17 energy bins for each day and night period ) and the 391day data .
the analysis includes not only the statistical errors , but also systematic uncertainties such as those of the eight solar neutrino fluxes .
kamland detects reactor anti - neutrinos at the kamiokande site by the process @xmath5 , where the delayed coincidence of the prompt energy from the positron and a characteristic gamma from the neutron capture allows an efficient reduction of backgrounds .
most of the incident @xmath6 s come from nuclear plants at distances of @xmath7 km from the detector , far enough to probe large mixing angle ( lma ) oscillations . to avoid large uncertainties associated with the geo - neutrino flux an energy cut at 2.6 mev prompt
energy is applied for the oscillation analysis .
the first kamland data correspond to a 162 ton - year exposure gave 54 anti - neutrino events in the final sample , after all cuts , while @xmath8 events are predicted for no oscillations with @xmath9 background events , consistent with the no disappearance hypothesis at less than 0.05% probability .
this gave the first evidence for the disappearance of reactor neutrinos before reaching the detector , and thus the first terrestrial confirmation of oscillations with @xmath10 .
additional kamland data with a somewhat larger fiducial volume of the detector were presented at neutrino 2004 , corresponding to an 766.3 ton - year exposure . in total 258 events
have been observed , versus @xmath11 reactor neutrino events expected in the case of no disappearance and @xmath12 background events .
this leads to a confidence level of 99.995% for @xmath13 disappearance .
moreover evidence for spectral distortion consistent with oscillations is obtained . a very convenient way to bin the latest kamland data is in terms of @xmath14 , rather than the traditional bins of equal size in @xmath15 .
various systematic errors associated to the neutrino fluxes , backgrounds , reactor fuel composition and individual reactor powers , small matter effects , and improved @xmath6 flux parameterization are included @xcite .
this singles out the lma solution from the previous `` zoo '' of alternatives @xcite .
the stronger evidence for spectral distortion in these data also leads to improved @xmath10 determination , substantially reducing the allowed region of oscillation parameters . from this point of view kamland
has played a key role in the resolution of the solar neutrino problem .
assuming cpt invariance one can directly compare the information obtained from solar neutrino experiments with the kamland reactor results .
the first evidence for neutrino oscillations was the zenith angle dependence of the @xmath16-like atmospheric neutrino data from the super - k experiment in 1998 , an effect also seen in other atmospheric neutrino experiments .
however , though appealing , the original oscillation interpretation was certainly not unique @xcite . today , thanks to the accumulation of upgoing muon data , and the observation of the dip in the @xmath17 distribution of the atmospheric @xmath18 survival probability , the signature for atmospheric neutrino oscillations has become clear .
the data include super - k charged - current atmospheric neutrino events , with the @xmath19-like and @xmath16-like data samples of sub- and multi - gev contained events grouped into 10 zenith - angle bins , with 5 angular bins of stopping muons and 10 through - going bins of up - going muons .
we do not use @xmath20 appearance , multi - ring @xmath16 and neutral - current events , since an efficient monte - carlo simulation of these data would require further details of the super - k experiment , in particular of the way the neutral - current signal is extracted from the data .
we employ the latest three dimensional atmospheric neutrino fluxes given in @xcite .
@xmath18 disappearance over a long - baseline probing the same @xmath21 region relevant for atmospheric neutrinos is now available from the kek to kamioka ( k2k ) neutrino oscillation experiment .
neutrinos produced by a 12 gev proton beam from the kek proton synchrotron consist of 98% muon neutrinos with a mean energy of 1.3 gev .
the beam is controlled by a near detector 300 m away from the proton target .
comparing these near detector data with the @xmath18 content of the beam observed by the super - k detector at a distance of 250 km gives information on neutrino oscillations . the data k2k - i sample ( @xmath22 protons on target ) gave 56 events in super - k , whereas @xmath23 were expected for no oscillations .
the k2k - ii data correspond to @xmath24 protons on target , comparable to the k2k - i sample .
altogether they give 108 events in super - k , to be compared with @xmath25 expected for no oscillations . out of the 108 events 56
are so - called single - ring muon events .
this data sample contains mainly muon events from the quasi - elastic scattering @xmath26 , and the reconstructed energy is closely related to the true neutrino energy .
the k2k collaboration finds that the observed spectrum is consistent with the one expected for no oscillation only at a probability of 0.11% , whereas the best fit oscillation hypothesis spectrum has a probability of 52% .
one finds that the neutrino mass - squared difference inferred from the @xmath18 disappearance in k2k agrees with atmospheric neutrino results , providing the first confirmation of oscillations with @xmath27 with accelerator neutrinos .
unfortunately in the current data sample k2k gives a rather weak constraint on the mixing angle , due to low statistics .
however , although the determination of @xmath28 is completely dominated by atmospheric data , k2k data already start constraining the allowed @xmath27 region @xcite . in particular
, there is a constraint on @xmath27 from below , which is important for future long - baseline experiments , since these are drastically affected if @xmath27 lies in the lower part of the 3@xmath29 range indicated by current atmospheric data alone .
the first systematic study of the effective lepton mixing matrix in gauge theories of massive neutrinos was given in @xcite .
for some models this matrix can be taken as approximately unitary . for three neutrinos , this gives @xmath30 where each factor is effectively @xmath31 and contains an angle and a cp phase .
two of the three angles are involved in solar and atmospheric oscillations , so we set @xmath32 and @xmath33 .
the last angle in the three
neutrino leptonic mixing matrix is @xmath34 , @xmath35 for which only an upper bound currently exists .
all three phases are physical @xcite , one corresponds to the one present in the quark sector ( dirac - phase ) and affects neutrino oscillations , while the other two are associated to the majorana nature of neutrinos and show up in neutrinoless double beta decay and other lepton - number violating processes , but not in conventional neutrino oscillations @xcite .
current neutrino oscillation experiments are insensitive to cp violation , thus we neglect all phases . in this approximation
three - neutrino oscillations depend on the three mixing parameters @xmath36 and on the two mass - squared differences @xmath37 and @xmath38 characterizing solar and atmospheric neutrinos .
the hierarchy @xmath39 implies that one can set , to a good approximation , @xmath40 in the analysis of atmospheric and k2k data , and @xmath27 to infinity in the analysis of solar and kamland data .
apart from the data already mentioned , the global oscillation analysis also includes the constraints from the chooz and palo verde reactor experiments .
the results of the global three
neutrino analysis are summarized in fig .
[ fig : global ] and in tab .
[ tab : summary ] , taken from ref . @xcite . in the upper panels of the figure the @xmath41
is shown as a function of the parameters @xmath42 , minimized with respect to the undisplayed parameters .
the lower panels show two - dimensional projections of the allowed regions in the five - dimensional parameter space .
the best fit values and the allowed 3@xmath29 ranges of the oscillation parameters from the global data are summarized in tab .
[ tab : summary ] .
this table gives the current status of neutrino oscillation parameters .
three neutrino regions allowed by the world s oscillation data at 90% , 95% , 99% , and 3@xmath29 c.l . for 2 d.o.f
.. in top panels @xmath41 is minimized wrt undisplayed parameters . ] ? = ?
.[tab : summary ] current oscillation parameters . [ cols="<,^,^",options="header " , ] as it has long been noted , in a three neutrino scheme cp violation disappears when two neutrinos become degenerate @xcite or when one angle vanihes , such as @xmath34 @xcite .
all genuine three
flavour effects involve the mass hierarchy parameter @xmath0 and the mixing angle @xmath34 .
determination of @xmath43 and bound on @xmath1 from current neutrino oscillation data.,title="fig:",height=132 ] determination of @xmath43 and bound on @xmath1 from current neutrino oscillation data.,title="fig:",height=132 ] the left panel in fig .
[ fig : alpha ] gives the parameter @xmath44 , namely the ratio of solar over atmospheric splittings , as determined from the global @xmath45 analysis of @xcite .
the right panel in fig .
[ fig : alpha ] gives @xmath46 as a function of @xmath1 for different data samples .
one finds that the kamland-2004 data have a surprisingly strong impact on this bound . before kamland-2004 the bound on @xmath1 from global data was dominated by the chooz reactor experiment , together with the determination of @xmath47 from atmospheric data .
however , including kamland-2004 the bound becomes comparable to the reactor bound .
note also that , since the reactor bound on @xmath1 deteriorates quickly as @xmath27 decreases ( see fig .
[ fig : t13-solar - chooz ] ) , the improvement is especially important for lower @xmath27 values . in fig .
[ fig : t13-solar - chooz ] we show the upper bound on @xmath1 as a function of @xmath27 from chooz data alone compared to the bound from an analysis including solar and reactor neutrino data .
one sees that , although for larger @xmath27 values the bound on @xmath1 is dominated by chooz , for @xmath48 the solar and kamland data become relevant .
altogether , the bound on @xmath1 contributes significantly to the overall global bound 0.047 at 3@xmath29 for 1 d.o.f .
as shown in @xcite such an improved @xmath1 bound follows mainly from the strong spectral distortion found in the 2004 sample .
upper bound on @xmath1 ( 1 d.o.f . ) from solar and reactor data versus @xmath27 . dashed ( solid )
curves correspond to 90% ( 3@xmath29 ) c.l .
bounds , thick curves include kamland-2004 data , thin ones do not .
light ( dark ) regions are excluded by chooz at 90% ( 3@xmath29 ) c.l .
the horizontal line corresponds to the current @xmath27 best fit value , hatched regions are excluded by atmospheric + k2k data at 3@xmath29.,height=151 ] future long baseline reactor and accelerator neutrino oscillation searches @xcite , as well as studies of the day / night effect in large water cerenkov solar neutrino experiments such as uno or hyper - k @xcite could bring more information on @xmath1 @xcite . with neutrino physics entering the precision age
it is necessary to scrutize also the validity of the unitary approximation of the lepton mixing matrix in future experiments , given its theoretical fragility @xcite .
on general grounds neutrino masses are expected to be majorana @xcite , a fact that may explain their relative smallness with respect to other fermion masses .
neutrino oscillation data are insensitive to the absolute scale of neutrino masses and also to the fundamental issue of whether neutrinos are dirac or majorana particles @xcite .
hence the importance of neutrinoless double beta decay @xcite .
the significance of the decay is given by the fact that , in a gauge theory , irrespective of the mechanism that induces , it is bound to also yield a majorana neutrino mass @xcite , as illustrated in fig .
[ fig : bbox ] . quantitative implications of the `` black - box '' argument are model - dependent , but the theorem itself holds in any `` natural '' gauge theory .
now that oscillations are experimentally confirmed we know that must be induced by the exchange of light majorana neutrinos .
the corresponding amplitude is sensitive both to the absolute scale of neutrino mass as well as the two majorana cp phases that characterize the minimal 3-neutrino mixing matrix @xcite .
[ fig : nbbfut ] shows the estimated average mass parameter characterizing the neutrino exchange contribution to versus the lightest neutrino mass .
the calculation takes into account the current neutrino oscillation parameters from @xcite and the nuclear matrix elements of @xcite and compares with experimental sensitivities .
the upper ( lower ) panel corresponds to the cases of normal ( inverted ) neutrino mass spectra . in these plots
the `` diagonals '' correspond to the case of quasi - degenerate neutrinos @xcite , which give the largest amplitude .
in contrast to the normal hierarchy , where a destructive interference of neutrino amplitudes is possible , the inverted neutrino mass hierarchy implies a `` lower '' bound for the amplitude .
an exception to the rule that there is no lower bound on in normal hierarchy models is provided by the model in @xcite .
future experiments @xcite will extend the sensitivity and provide an independent check of the heidelberg - moscow claim @xcite .
more information on the absolute scale of neutrino mass will also come from future beta decays searches @xcite and cosmology @xcite .
neutrino mass arise from the dimension - five operator @xmath49 where @xmath50 the 1 higgs doublet and @xmath51 is a lepton doublet @xcite .
nothing is known from first principles about the mechanism that induces this operator , its associated mass scale or flavour structure .
the neutrino masses that result from it once the electroweak symmetry breaks down are expected to be majorana .
this may explain why neutrino masses are much smaller than those of the other fermions .
this may happen either because the operator is suppressed by a large scale in the denominator , or else suppressed by a small scale in the numerator .
both ways are viable and can be made natural .
the most popular case is the seesaw mechanism which induces small neutrino masses from the exchange of heavy states that may come from unification .
the seesaw mechanism.,title="fig:",height=75 ] the seesaw mechanism.,title="fig:",height=83 ] small neutrino masses are induced either by heavy 1 singlet `` right - handed '' neutrino exchange ( type i ) or heavy scalar bosons exchange ( type ii ) , in a nomenclature opposite from the original one in @xcite .
the effective triplet seesaw term has a flavor structure different from the type - i term , contributing to the lack of predictivity of seesaw schemes .
an attempt to recover predicitvity within the seesaw approach by appealling to extra symmetries @xcite is given in @xcite .
the model predicts maximal @xmath52 , @xmath53 , and naturally large @xmath54 , though unpredicted .
moreover , if cp is violated @xmath34 becomes arbitrary but the dirac cp violation phase is maximal @xcite .
the model gives a lower bound on the absolute neutrino massd @xmath55 ev , requires a light slepton below 200 gev , and gives large rates for flavour violating processes .
amongst `` bottom - up '' models we mention those where neutrino masses are given as radiative corrections @xcite and models where low energy supersymmetry is the origin of neutrino mass @xcite .
the latter are based on the idea that r parity spontaneously break @xcite , leading to a very simple effective bilinear r parity violation model @xcite . in this case
the neutrino spectrum is typically hierarchical , with the atmospheric scale generated at the tree level and the solar scale radiatively `` calculable '' @xcite . for the parameters that reproduce the masses indicated by current data , typically the loghtest supersymmetric particle decays in the detector , and
its decay properties correlate with neutrino mixing angles , a test that can be made , e. g. at the lhc .
the general effective model - independent description of the seesaw at low - energies was given in @xcite .
it is characterized by ( n , m ) , n the number of 1 isodoublet and m the number of 1 isosinglet leptons . in the mass basis
a the ( 3,3 ) seesaw model has 12 mixing angles and 12 cp phases ( both dirac and majorana - type ) characterizing its full 3@xmath566 charged current seesaw lepton mixing matrix and non diagonal neutral current @xcite .
the nontrivial structure of charged and neutral current weak interactions with non - unitary lepton mixing matrix is a general feature of seesaw models @xcite and lead to dimension-6 terms non - standard neutrino interactions ( nsi ) , as illustrated in fig .
[ fig : nunsi ] .
such sub - weak strength @xmath57 operators can be of two types : flavour - changing ( fc ) and non - universal ( nu ) . in inverse
seesaw - type models @xcite the non - unitary piece of the lepton mixing matrix can be sizable and hence the induced nsi may be phenomenologically important @xcite .
sizable nsi strengths may also arise in radiative neutrino mass models and in supersymmetric unified models @xcite .
non - standard physics may in principle affect neutrino propagation properties and detection cross sections @xcite . in their presence , the hamiltonian describing atmospheric neutrino propagation has , in addition to the standard oscillation part , another term @xmath58 @xmath59 here @xmath60 holds for neutrinos ( anti - neutrinos ) and @xmath61 and @xmath62 parameterize the nsi : @xmath63 is the forward scattering amplitude for the fc process @xmath64 and @xmath65 represents the difference between @xmath66 and @xmath67 elastic forward scattering . here
@xmath68 is the number density of the fermion @xmath69 along the neutrino path . in the 2neutrino approximation ,
the determination of atmospheric neutrino parameters @xmath27 and @xmath28 was shown to be practically unaffected by the presence of nsi on down - type quarks ( @xmath70 ) @xcite .
future neutrino factories will substantially improve this bound @xcite .
in contrast , the oscillation interpretation of solar neutrino data is `` fragile '' in the presence of non - standard interactions @xcite , with a new `` dark side '' solution ( with @xmath71 @xcite ) , essentially degenerate with the conventional one , present even after combining with data from reactors . on the other hand , it has been shown @xcite that , even a small residual non - standard interaction of neutrinos in the @xmath72 channel leads to a drastic loss in sensitivity in the @xmath34 determination at a neutrino factory .
it is therefore important to improve the sensitivities on nsi , another window of oportunity for neutrino physics in the precision age .
m. maltoni , t. schwetz , m. a. tortola and j. w. f. valle , new j. phys . 6
( 2004 ) 122 [ hep - ph/0405172 ] .
this reviews the status of global fits to neutrino oscillations with references to the key experimental data .
j. schechter and j. w. f. valle , phys . rev .
* d22 * ( 1980 ) 2227 ; ( 1982 ) 774 j. n. bahcall and m. h. pinsonneault , phys .
* 93 * ( 2004 ) 121301 m. honda , t. kajita , k. kasahara and s. midorikawa , astro - ph/0404457
. s. p. mikheev and a. y. smirnov , sov . j. nucl . phys .
* 42 * ( 1985 ) 913 l. wolfenstein , phys
* d17 * ( 1978 ) 2369 m. maltoni et al , phys . rev . * d68 * ( 2003 ) 113010 ; phys
* d67 * ( 2003 ) 013011 m. c. gonzalez - garcia _ et al .
_ , phys .
lett . * 82 * ( 1999 ) 3202 [ hep - ph/9809531 ] .
j. schechter , j. w. f. valle , phys .
* d23 * ( 1981 ) 1666 m. doi et al , phys . lett . *
b102 * ( 1981 ) 323 ; s. bilenky et al , phys .
b * 94 * ( 1980 ) 49 j. schechter , j. w. f. valle , phys .
* d21 * ( 1980 ) 309 m. lindner and s. schonert , nucl .
* 145 * ( 2005 ) 361 . c. yanagisawa , proc . of ahep2003 , http://ahep.uv.es/
e. k. akhmedov et al , jhep * 0405 * ( 2004 ) 057 ; m. blennow et al , phys
* d69 * ( 2004 ) 073006 l. wolfenstein , phys .
b * 107 * ( 1981 ) 77 ; j. schechter and j. w. f. valle , phys . rev .
* d24 * ( 1981 ) 1883 , err .
d25 ( 1982 ) 283 ; see also talks by l. wolfenstein and j. w. f. valle at joe s fest , aip conf .
* 687 * ( 2003 ) 16 j. schechter , j. w. f. valle , phys
* d25 * ( 1982 ) 2951 s. m. bilenky , a. faessler and f. simkovic , phys .
* d70 * ( 2004 ) 033003 d. caldwell , r. mohapatra , phys .
* d48 * ( 1993 ) 3259 ; a. joshipura , z. phys .
c64 ( 1994 ) 289 ; a. ioannisian and j. w. f. valle , phys .
b * 332 * ( 1994 ) 93 k. babu , e. ma and j. w. f. valle , phys .
b552 * ( 2003 ) 207 ; m. hirsch et al , phys .
d * 69 * ( 2004 ) 093006 m. hirsch , et al , hep - ph/0507148 .
c. e. aalseth _
et al . _ [ igex ] , phys .
d * 65 * ( 2002 ) 092007 [ hep - ex/0202026 ] ; i. abt _ et al .
_ , hep - ex/0404039 ; r. ardito _ et al . _ [ cuore ] , hep - ex/0501010 ; r. gaitskell _ et al .
_ [ majorana ] , nucl - ex/0311013 ; d. akimov _ et al .
_ , [ exo ] , nucl .
* 138 * ( 2005 ) 224 h. klapdor et al , phys .
b * 586 * ( 2004 ) 198 a. osipowicz , et al , hep - ex/0109033 . s. pastor , hep - ph 0404239 ; s. hannestad , hep - ph/0404239 ; g. fogli etal , hep - ph/0408045 s. weinberg , phys .
* d22 * ( 1980 ) 1694 for a review see g. altarelli and f. feruglio , new j. phys .
* 6 * ( 2004 ) 106 [ hep - ph/0405048 ] and references therein .
w. grimus and l. lavoura , phys .
b * 579 * ( 2004 ) 113 a. zee , phys .
lett . * b93 * ( 1980 ) 389 k. s. babu , phys .
b203 * ( 1988 ) 132 m. hirsch and j. w. f. valle , new j. phys . * 6 * ( 2004 ) 76 [ hep - ph/0405015 ] a. masiero and j. w. f. valle , phys .
b * 251 * ( 1990 ) 273 .
m. a. diaz , j. c. romao and j. w. f. valle , nucl .
b * 524 * ( 1998 ) 23 [ hep - ph/9706315 ] .
m. hirsch et al , phys .
d * 62 * ( 2000 ) 113008 [ err .- ibid .
d * 65 * ( 2002 ) 119901 ] [ hep - ph/0004115 ] ; m. a. diaz et al , phys .
d * 68 * ( 2003 ) 013009 [ err - ibid .
d * 71 * ( 2005 ) 059904 ] [ hep - ph/0302021 ] .
r. mohapatra , j. w. f. valle , phys . rev .
* d34 * ( 1986 ) 1642 j. bernabeu et al , phys . lett .
* b187 * ( 1987 ) 303 ; g. c. branco , m. rebelo and j. w. f. valle , phys . lett .
* b225 * ( 1989 ) 385 ; n. rius and j. w. f. valle , phys . lett . *
b246 * ( 1990 ) 249 ; f. deppisch and j. w. f. valle , phys .
d * 72 * ( 2005 ) 036001 [ hep - ph/0406040 ] .
l. j. hall , v. a. kostelecky and s. raby , nucl .
b267 * ( 1986 ) 415 s. pakvasa and j. w. f. valle , proc . of the indian nat .
acad . of sciences ,
70a , no.1 , p.189 - 222 ( 2004 ) , [ hep - ph/0301061 ] n. fornengo et al , phys .
* d65 * ( 2002 ) 013010 p. huber and j. w. f. valle , phys . lett .
* b523 * ( 2001 ) 151 [ hep - ph/0108193 ] . o. g. miranda , m. a. tortola and j. w. f. valle , hep - ph/0406280 .
p. huber , t. schwetz and j. w. f. valle , phys .
* d66 * ( 2002 ) 013006 ; phys .
* 88 * ( 2002 ) 101804
|
i summarize the status of three neutrino oscillations that follow from combining the relevant world s data .
the discussion includes the small parameters @xmath0 and @xmath1 , which characterize the strength of cp violation in neutrino oscillations , the impact of oscillation data on the prospects for probing the absolute scale of neutrino mass in and the robustness of the neutrino oscillation interpretation itself in the presence of non - standard physics .
i also comment on the theoretical origin of neutrino mass , mentioning recent attemps to explain current oscillation data . address= ahep group , instituto de fsica corpuscular , c.s.i.c .
universitat de valncia + edificio de institutos de paterna , apartado 22085 , e46071 valncia , spain +
|
recently , non - orthogonal multiple access ( noma ) @xcite has been considered as a promising technique for fifth generation ( 5 g ) and beyond 5 g ( b5 g ) cellular networks .
the key idea of noma is to simultaneously serve multiple users ( ideally all active users in a serving cell ) over same radio resources at the expense of minimal inter - user interference .
noma not only allows serving individual users with higher effective bandwidth but also allows scheduling more users than the number of available resources .
in contrast to conventional orthogonal multiple access ( oma ) , where every user is served on exclusively allocated radio resources , noma superposes the message signals of multiple users in power domain by exploiting their respective channel gain differences . successive interference cancellation ( sic ) is then applied at the receivers for multi - user detection and decoding . for example , in downlink noma , the base station ( bs ) schedules different users over same resources but their respective message signals are transmitted using different power levels . by exploiting the power differences , each user equipment ( ue ) can apply sic and in turn decode its desired signal .
recently , numerous research activities have been initiated across the globe to identify the potential gains of noma in both the downlink and uplink transmissions . here
we review the most recent and relevant research studies for uplink and downlink noma transmissions .
the basic concept of noma was exploited in @xcite-@xcite for downlink transmissions .
the authors in @xcite-@xcite proposed power domain user multiplexing at the bss and sic - based signal reception at ue terminals . in @xcite ,
the authors discussed various practical challenges of noma systems , such as multi - user power allocation and user scheduling schemes , error propagation in sic , overall system overhead , user mobility , and the combination of noma with multiple - input multiple - output ( mimo ) .
system - level and link - level simulations in @xcite indicated clear benefits of noma over oma in terms of overall system throughput as well as individual user s throughput . in @xcite ,
closed - form expressions for ergodic sum - rate and outage probability were presented for two users considering static power allocations .
the impact of user pairing was studied in @xcite for a two - user noma system .
the authors proposed fixed and opportunistic user pairing schemes by statically allocating transmission powers among noma users . on the other hand ,
the impact of power allocation on the fairness of the downlink noma was investigated in @xcite , considering perfect channel state information ( csi ) feedback as well as average csi feedback . in @xcite ,
a cooperative noma system was studied , where the authors advocated the idea of pairing weak channel users with the strong channel users for cooperative data transmission .
a test - bed for two - user noma system was presented in @xcite .
the experiments were performed by setting @xmath0 mhz bandwidth for noma users .
the results were compared with those for a two - user oma system where each user has a transmission bandwidth of @xmath1 mhz @xcite .
the results showed significance of noma over oma in terms of aggregate as well as individual user s throughput . for uplink transmissions ,
noma was first investigated in @xcite where power control was applied at ue transmitter and minimum mean squared error ( mmse)-based sic decoding was utilized at bs receiver .
a joint subcarrier and power allocation problem was studied in @xcite .
specifically , a sub - optimal solution was designed to maximize the sum - rate of a noma cluster .
closed - form expressions for sum - throughput and outage probability were derived for a two - user uplink noma system in @xcite assuming static powers of different users .
the authors in @xcite also compared their results with tdma - based oma system and concluded that without proper selection of target data rate for each noma user , a user can always be in outage .
this conclusion was also mentioned in @xcite for downlink noma .
apart from these , a robust user scheduling algorithm for uplink noma with sc - fdma was designed in @xcite , where the distinct channel gains of different users were exploited to obtain efficient user grouping . for both uplink and downlink noma systems , efficient user clustering and power allocation among users are the most fundamental design issues . to date ,
most of the research investigations have been conducted either for downlink or for uplink scenario while considering two users in the system with fixed power allocations .
in particular , * there is no comprehensive investigation to precisely analyze the differences in uplink and downlink noma systems and their respective impact on the user grouping and power allocation problems*. in this context , this paper focuses on developing efficient user clustering and power allocation solutions for multi - user uplink and downlink noma systems .
the contributions of this paper are outlined as follows : * we briefly review and describe the differences in the working principles of uplink and downlink noma .
* for both uplink and downlink noma , we formulate cell - throughput maximization problem such that user grouping and power allocations in noma cluster(s ) can be optimized under transmission power constraints , minimum rate requirements of the users , and sic constraints . * due to the combinatorial nature of the formulated mixed integer non - linear programming ( minlp ) problem , we propose a low - complexity sub - optimal user grouping scheme .
the proposed scheme exploits the channel gain differences among users in a noma cluster and group them either in a single cluster or multiple clusters to enhance the sum throughput of the uplink and downlink noma systems .
* for a given set of noma clusters , we derive optimal power allocation that maximizes the sum throughput of all users in a cluster and in turn maximizes the overall system throughput . using kkt optimality conditions , for both uplink and downlink noma
, we derive closed - form optimal power allocations for any cluster size . *
we evaluate the performances of different uplink and downlink noma systems using the proposed user grouping and power allocation solutions .
numerical results compare the performance of noma over oma and illustrate the significance of noma in various network scenarios .
important guidelines related to the key design factors of noma systems are obtained .
the rest of the paper is organized as follows : section ii discusses the fundamentals of downlink and uplink noma systems .
iii presents the system model , assumptions , and the joint problem formulation for optimal user clustering and power allocation in noma systems .
section iv and section v , respectively , discuss the proposed sub - optimal user clustering solution and the optimal power allocation solutions for uplink and downlink noma systems .
section vi evaluates the performance of the proposed solutions numerically and section vii concludes the article .
in this section , we discuss the basic concepts of downlink and uplink noma considering @xmath2 users with distinct channel gains in a cluster .
the power domain multiplexing is applied to superpose multiple signals , whereas an sic mechanism is applied at the receiver(s ) to decode the superposed signals .
let us consider a general @xmath2-user downlink noma system , where all users experience distinct channel gains . in @xmath2-user downlink noma , a single transmitter ( i.e. , bs ) non - orthogonally transmits @xmath2 different signals over the same radio resources ; whereas , all @xmath2 receivers ( i.e. , ues ) receive their desired signals along with the interferences caused by the messages of other ues . to obtain the desired signal ,
each sic receiver first decodes the dominant interferences and then subtracts them from the superposed signal
. therefore , the received signal strength of the interference signals needs to be sufficiently higher in comparison to the desired signal in order to cancel them by sic processing at the receiver end .
since each ue receives all signals ( desired and interfering signals ) over the same radio spectrum ( i.e. , channel ) , multiplexing of different signals with different power levels is crucial to diversify each signal and to perform sic at a given ue receiver . in downlink noma
, the messages of high channel gain users are transmitted with low power levels whereas the messages of low channel gain users are transmitted with high power levels . as such , at a given receiver , the strong interfering signals are mainly due to the information of low channel gain users .
the weakest channel user ( who receives low interferences due to relatively low powers of the messages of high channel gain users ) can not suppress any interferences .
however , the highest channel gain user ( who receives strong interferences due to relatively high powers of the messages of low channel gain users ) can suppress all interfering signals .
-user downlink noma cluster with sic at bs.,width=3 ] * illustration : * fig .
[ fig : alpha3 ] illustrates a @xmath3-user downlink noma system where @xmath4 , @xmath5 , and @xmath6 are the channel gains of @xmath7 , @xmath8 , and @xmath9 , respectively .
also , it is assumed that @xmath10 , @xmath11 , and @xmath12 are the desired messages of @xmath7 , @xmath8 , and @xmath9 , respectively , while @xmath13 , @xmath14 , and @xmath15 denote the respective additive white gaussian noise ( awgn ) .
if @xmath16 , then @xmath7 can perform sic to cancel interference from both @xmath8 and @xmath9 , whereas @xmath8 can only cancel interference from @xmath9 .
also , @xmath9 experiences interference from both @xmath7 and @xmath8 , but can not cancel any of them . therefore , the achievable throughput for @xmath17 , in a @xmath3-user downlink noma cluster can be expressed as @xmath18 where @xmath19 is the normalized channel gain with @xmath20 being the noise power , @xmath21 is the number of radio channels ( e.g. , resource blocks in an lte / lte - a system ) assigned for the cluster , @xmath22 is the transmission bandwidth of each channel ( e.g. , resource block ) . to perform sic , transmission power for each noma
user needs to be selected properly .
if @xmath23 , @xmath24 , and @xmath25 are the transmission powers for @xmath7 , @xmath8 , and @xmath9 , respectively , then the power allocations need to satisfy the following conditions for efficient sic at @xmath7 receiver : @xmath26 where @xmath27 , @xmath28 is the downlink power budget for this @xmath3-user noma cluster , and @xmath29 is the minimum power difference needed to distinguish between the signal to be decoded and the remaining non - decoded message signals . and represent the power allocation conditions to cancel interference of @xmath9 and @xmath8 , respectively , at @xmath7 receiver . from and , it is evident that the transmit power for any user must be the greater than sum transmit power for all users with relatively stronger channel gains .
that is , the transmit power for @xmath9 must be greater than the sum transmit power for @xmath7 and @xmath8 , while the transmit power for @xmath8 needs to be greater the transmit power for @xmath7 .
subsequently , the power allocation condition to cancel the interference of @xmath9 at @xmath8 receiver can be given as @xmath30 note that , since @xmath31 , is automatically satisfied if holds .
therefore , the necessary power constraints for efficient sic in a @xmath3-user cluster can be given by and .
based on the above illustration , the necessary power constraints for efficient sic in an @xmath2-user noma cluster can be expressed as follows : @xmath32 consequently , an important conclusion about the transmit power of highest channel gain user in a noma cluster can be derived as given in the following .
the maximum transmission power allocation to the highest channel gain user in the downlink noma cluster must be smaller than @xmath33 , where @xmath2 is the number of users in the cluster and @xmath28 is the total transmission power budget for the given noma cluster .
the proof follows by induction .
let us consider a @xmath34-user downlink noma cluster where @xmath35 and @xmath36 are the normalized channel gains of the high and low channel gain users , respectively .
as per the sic constraints in , we have @xmath37 where @xmath23 and @xmath24 are the allocated powers for high and low channel users , respectively .
the maximum allocated power to the highest channel gain user can thus be derived as @xmath38 where @xmath39 is the minimum power difference needed for sic . note that the value of @xmath40 can be very small when the value of @xmath35 is very high , which is usually the case .
similarly , for @xmath3-user downlink noma cluster , the maximum allocated transmit powers for second higher and highest channel gain user can be expressed , respectively , as @xmath41 consequently , for an @xmath2-user cluster , we have @xmath42 this is the same result as given in * lemma 1*. the operation of uplink noma is quite different from that of downlink noma . in uplink noma , multiple transmitters ( ues )
non - orthogonally transmit to a single receiver ( bs ) on the same radio spectrum ( i.e. , channel ) .
each ue independently transmits its own signal at either maximum transmit power or controlled transmit power .
all received signals at the bs are the desired signals , although they cause interference to each other .
since the transmitters are different , each received signal at sic receiver ( i.e. , the bs ) experiences distinct channel gain .
note that , to apply sic and decode signals at bs , we need to maintain the distinctness among various message signals .
as such , conventional transmit power control ( typically intended to equalize the received signal powers of all users ) may not be feasible in noma - based systems .
let us consider a general @xmath2-user uplink noma system in which @xmath2 users transmit to a common bs over the same radio channel , at either maximum transmit power or controlled transmit power .
the bs receives the superposed message signal of @xmath2 different users and applies sic to decode each signal .
since the received signal from the highest channel gain user is likely the strongest at the bs ; therefore , this signal is decoded first . consequently ,
the highest channel gain user experiences interference from all other users in the cluster .
then , the signal for second highest channel gain user is decoded and so on . as a result , the highest channel gain user experiences interference from all users and
the lowest channel gain user enjoys interference - free data rate .
-user uplink noma cluster with sic at ues.,width=3 ] * illustration : * fig .
[ fig : alpha3-up ] illustrates a @xmath3-user uplink noma cluster in which @xmath7 , @xmath8 , and @xmath9 experience channel gains of @xmath4 , @xmath5 , and @xmath6 , respectively , where @xmath16 . in uplink noma
, the user s signal with the highest channel gain is decoded first at the bs .
thus , the achievable data rate of @xmath7 depends on the interference from @xmath8 and @xmath9 , whereas @xmath9 achieves interference free data rate .
the data rate of @xmath8 depends on the interference from @xmath9 .
consequently , the achievable data rate ( or throughput ) for @xmath17 , in a @xmath3-user uplink noma cluster can be expressed as @xmath43 where @xmath44 and @xmath22 are defined similarly as in downlink noma .
if @xmath23 , @xmath24 , and @xmath25 are the transmission powers of @xmath7 , @xmath8 , and @xmath9 , respectively , then the following conditions need to be satisfied for efficient sic at bs , i.e. , @xmath45 where @xmath46 and @xmath47 is the maximum transmit power budget of each ue . and represent the necessary conditions for efficient decoding of @xmath7 and @xmath8 signals , respectively , prior to decoding the signal of @xmath9 . based on the above example , the necessary power constraints for efficient sic , in an @xmath2-user uplink noma cluster can be expressed as follows : @xmath48
we consider a macro base station ( bs ) serving @xmath49 uniformly distributed ues for uplink as well as downlink .
the bs and each of the ues use a single antenna configuration .
the available system bandwidth @xmath50 is divided into frequency resource blocks , each of bandwidth @xmath22 .
that is , the total number of frequency resource blocks are given as @xmath51 .
users who are non - orthogonally scheduled over the same resource blocks form a noma cluster
. however , each noma cluster operates on orthogonal frequency resource blocks . the number of users per noma cluster is represented by @xmath2 which ranges from @xmath52 .
also , the resource blocks allocated per cluster are represented by @xmath21 , where @xmath53 .
provided the range of @xmath2 , the number of clusters can vary between 1 and @xmath54 .
the maximum bs transmission power budget is @xmath55 , the maximum transmission power budget per downlink noma cluster is @xmath28 , and the maximum user transmit power is @xmath47 . the normalized channel gain between @xmath56-th ue and the bs is represented by @xmath57 which accounts for both distance - based path - loss and shadowing .
the users are sorted according to the descending order of their normalized channel gains as @xmath58 .
now , let us define a variable @xmath59 as follows : @xmath60 where @xmath61 and @xmath62 .
the joint user clustering ( i.e. , grouping of users into clusters ) and power allocation problem for the throughput maximization in downlink noma can be formulated as @xmath63 where @xmath64 is the minimum data rate requirement for @xmath56-th user .
the * constraint @xmath65 * denotes the total power constraint of the bs , * constraint @xmath66 * ensures the minimum downlink data rate requirements of the users , * constraint @xmath67 * denotes the sic constraints as discussed in section ii.a , * constraint @xmath68 * ensures that one user can be assigned to at most one cluster , while at least two users are grouped into each downlink noma cluster , * constraint @xmath69 * provides the total downlink frequency resource constraint .
in addition , * constraint @xmath70 * and @xmath71 demonstrate that @xmath21 and @xmath72 are integer variables .
similarly , the joint user clustering and power allocation problem for throughput maximization of an uplink noma system can be formulated as follows : @xmath73 @xmath74 where @xmath75 is the minimum uplink data rate requirement for @xmath56-th user .
* constraint @xmath76 * ensures the minimum rate requirements of the users , * constraint @xmath77 * ensures the minimum uplink data rate requirements of the users , * constraint @xmath78 * ensures the sic constraint as discussed in section ii.b , and * constraints @xmath79 * are same as defined in section iii.b .
as can be seen , the formulated problems are mixed integer non - linear programming ( minlp ) problems whose solution is combinatorial by nature .
specifically , for throughput maximization , the optimal user clustering solution requires an exhaustive search to form a noma cluster @xcite .
that is , for every single user , we need to consider all possible combinations of user grouping .
for example , let us consider an uplink / downlink noma system with @xmath49 users .
in such a system , the number of possible combinations for optimal user clustering , can be expressed as follows : @xmath80 evidently , the computational complexity of optimal user clustering may not be affordable for practical systems with a large number of active users . as such , we resort to solve the problem in two steps , i.e. , by developing a less complex solution for grouping users into different noma clusters and then optimizing their respective powers to maximize the sum throughput per cluster .
subsequently , section iv details the proposed low - complexity user clustering scheme . given the user clustering ,
we derive optimal power allocations for users in section v.
in this section , we propose a low - complexity sub - optimal user clustering scheme for both uplink and downlink noma systems .
the proposed scheme exploits the channel gain differences among users and aims at enhancing the sum - throughput of the considered cell .
prior to user grouping , this scheme relies on selecting a feasible number of clusters , i.e. , decides the number of clusters and in turn the number of users per cluster .
once the number of users in a cluster is decided , user grouping is performed . in the following
, the key concepts of our proposed user paring policies and their algorithmic presentation are detailed .
let us consider an @xmath2-user downlink noma cluster to which @xmath21 units of resource blocks are allocated .
for such a system , the achievable per - user throughput in ( [ ugdl2 ] ) , where @xmath81 , provides necessary insights to group users into a cluster .
these insights are discussed below : * after sic , the throughput of the highest channel gain user in a cluster is not subject to the intra - cluster interference ; instead , its throughput depends on its own channel gain and power .
although the allocated transmit power for the highest channel gain user is low ( as also mentioned in section ii.a ) , its impact on the throughput is minimal .
subsequently , if the gain of the highest channel is sufficiently high , then the achievable data rate negligibly depends on the transmission power , unless the power is very low .
thus it is beneficial to distribute the high channel gain users in a cell into different noma clusters , as they can significantly contribute to the sum - throughput of a cluster . * to increase the throughput of the users with low channel gains , it is useful to pair them with the high channel gain users .
the reason is that the high channel gain users can achieve a higher rate even with the low power levels while making the large fraction of power available for weak channel users . as such
, the key point of our proposed user clustering in downlink noma is to pair the highest channel gain user and the lowest channel gain user into same noma cluster , while the second highest channel gain user and the second lowest channel gain user into another noma cluster , and so on . *
the throughput of the remaining users in a noma cluster follows the same format .
that is , the sinr contains same channel in both the denominator and numerator , whereas the transmit power in the numerator is greater than the sum power in the denominator ( this is given by the sic constraint ) . as such , the throughput of the remaining users in a noma cluster depends mainly on the distribution of the transmit power levels .
let us consider an @xmath2-user uplink noma cluster to which @xmath21 units of resources are allocated .
for such a system , the achievable per - user throughput in ( [ ugul1 ] ) , where @xmath81 , provides necessary insights to group users .
these insights are discussed below : * in an uplink noma cluster , all users signals experience distinct channel gains . to perform sic at bs , we need to maintain the distinctness of received signals .
as such , the conventional transmit power control may not be feasible in a noma cluster .
further , contrary to downlink noma , the power control at any user does nt increase the power budget for any other user in a cluster .
as such , the ultimate result of power control is sum - throughput degradation . *
the distinctness among the channels of different users within a noma cluster is crucial to minimize inter - user interference and thus to maximize the cluster throughput . * in uplink noma
, the highest channel gain user does not interfere to weak channel users ( actually his interference is canceled by sic ) .
therefore , this user can transmit with maximum power to achieve a higher throughput .
it is thus beneficial to include the high channel gain users transmitting with maximum powers in each noma cluster , as they can significantly contribute to the throughput of a cluster .
based on the above discussions , let we classify the users into two classes : class - a and class - b .
the number of users in class - a , denoted as @xmath82 , have much higher channel gains compared to users in class - b , i.e. , @xmath83 then , the proposed sub - optimal user clustering algorithm can be given as in * algorithm 1 * below
. + ' '' '' ' '' '' 1 .
* sort users : * @xmath84 .
2 . * select no . of clusters
: * @xmath85 3 . * ( a ) group users into clusters for downlink noma : * + @xmath86st cluster @xmath87 , + @xmath34nd cluster @xmath88 , + @xmath89-th cluster @xmath90 . + * ( b ) group users into clusters for uplink noma : * + @xmath86st cluster @xmath91 + @xmath34nd cluster @xmath92 + @xmath89-th cluster @xmath93 .
* cluster size : * @xmath94 ' '' '' to illustrate , in fig . 3 and fig . 4
, we show user grouping for @xmath34-user , @xmath3-user , and @xmath95-user noma clusters in downlink and uplink , respectively , where the total number of users are taken as 12 .
-user , @xmath3-user , and @xmath95-user noma clustering for downlink transmission to 12 active users in a cell.,width=3 ] -user , @xmath3-user , and @xmath95-user noma for uplink transmission of 12 active users in a cell.,width=3 ]
given the noma clusters obtained from section iv , in this section , we derive optimal power allocations for a noma cluster with @xmath2 users in the both downlink and uplink transmission scenarios , where @xmath96 .
closed - form expressions for the optimal power allocations are derived using kkt optimality conditions .
let us consider an @xmath2-user downlink noma cluster , where the normalized channel gains of @xmath97 are assumed as @xmath98 , respectively , and their respective minimum rate requirements are @xmath99 , where @xmath100 .
it is assumed that @xmath21 resource blocks , each of bandwidth @xmath22 , are allocated to the downlink noma cluster .
if @xmath101 are the transmission powers for @xmath102 , respectively , then the optimal power allocation problem can be given as @xmath103 where @xmath104 is the inter - user interference for @xmath56-th user in downlink noma cluster .
* constraint @xmath105 * is the total power constraint , * constraint @xmath106 * is the minimum rate requirement per user , and * constraint @xmath107 * denotes the sic constraints .
note that the aforementioned problem is convex under * constraints @xmath108*. for the aforementioned problem , the lagrangian can be expressed as : @xmath109 where @xmath110 , @xmath111 , and @xmath112 are the lagrange multipliers , @xmath113 and @xmath114 .
also , @xmath115 .
taking derivatives of ( [ opd1 ] ) w.r.t .
@xmath116 , @xmath110 , @xmath111 , @xmath112 , we can write karush - kuhn - tucker ( kkt ) conditions as follows : @xmath117 @xmath118 @xmath119 @xmath120 @xmath121 in addition , we have several kkt complementarity conditions whose treatment is detailed in the following . in an @xmath2-user cluster , there are @xmath122 lagrange multipliers .
thus there are @xmath123 combinations of lagrange multipliers that need to be checked for satisfying the kkt conditions @xcite . however
, checking @xmath123 combinations is computationally complex .
for example , if @xmath124 , then the number of combinations are @xmath125 , respectively . in our problem @xmath126 ; therefore , to obtain a fixed number of solutions for @xmath2 decision variables we need exactly @xmath2 equations @xcite .
thus , all @xmath123 combinations need not to be checked , rather we need to check only @xmath127 combinations . after solving for @xmath34- , @xmath3- , @xmath95- , and @xmath128-user noma clusters , we find that the lagrange multiplier combinations satisfying kkt conditions are @xmath129 , respectively , thus in general @xmath130 .
the lagrange multipliers for @xmath2-user downlink noma cluster belong to three sets of constraints .
these sets are the total transmit power constraints , minimum data rate constraints , and sic constraints , given mathematically as , @xmath131 , respectively .
therefore , the solution set is , @xmath132 .
for example , for @xmath34-user downlink noma , the satisfied kkt conditions are @xmath133 and @xmath134 . for @xmath3-user noma ,
the satisfied kkt conditions are @xmath135 , @xmath136 , @xmath137 , and @xmath138 .
now let us define two additional sets of lagrange multipliers , @xmath139 and @xmath140 .
then the closed - form solution of optimal power allocation to @xmath2-user downlink noma cluster can be given as in the following .
the closed - form solution of the optimal power allocation for the highest channel gain user in downlink noma cluster can be given as follows : @xmath141 on the other hand , the optimal power allocations for remaining users ( except the highest channel gain user ) are given as @xmath142\times ( \varphi_i - 1).\end{aligned}\ ] ] @xmath143 see * appendix a*. the optimal transmission powers and the corresponding necessary conditions for @xmath34- , @xmath3- , and @xmath95-user downlink noma clusters are provided in * table [ dlcs]*. [ cols="^ " , ] in this subsection , we compare the performance of noma with ofdma in terms of sum - throughput and individual users throughput .
further , we compare the overall throughput performance of @xmath34-user , @xmath3-user , and @xmath95-user downlink noma systems by considering @xmath144 active downlink users
. + -user downlink noma and oma systems assuming 100 kbps minimum data rate .
normalized channel gain of @xmath7 is 40 db.,width=3 ] -user downlink noma and oma systems assuming 1 mbps minimum data rate .
normalized channel gain of @xmath7 is 40 db.,width=3 ] -user downlink noma and oma systems .
initial normalized channel gains of @xmath7 , @xmath8 , @xmath9 , and @xmath145 are 40 db , 32 db , 24 db and 16 db , respectively .
we assume @xmath146 100 kbps and @xmath146 500 kbps , where @xmath147 .
, width=3 ] fig .
[ fig:1 ] and fig .
[ fig:2 ] show the sum - throughput and individual throughput of @xmath34-user downlink noma cluster as well as its corresponding oma system for minimum data rate requirements of @xmath148 kbps and @xmath86 mbps , respectively .
the channel gain of the higher channel gain user is fixed at 40 db whereas the channel gain of weak user varies .
further , fig .
[ fig:4 ] represents the sum - throughput of @xmath95-user downlink noma cluster and its corresponding oma system as a function of each user channel variations . in fig .
[ fig:4 ] , the initial channel gains of @xmath7 , @xmath8 , @xmath9 , and @xmath145 are set as @xmath149 db , @xmath150 db , @xmath151 db , and @xmath152 db , respectively , and in each sub - figure only one user s channel is varied by ensuring @xmath153 . from these simulation results we have the following observations : * sum - throughput of downlink noma is always better than oma at any channel conditions .
however , a significant throughput gain can be achieved for more distinct channel conditions of users in a cluster . *
individual throughput of the highest channel gain user in a noma cluster is significantly higher than that in oma .
however , the lowest channel gain user s throughput is limited by its minimum rate requirement . to address this issue in noma
, the minimum rate requirements of different users can be dynamically adjusted to enhance the fairness among users .
* sum - throughput of downlink noma strongly depends on the highest channel gain user within a cluster .
the reason behind is that the strongest channel gain user can cancel all interfering signals before decoding its own signal , thus its achievable data rate does not depend on inter - user interference .
* the impact of the lowest channel gain user is minimal on the cluster sum - throughput , unless the channel gain is so small that a huge power is required by the lowest channel gain user . at this point , a sharp decay of sum - throughput is observed .
note that the traditional oma is unable to operate at such poor channel gains . *
the channel variations of all users , except the highest and lowest channel gain users , in a downlink noma cluster do not considerably affect the sum - throughput of a noma cluster ( see fig .
[ fig:4](b ) and fig .
[ fig:4](c ) ) . as given in * lemma 1 * , the higher number of users in a downlink noma cluster significantly reduces the amount of available power for the strongest channel user , who generally contributes maximum throughput in a cluster .
therefore , it is crucial to select the correct cluster size .
however , the throughput performance of a noma cluster depends significantly on three parameters , i.e. , cluster size , transmit power , and channel gains of users . for a particular set of transmit powers and channel gains of users
, we can find a cluster size that generally maximizes the sum throughput .
* table [ dl - compr ] * represents the sum - throughput of different noma and oma systems with @xmath144 downlink users for various channel conditions while the transmit power is fixed in all cases . in * table [ dl - compr ] * , we order the users according to their channel gains in descending order .
there are 6 , 4 , and 3 clusters with @xmath34 , @xmath3 , and @xmath95 users in a cluster , respectively .
the throughput performances of the aforementioned downlink noma clusters and their respective oma counterparts are demonstrated in * table [ dl - compr]*. the main observations are as follows : * * _ less distinct channel gains of users _ ( cases 12 and 14 ) _ _ * : in this case , the throughput of different downlink noma systems is nearly the same .
however , the @xmath95-user noma achieves a higher throughput at better channel gains _ ( case 12 ) _ , while the @xmath34-user noma obtains higher throughput at lower channel gains _
( case 14)_. as such , we can conclude that higher cluster size is preferred for higher channel gains of the users and lower cluster size should be preferred for lower channel gains of the users .
the overall throughput gains of downlink noma over oma are very limited . * * _ more distinct channel gains of users ( case 13 ) : _ * in this case , noma systems outperform their oma counterparts .
it can be seen that the @xmath95-user noma achieves a better throughput than @xmath34-user and @xmath3-user noma systems .
therefore , higher cluster size can be selected in such cases as long as the power allocation to the highest channel gain does not decrease significantly * ( lemma 1)*. * * _ number of higher channel gain users equals to the number of clusters _ ( cases 3 , 4 , 6 ) : _ _ * in such a case , each downlink noma system achieves maximum relative throughput gain compared to oma . in * table [ dl - compr ] * , the @xmath95-user downlink noma system ( i.e. , @xmath3 clusters ) achieves maximum 106.3% throughput gain over oma system when the number of good channels are equal to @xmath3 ( _ case 3 _ ) .
however , the @xmath3-user ( i.e. , @xmath95 clusters ) and @xmath34-user ( i.e. , @xmath128 clusters ) systems achieve maximum throughput gains of 86.6% and 52.8% , respectively , over oma systems ( _ case 4 _ and _ case 6 _ ) .
thus , a noma system with the number of higher channel gain users equal to the number of clusters either achieves maximum or close to maximum throughput among all noma systems . * in general , the throughputs of all noma systems are quite similar when 50% or more users experience good channels ( _ cases 6 to 12 _ ) . however ,
if the higher channel gain users become limited , then the higher cluster sizes should be selected to enhance throughput ( _ cases 1 to 3 _ ) .
-user downlink noma system , for @xmath154 100 kbps and 2 mbps .
normalized channel gains of @xmath7 and @xmath8 are 40 db and 30 db , respectively.,width=3 ] finally , fig .
[ fig : f ] shows the individual users throughput for @xmath3-user downlink noma cluster considering two different minimum rate requirements of users , i.e. , 100 kbps and 2 mbps . in general
, our optimal power allocations maximize the transmit power of the highest channel gain user while maintaining the minimum rate requirements of all users in a noma cluster .
however , in such a case , the lower channel gain users may experience significant throughput difference when compared to the highest channel gain user , as discussed in section vi.a.1 . as such , to improve the throughput fairness among users , the minimum data rate requirements of the users can be adjusted further to balance the trade - off between fairness and the overall system throughput . in this subsection , we compare the throughput performances of uplink noma and oma systems .
we also compare the overall and individual user s throughput of @xmath34-user , @xmath3-user , @xmath95-user , and @xmath128-user uplink noma systems for @xmath144 uplink users .
+ the sum - throughput and individual throughput of @xmath34-user uplink noma cluster is shown in fig .
[ fig:6 ] , where the minimum user rate requirement is @xmath148 kbps .
it can be seen that the sum throughput of noma outperforms oma with more distinct channel gains of users in a cluster . also , the noma sum - throughput remains higher than oma , regardless of the weakest channel . on the other hand ,
when @xmath7 and @xmath8 have nearly the same channel gains , @xmath8 achieves better data rate .
however , due to high interference of @xmath8 , @xmath7 obtains very low data rate . as the channel gain of @xmath8 reduces , its throughput gradually reduces
however , the interference on @xmath7 also reduces which improves the achievable throughput of @xmath7 . as such
, the sum throughput of uplink noma cluster remains almost unchanged .
when the minimum rate requirement of both users is set as high as @xmath86 mbps , fig .
[ fig:7 ] depicts the sum - throughput and individual throughput of @xmath34-user noma cluster as well as the sum throughput of @xmath34-user oma system .
it is observed that the sum - throughput of noma becomes less than the corresponding oma system in the region when the channel gains are less distinct and power control is applied .
the reason is that , without power control , the sum data rate of @xmath34-user uplink noma system and corresponding oma system is nearly similar . therefore , after applying power control in noma , sum data rate is further reduced and goes below oma .
we further note that , since the power control is only applicable at the weakest channel gain user , its impact keeps diminishing gradually for @xmath3-user and beyond uplink noma clusters ( see in fig . [ fig:7new ] ) .
[ fig:7new ] shows the sum - throughput and individual throughput of @xmath3-user uplink noma cluster and corresponding sum throughput of oma system , where the minimum individual rate requirement is @xmath86 mbps .
it is evident that the sum - throughputs of @xmath3-user and beyond uplink noma clusters remain always better than those of the oma systems .
-user uplink noma and oma systems assuming 100 kbps minimum data rate
. normalized channel gain of @xmath7 is 40 db.,width=3 ] -user uplink noma and oma systems assuming 1 mbps minimum data rate .
normalized channel gain of @xmath7 is 40 db.,width=3 ] -user uplink noma and oma systems assuming 1 mbps minimum data rate .
normalized channel gains of @xmath7 and @xmath8 are 40 db and 34 db , respectively.,width=3 ] fig . [ fig:8 ] demonstrates the sum - throughput of @xmath95-user uplink noma cluster as a function of individual user s channel gain variation .
the normalized channel gains of @xmath7 , @xmath8 , @xmath9 , and @xmath145 are set at 40 db , 32 db , 24 db and 16 db , respectively . in each sub - figure , the channel gain of each user is varied while others remain fixed .
for example , in case ( a ) , the normalized channel gain of @xmath7 is varied from 40 db to 32.5 db while that channel gains of other users are fixed .
it is observed that the sum - throughput of uplink noma mainly depends on the channel conditions of the highest channel gain user .
-user uplink noma and oma systems .
initial normalized channel gains of @xmath7 , @xmath8 , @xmath9 , and @xmath145 are 40 db , 32 db , 24 db , and 16 db , respectively .
we assume @xmath155 100 kbps and @xmath155 500 kbps , where @xmath156.,width=3 ] -user uplink noma cluster for 100 kbps minimum data rate .
normalized channel gains of @xmath7 to @xmath157 are 40 , 35 , 30 , 26 , and 22 db , respectively , while the normalized channel gain of @xmath158 varies.,width=3 ] the important feature of uplink noma system is that all lower channel gain users in a noma cluster interfere significantly to the higher channel gain users . note that , due to sic , low channel gain users do not experience any interference from high channel gain users . in contrast to downlink noma , the impact of transmission power control in not significant in uplink noma . in uplink noma , if more users need to be grouped into a cluster , more distinct channels are required . with different channel conditions of @xmath144 uplink active users
, we measure the performances of @xmath34-user , @xmath3-user , @xmath95-user , and @xmath128-user uplink noma systems as well as oma system , shown in * table [ ul - compr]*. the users are clustered according to the method discussed in section iv.c . the channel gains in table * [ ul - compr ] * are chosen in order to ensure the minimum channel distinctness required for @xmath128-user uplink noma system . in table * [ ul - compr ] * , we sort 12 users according to the descending order of their channel gains .
there are 6 , 4 , 3 , and 2 clusters available for @xmath34-user , @xmath3-user , @xmath95-user , and @xmath128-user uplink noma systems , respectively . from table
[ ul - compr ] , we have the following key observations : * * _ more distinct channel gains of users ( case 13 ) : _ * in this case , the performance of any uplink noma is much better than that of the oma system ( _ case 13 _ ) , whereas the higher order noma achieves higher throughput . * * _ less distinct channel gains of users _ ( cases 12 and 14 ) _ _ * : in this case , higher order uplink noma still performs better , although the throughput performance of various noma systems are quite similar .
however , in this case , the throughput gains of noma over oma are marginal . * * _ number of good channel users equals the number of clusters _ ( cases 2 , 3 , 4 , 6 ) : _ _ * in such a case , each uplink noma system achieves maximum relative throughput gain compared to oma . in * table [ ul - compr ] * , _ case 2 _ shows 105.4% throughput gain of @xmath128-user uplink noma system in comparison to oma system , while _ case 3 , 4 _ and _ case 6 _ show the maximum throughput gain of @xmath95-user , @xmath3-user , and @xmath34-user uplink noma systems , respectively , where the gains are 79.9% , 60.25% , and 33.3% , respectively .
thus , a noma system with the number of higher channel gain users equal to the number of clusters either achieves maximum or close to maximum throughput among all noma systems . * when one or couple of users experience higher channel gains , higher order uplink noma systems perform much better than oma systems _ ( case 1 , 2 and 3)_. as the number of users experiencing higher channel gains increases , different uplink noma systems perform nearly similar _ ( case 4 to 12)_. fig .
[ fig:8new ] shows the individual throughput of @xmath128-user uplink noma cluster , where @xmath7 to @xmath157 experience 40 , 35 , 30 , 26 , and 22 db channel gains , respectively .
[ fig:8new ] shows good throughput fairness among @xmath7 to @xmath145 , whilst , by selecting proper channel gains for @xmath157 and @xmath158 , all six users can get good throughput fairness .
therefore , exploitation of channel gain differences among the noma users is the key to designing efficient uplink noma systems .
efficient user clustering and power allocation among noma users are the key design issues for successful noma operations . in this paper , for both uplink and downlink noma , we have formulated a joint optimization problem for sum throughput maximization under the constraints of transmission power budget , minimum rate requirements of users , and sic receiver s operation constraints . due to the combinatorial nature of the problem
, we have developed a low - complexity sub - optimal user clustering scheme .
given the user clustering , we have derived closed - form optimal power allocations for @xmath2-user uplink / downlink noma systems . in both of downlink and uplink noma , user clusters with more distinctive channel gains
provides impressive throughput gain of noma over their counterpart oma systems .
numerical results show that the performance of downlink noma decreases if the cluster size increases beyond a certain threshold . despite the numerous benefits of noma over oma , the issues such as sic error processing and inter - cell interference , are still under investigation . in this paper , we have considered ideal sic ; however , the performance of noma may depend significantly on the sic errors .
in downlink noma , each signal needs to be identically encoded , modulated , and precoded at the bs while the sic receiver needs to successively demodulate , decode , and re - modulate stronger signals .
therefore , for large noma clusters , error propagation in sic may drastically reduce the noma performance . further
, inter - cell interference is another major challenge for practical noma applications .
the power allocation solutions need to be carefully extended for dense cellular networks in which inter - cell interference can be significant .
in this appendix , we show the verification of kkt conditions for a given lagrange multiplier combination in a general @xmath2-user downlink noma cluster considering a particular value of @xmath2 .
let us consider @xmath159 which is a @xmath95-user downlink noma cluster .
then , the set of lagrange multipliers can be expressed as @xmath160 .
the solution sets that satisfy kkt conditions are , @xmath161 .
now , consider one set of lagrange multipliers , say @xmath162 , that satisfies the kkt conditions , and thus the values of the lagrange multipliers are @xmath163 , @xmath164 , @xmath165 , @xmath166 , and @xmath167 .
for the aforementioned lagrange multipliers to ( [ opd6 ] ) can be expressed as @xmath168 @xmath169 @xmath170 @xmath171 equations ( [ eqn : a1])-([eqn : a2 ] ) provide the optimal solution of @xmath172 and @xmath173 , while equations ( [ eqn : a3])-([eqn : a4 ] ) provide the necessary conditions of these optimal solutions .
now , solving the equations ( [ eqn : a1])-([eqn : a2 ] ) , we obtain the optimal power allocations as follows : @xmath174 @xmath175 @xmath176 @xmath177 since @xmath178 , the solutions of @xmath179 and @xmath173 all are positive .
now , with the considered lagrange multipliers and positive @xmath180 , and after some algebraic operations , equations ( [ opd2 ] ) to ( [ opd3 ] ) can be expressed as @xmath181 @xmath182 after performing some algebraic operations into equations ( [ eqn : a5])-([eqn : a6 ] ) , we obtain the lagrange multipliers as follows : @xmath183 @xmath184 @xmath185 in our proposed dynamic power allocation solutions , we sort the ues according to the descending order of their normalized channel gains , i.e @xmath186 .
therefore , the solutions for @xmath110 and @xmath187 , of equation ( [ eqn : a7])-([eqn : a9 ] ) all are positives .
therefore , the set of lagrange multipliers @xmath162 satisfies the kkt conditions .
all the other cases can easily be verified by using a similar approach .
similar to downlink noma cluster , we show the kkt verification of a general @xmath2-user uplink noma cluster by considering a particular combination of lagrange multipliers in a particular cluster size .
let again consider @xmath159 , i.e. , a @xmath95-user uplink noma cluster .
then , the sets of lagrange multipliers can be expressed as @xmath188 .
therefore , the solution sets that satisfy the kkt conditions are : @xmath189 .
now , consider one set of lagrange multipliers , say , @xmath190 , satisfies the kkt conditions , and thus the value of the lagrange multipliers are @xmath191 , @xmath165 , and @xmath192 .
now , with the aforementioned lagrange multipliers , equations ( [ opu4])-([opu6 ] ) can be expressed as @xmath193 @xmath194 @xmath195 @xmath196 @xmath197 equations ( [ equ : b1])-([equ : b2 ] ) provide the optimal solution of @xmath172 and @xmath173 , while equations ( [ equ : b3])-([equ : b5 ] ) provide the necessary conditions of these optimal solutions .
now , solving equations ( [ equ : b1])-([equ : b2 ] ) , we obtain the optimal power allocations as follows : @xmath198 @xmath199 since @xmath200 , the solution of @xmath179 and @xmath173 all are positive .
now , with the considered lagrange multipliers and the resultant positive @xmath180 , equations ( [ opu2])-([opu3 ] ) can be expressed as follows : @xmath201 @xmath202 @xmath203 since the solutions of @xmath116 , @xmath204 , all are positive , equations ( [ equ : b6])-([equ : b8 ] ) show that the @xmath205 , @xmath206 , @xmath207 , and @xmath208 all are positive .
therefore , the lagrange multiplier set @xmath190 , satisfies the kkt conditions .
all the other cases can easily be verified by using a similar approach .
a. benjebbour , y. saito , y. kishiyama , a. li , a. harada and t. nakamura , `` concept and practical considerations of non - orthogonal multiple access ( noma ) for future radio access , '' in _ proc .
intelligent signal process . and commun
2013 , pp . 770774 .
a. benjebbour , k. saito , a. li , y. kishiyama , and t. nakamura , `` non - orthogonal multiple access ( noma ) : concept , performance evaluation and experimental trials , '' in _ proc .
wireless networks and mobile commun .
_ , 2015 .
z. ding , z. yang , p. fan , and h. v. poor , `` on the performance of non - orthogonal multiple access in 5 g systems with randomly deployed users '' , _ ieee signal process .
2014 , pp .
15011505 .
a. li , a benjebbour , x chen , h. jiang and h. kayama , `` uplink non - orthogonal multiple access ( noma ) with single - carrier frequency division multiple access ( sc - fdma ) for 5 g systems '' , _ ieice trans .
e98-b , no . 8 , aug .
2015 , pp . 14261436 .
s. sen , n. santhapuri , r.r .
choudhury , and s. nelakuditi , `` successive interference cancellation : a back - of - the - envelope perspective , '' in _ proc .
9th acm workshop on hot topics in networks ( hotnets - ix ) _ , oct .
|
non - orthogonal multiple access ( noma ) has recently been considered as a key enabling technique for 5 g cellular systems . in noma , by exploiting the channel gain differences multiple users are multiplexed into transmission power domain and then non - orthogonally scheduled on the same spectrum resources .
successive interference cancellation ( sic ) is then applied at the receiver(s ) to decode the message signals . in this paper , first we briefly describe the differences in the working principles of uplink and downlink noma transmissions .
then , for both uplink and downlink noma , we formulate a sum - throughput maximization problem in a cell such that the user clustering ( i.e. , grouping users into a single cluster or multiple clusters ) and power allocations in noma cluster(s ) can be optimized under transmission power constraints , minimum rate requirements of the users , and sic constraints . due to the combinatorial nature of the formulated mixed integer non - linear programming ( minlp ) problem
, we solve the problem in two steps , i.e. , by first grouping users into clusters and then optimizing their respective power allocations .
in particular , we propose a low - complexity sub - optimal user grouping scheme .
the proposed scheme exploits the channel gain differences among users in a noma cluster and group them into a single cluster or multiple clusters in order to enhance the sum - throughput of the system . for a given set of noma clusters
, we then derive the optimal power allocation policy that maximizes the sum throughput per noma cluster and in turn maximizes the overall system throughput . using kkt optimality conditions , closed - form solutions for optimal power allocations
are derived for any cluster size , considering both uplink and downlink noma systems .
numerical results compare the performance of noma over orthogonal multiple access ( oma ) and illustrate the significance of noma in various network scenarios .
5 g cellular , non - orthogonal multiple access ( noma ) , orthogonal multiple access ( oma ) , power allocation , throughput maximization , user grouping .
|
the periodically kicked rotor is one of the simplest systems whose classical motion exhibits chaotic dynamics , leading to an unbounded diffusive growth of its energy with the number of kicks .
in contrast , the energy growth of a quantum kicked rotor ( qkr ) is determined by the interference of quantum interaction pathways @xcite . in the quantum limit ,
the rotational excitation is either enhanced due to quantum resonances @xcite or suppressed due to the effect of dynamical localization @xcite , which has been linked to anderson localization in solids @xcite .
similarly to the localization of a quantum particle in a disordered one - dimensional lattice , the qkr localizes in a `` rotational lattice '' of angular momentum states .
experimental work has mostly been conducted in substitute systems , imitating the qkr behavior with rydberg atoms in microwave fields @xcite or ultracold atoms in optical lattices @xcite . in a series of recent articles @xcite ,
averbukh and coworkers proposed a strategy to study a number of qkr phenomena in an ensemble of diatomic molecules exposed to a periodic sequence of ultrashort laser pulses . in this system of true quantum rotors ,
the effects of a quantum resonance @xcite and dynamical localization ( dl ) @xcite have recently been demonstrated .
the recovery of classical diffusion under the influence of noise and decoherence has also been shown experimentally with atoms @xcite and molecules @xcite .
the discreteness of the qkr spectrum results in periodic dynamics with a revival time @xmath0 , determined by the rotational constant @xmath1 of the molecule , with @xmath2 being the speed of light . matching the period @xmath3 of a pulse sequence with the so - called quantum resonance at @xmath4 enables an efficient excitation of multiple rotational states with growing ( from kick to kick ) rotational energy . on the other hand ,
away from all full and fractional quantum resonances ( @xmath5 , where @xmath6 and @xmath7 are integers ) dynamical localization suppresses the rotational energy growth .
this quantum suppression of classical diffusion is accompanied by the exponential line shape of the kicked rotor s quasienergy eigenstates in the angular momentum space .
the shape of the observable wave packet depends on the overlap between these quasienergy states , whose spectrum is dictated by the pulse train period , and the initial rotational distribution of molecules @xcite . in a recent experimental demonstration of dl in laser - kicked molecules , we showed an exponentially localized wave function near the ( initially most populated ) rotational ground state @xcite .
we also demonstrated a shifting of the localization center towards higher energies by applying a periodic pulse train to a previously prepared coherent superposition of rotational states and controlling the relative phases between the states @xcite . here , we study an alternative route to control the final rotational distribution by varying the period of the applied sequence of laser pulses .
oxygen molecules , cooled down to 25 k in a supersonic expansion , are exposed to a series of 13 laser pulses .
we measure the shape of the created rotational wave packet by means of state - resolved coherent raman spectroscopy .
an exponentially localized distribution of the molecular angular momentum is observed in the case of periodic excitation , while a gaussian line shape , characteristic of classical diffusion , is found for periodic pulse sequences subject to timing noise . in both cases ,
we investigate the effect of the pulse train period on the rotational distribution - its center , width and shape .
we demonstrate a controlled shift of the quantum localization center and , similarly , the center of the classical distribution , which depends on the time separation between the pulses in a sequence and the nearest fractional quantum resonances .
the control mechanism is associated with the finite duration of the laser pulses , which results in a partially resonant excitation of molecules away from their ground rotational state .
the interaction of a diatomic molecule with a periodic train of @xmath8 linearly polarized laser pulses , not resonant with any electronic transition , is described by the following hamiltonian : @xmath9 where @xmath10 is the angle between the molecular axis and the vector of laser polarization , @xmath11 is the angular momentum operator , @xmath12 is the molecular moment of inertia and @xmath13 is the reduced planck constant .
the degree of stochasticity in the dynamics of a kicked rotor is governed by a single parameter @xmath14 , where @xmath15 is the effective planck constant and @xmath16 is a kick strength , @xmath17 is the molecular polarizability anisotropy and @xmath18 is the temporal envelope of the pulse .
all experiments discussed in this report satisfy the condition of @xmath19 , for which the phase space of the underlying classical motion is fully chaotic and supports an unbounded diffusive growth of rotational energy @xcite . for the periodically driven quantum system described by hamiltonian ( [ eq - hamiltonian ] ) , the solutions of the time - dependent schrdinger equation are the well - known floquet states @xmath20 , which are periodic in time , @xmath21 , up to a phase factor determined by the quasienergy @xmath22 . finding quasienergy states of the quantum kicked rotor is equivalent to solving a one - dimensional tight - binding model of a solid - state lattice , whose sites correspond to the rotational states @xmath23 @xcite .
owing to the selection rules for a rotational raman process , laser kicks couple only states of the same parity , resulting in an effective `` rotational lattice '' with a lattice constant @xmath24 .
the on - site energy is @xcite @xmath25 where @xmath26 is the energy of a molecule , which can be approximated by a rigid rotor due to the low degree of rotational excitation considered here .
the coupling between different lattice sites is defined by the kick strength @xmath27 , as described in detail in ref ..
the solution of the rotational tight - binding model critically depends on the site - to - site energy variation , which in turn is determined by the period of the pulse train .
it is instructive to express @xmath28 through the molecular revival time @xmath29 : @xmath30 on quantum resonance ( @xmath4 ) , the on - site energy does not depend on the site number , @xmath31 .
all quasienergy states in this _ periodic _ lattice are _ extended states_. the rotational population will therefore spread along the lattice , until it is stopped by the centrifugal distortion which destroys the periodicity @xcite .
the same phenomenon persists in a weaker form at fractional resonances ( @xmath32 ) @xcite , i.e. the on - site energies of the rotational lattice remain periodic with a period proportional to @xmath7 . if , however , all resonances are avoided , the argument of the tangent function in eq.([eq - tj ] ) changes from site to site in a way that makes the energy @xmath33 a _ pseudo - random _ function of @xmath23 .
the quasienergy states of such a disordered lattice are _ localized states _ @xcite .
the molecular kicked rotor is unique in the way the disorder strength depends on the rotational quantum number @xmath23 .
while anderson localization in real space is typically studied in a lattice with spatially uniform disorder , the randomness of the rotational lattice depends on the site number .
importantly , this dependence can be controlled by the excitation period @xmath3 through its proximity to quantum resonances , as we demonstrate in this work .
-.1truein [ fig : setup ] a high - intensity train of femtosecond pulses is generated in an optical system @xcite , shown schematically at the top of fig .
[ fig : setup ] .
we use a ti : sapphire laser system producing pulses of 130 fs full width at half maximum ( fwhm ) at a central wavelength of 800 nm , 1 khz repetition rate , and 2 mj per pulse .
part of the beam ( @xmath34 in energy ) is sent through a standard ` @xmath35 ' pulse shaper @xcite , which generates a sequence of 13 pulses of equal amplitudes separated by arbitrary time intervals in a total window of 50 ps .
the revival time of oxygen , @xmath36 ps , is long enough for a 130 fs pulse to act as a @xmath37-kick , yet short enough to generate sufficiently many pulses within the maximum time window accessible by the pulse shaper .
the pulse sequence is amplified by a home - built multi - pass amplifier to reach a kick strength of up to @xmath38 per pulse ( @xmath39 ) at 10 hz repetition rate .
the standard deviation of the pulse energy fluctuations is about @xmath40 .
the remaining part of the 800 nm beam is used as a probe .
its spectrum is narrowed down in a separate pulse shaper , before its central wavelength is shifted to @xmath41 nm by means of second harmonic generation in a nonlinear optical crystal .
the experimental setup is illustrated at the bottom of fig .
[ fig : setup ] .
the probe pulse of 0.15 nm spectral width ( fwhm ) is linearly polarized at 45 with respect to the pulses in the pump train .
both beams are focused into a vacuum chamber , where they are combined on a dichroic beamsplitter and intersect a supersonic jet of oxygen molecules .
the pump pulses produce coherent molecular rotation , which modulates the refractive index of the gas . as a result
, raman sidebands appear in the narrow - band spectrum of the weak probe pulse , polarized orthogonally to its initial polarization @xcite .
each raman peak is shifted from the central probe frequency by an amount that depends on the rotational quantum number @xmath23 , while its intensity @xmath42 allows the retrieval of the corresponding population @xmath43 @xcite .
critical to these experiments , we achieve a dynamic range of four orders of magnitude in raman intensity ( two orders in rotational population ) .
special care is taken to avoid detrimental effects of spatial averaging by making the probe beam significantly smaller than the pump ( fwhm beam diameters of @xmath44 and @xmath45 , respectively ) .
we use a @xmath46 diameter pulsed nozzle , operating at the repetition rate of 10 hz and the stagnation pressure of 33 bar , to lower the rotational temperature of oxygen down to 25 k at a distance of 2 mm from the nozzle .
markers indicate the values of @xmath3 , for which the relative phase between the two neighboring sites @xmath23 and @xmath48 of the rotational lattice is equal to an integer multiple of @xmath49 .
two sets of ten vertical lines each indicate the experimental pulse train periods , ( * 1 * ) @xmath50 and ( * 2 * ) @xmath51 . ]
-.1truein [ fig : map_periodic ] observing the effect of dynamical localization relies on avoiding fractional quantum resonances .
figure [ fig : map_periodic ] shows the location of resonances ( crosses ) and illustrates our strategy in choosing the appropriate pulse train periods ( vertical lines ) .
each cross represents the value of @xmath3 , for which the relative phase between the two neighboring sites @xmath23 and @xmath48 of the rotational lattice , @xmath52 , is equal to an integer multiple of @xmath49 , meaning that the two on - site energies are the same .
( equivalently , these are also the times when a rotational wave packet , consisting of the states @xmath53 and @xmath54 , completes an integer number of half - rotations ) .
note , that only odd rotational states are allowed because of the nuclear spin statistics of @xmath55o@xmath47 .
, ( * b * ) @xmath56 and ( * c * ) @xmath38 . shown
are the average distributions obtained from ten pulse trains with the mean period of ( * 1 * ) @xmath57 and ( * 2 * ) @xmath58 , see fig .
[ fig : map_periodic ] .
the populations are fitted with an exponential function ( thick green line ) and compared to the initial thermal distribution at @xmath59 k ( dashed grey line ) .
the dotted vertical line represents the excitation limit due to the finite pulse duration . ]
-.1truein [ fig : population_periodic ] we start by choosing a set of ten evenly spaced values of @xmath3 in the interval @xmath50 , marked with ( * 1 * ) in fig . [
fig : map_periodic ] . here , all resonances belonging to low - lying @xmath23-states , which are initially populated , are far off - resonant .
the upper row in fig .
[ fig : population_periodic ] shows the measured population distribution averaged over these ten _ periodic _ trains of 13 pulses each
. plots ( * 1a-1c * ) correspond to the individual kick strength of @xmath60 , respectively .
the rotational population has been extracted from the detected raman signal as @xmath61 , with the coefficient @xmath62 found from normalizing the total population to unity .
we note , that this method of calculating @xmath43 neglects the initial distribution of molecules among more than one rotational state , including the degenerate magnetic sub - levels , which results in a systematic underestimation of the amount of molecules at low @xmath23 s @xcite . in the analysis of our results below
, we therefore fit the extracted distributions starting from @xmath63 and up to the level containing more than 1% of the total population . the raman frequency shift ( horizontal axis )
has been converted to the rotational quantum number @xmath23 . in fig .
[ fig : population_periodic ] ( * 1a * ) , the angular momentum distribution decays exponentially away from the initially populated states centered at @xmath64 down to the instrumental noise floor around @xmath65 .
it can clearly be distinguished from the initial boltzmann distribution ( grey dashed line ) , which has been recorded after a single weak pulse .
the evident exponential shape ( fit indicated by the thick green line ) , with a localization length ( @xmath66 width ) @xmath67 , is a hallmark of dynamical localization .
note that an unavoidable ( due to the finite pulse length ) overlap with higher fractional resonances @xmath68 ( fig.2 ) is not only weak , but also averages out between different pulse trains , and hence does not distort the overall exponential line shape . for stronger kicks of @xmath56 and 8 (
* 1b,1c * ) , the localization length increases to @xmath69 and @xmath70 , respectively .
the distributions become sub - exponential at higher values of angular momentum due to the finite duration of the laser pulses ( 130 fs fwhm ) . beyond the limit of @xmath71 ( i.e. to the right of the dotted vertical line ) , an oxygen molecule rotates by @xmath72 during the length of the pulse , which lowers its effective kick strength and suppresses further rotational excitation .
each exponential fit only includes experimental values up to @xmath73 . + to demonstrate the control over the localized distribution and its localization center
, we now select a set of ten evenly spaced pulse train periods in the second interval @xmath51 , marked with ( * 2 * ) in fig .
[ fig : map_periodic ] .
although the selected periods are again nonresonant , as required for dl , the proximity of _ all _ pulse trains to the two low - lying fractional resonances associated with @xmath74 at @xmath75 and @xmath76 at @xmath77 , alters the shape of the localized distribution . in oxygen ,
the separation between these two resonances is about 300 fs , such that the finite pulse duration results in a partial overlap with both resonances for all chosen periods .
therefore , a population transfer to higher rotational states is facilitated before the qkr dynamically localizes .
this mechanism is unique to finite pulse durations and could not be replicated in our numerical simulations with @xmath78-kicks . in fig .
[ fig : population_periodic ] ( * 2a-2c * ) , the localization center @xmath79 is significantly shifted away from the edge , with its position being independent of the kick strength . an exponential fit of the form @xmath80 yields @xmath81 for @xmath60 , respectively ( green lines ) . in comparison to the far off - resonant case ( * 1a-1c * ) ,
the different localization lengths @xmath82 remain the same , as they are solely determined by the kick strengths of the pulse train . only for @xmath38 ( * 2c * ) we observe a larger localization length than in ( * 1c * ) , caused by a local flattening of the distribution around @xmath83 .
this deviation reflects the proximity of the train period to the next higher lying resonances ( see fig .
[ fig : map_periodic ] ) .
the effect is more pronounced for stronger kicks when such higher rotational states become populated .
dynamical localization relies on quantum coherences .
both timing and amplitude noise were shown to destroy localization and recover classical diffusion of the qkr angular momentum @xcite . here
, we investigate how fractional resonances affect the noise - induced classical diffusion . to make the proper comparison with the localized scenario
, we left the effective planck constant @xmath84 , the kick strength @xmath27 and , accordingly , the stochasticity parameter @xmath85 the same . . here
, 120 vertical lines indicate the random periods of ten different pulse sequences , 13 pulses each , following a gaussian distribution with a mean and a standard deviation of ( * 1 * ) @xmath86 and @xmath87 , ( * 2 * ) @xmath88 and @xmath89 , respectively . in case ( * 1 * ) , no period is within 150 fs of any fractional resonance associated with @xmath90 or 5 .
no such restriction was imposed in case ( * 2 * ) .
the dashed red lines mark the two mean periods . ]
-.1truein [ fig : map_random ] we introduce timing noise by randomly varying the time intervals between the 13 pulses in each train ( using the pulse shaping technique , described earlier ) with a standard deviation of @xmath91 around the mean period @xmath92 , which is chosen to be similar to its value in the case of periodic trains .
the upper plot in fig .
[ fig : map_random ] shows the first set of 120 random periods ( for 10 pulse trains ) , which follow a gaussian distribution with @xmath86 and @xmath93 , but are engineered to avoid all quantum resonances with low - lying rotational states @xmath90 or 5 . in case ( * 2 * ) , shown in the lower plot of fig .
[ fig : map_random ] , the distribution is truly random with no restrictions , @xmath88 and @xmath94 . , ( * b * ) @xmath56 and ( * c * ) @xmath38 . shown
are the average distributions obtained from ten pulse trains with a mean and a standard deviation of ( * 1 * ) @xmath86 and @xmath87 , ( * 2 * ) @xmath88 and @xmath89 , respectively , see fig . [
fig : map_random ] . in case (
* 1 * ) , no period is within 150 fs of any fractional resonance associated with @xmath90 or 5 .
the populations are fitted with a gaussian function ( thick green line ) and compared to the initial thermal distribution at @xmath59 k ( dashed grey line ) .
the dotted vertical line represents the excitation limit due to the finite pulse duration . ]
-.1truein [ fig : population_random ] figure [ fig : population_random ] presents the population distributions , obtained from the average over ten _
nonperiodic _ pulse trains , for the same six cases exhibited in fig .
[ fig : population_periodic ] .
we observe a qualitatively different nonexponential shape , expected for a classical kicked rotor .
in contrast to the exponential line shapes of dl , now destroyed by noise , the angular momenta follow a gaussian distribution ( thick green line ) , which is characteristic of classical diffusion .
its @xmath66 width @xmath95 is wider than the corresponding localized counterpart @xmath96 for all six sets of experimental parameters . in plots ( * 1a-1c * ) , when the random periods are designed to avoid low - lying resonances , a gaussian fit yields the distribution center at @xmath97 and a width of @xmath98 for @xmath60 , respectively . the distribution becomes broader with the increasing kick strength . in comparison ,
when the train periods randomly overlap with fractional resonances [ plots ( * 2a-2c * ) ] , we detect a noticeable shift of the gaussian center towards higher rotational states @xmath99 , which becomes larger with the increasing kick strength @xmath27 .
the distribution widths of @xmath100 remain similar , within the errors of the fit , to the ones measured with set ( * 1 * ) of the noisy pulse trains . with a kick strength @xmath101 ( red circles ) , @xmath56 ( blue triangles ) and @xmath38 ( black squares ) .
compared are the energies for ( * a * ) periodic and ( * b * ) nonperiodic sequences , with periods that ( * 1 * ) avoid or ( * 2 * ) allow overlap with low - lying quantum resonances , as shown in fig .
[ fig : map_periodic ] and fig .
[ fig : map_random ] . ]
-.1truein [ fig : energy ] the ability to resolve individual rotational states allows us to determine the rotational energy , absorbed by the molecules , as @xmath102 .
the absorbed energy is plotted in fig .
[ fig : energy ] as a function of the number of kicks @xmath8 for all the discussed excitation scenarios . in the case of nonresonant
_ periodic _ pulse trains ( * 1a * ) , the rotational energy ceases to grow after as few as three pulses , after which the dynamical localization sets in .
larger localization lengths for the increasing kick strength , from @xmath101 ( red circles ) to @xmath56 ( blue triangles ) to @xmath38 ( black squares ) , are reflected in a higher saturation level of the rotational energy .
periodic pulse trains that promote the initial population transfer via the @xmath74 and 5 resonances lead to a greater absorption of energy ( * 2a * ) .
it takes more pulses for the dynamical localization to set in and the total energy to saturate . to observe a completely suppressed energy growth
one would need a sequence of more than 13 pulses .
in contrast , _ nonperiodic _ sequences result in a continuous increase of rotational energy , exceeding the energy achieved by the corresponding periodic sequences .
the growth rate is faster for truly random periods ( * 2b * ) compared to random periods that are tailored to avoid fractional resonances ( * 1b * ) .
the continuous absorption of energy and the gaussian line shape are both manifestations of classical diffusion .
the sublinear growth rate is due to the finite duration of the laser pulses , discussed earlier in the text .
we investigated how the period of a pulse train affects the shape of the angular momentum distribution of a molecular rotor .
off - resonant excitation , which can be mapped onto pseudo - random on - site energies in a rotational lattice , is shown to lead to dynamical localization . due to the finite duration of the experimental pulses , the shape of the localized distribution
is affected by the partial overlap of the train period with fractional quantum resonances .
the average distribution obtained from ten periodic sequences with different nonresonant periods revealed an exponentially localized spectrum with a localization length @xmath96 that depended only on the kick strength @xmath27 .
we measured @xmath103 for values of @xmath60 , respectively . on the other hand , by matching the periodicity of the pulse train to low - lying resonances , we transferred the population to higher rotational states before it dynamically localized .
we demonstrated a localized distribution centered around @xmath104 with the anticipated localization length according to different kick strengths . breaking the periodicity by introducing timing noise
destroys the localization of the quantum rotor s wave function and leads to classical diffusion .
we showed that the resulting gaussian population distribution becomes wider with the number of pulses as well as their kick strength .
the distribution center is equally affected by the relative position of the random periods with respect to fractional resonances . omitting resonances resulted in a slower spread to higher @xmath23-states .
the rotational energy of the qkr has been studied as a function of the number of kicks .
it reflected both processes , the one of dynamical localization , when the energy growth is shown to cease completely after as few as three pulses , and the one of classical diffusion with an unbounded increase in energy .
|
we experimentally study a system of quantum kicked rotors - an ensemble of diatomic molecules exposed to a periodic sequence of ultrashort laser pulses . in the regime , where the underlying classical dynamics is chaotic , we investigate the quantum phenomenon of dynamical localization by means of state - resolved coherent raman spectroscopy .
we examine the dependence of the exponentially localized angular momentum distribution and of the total rotational energy on the time period between the pulses and their amplitude .
the former parameter is shown to provide control over the localization center , whereas the latter one controls the localization length .
similar control of the center and width of a nonlocalized rotational distribution is demonstrated in the limit of classical diffusion , established by adding noise to the periodic pulse sequence .
|
gamma - ray bursts ( grbs ) are the brightest phenomena in the universe , releasing as much electromagnetic energy in tens of seconds as the entire milky way galaxy does in a few years @xcite .
they typically reach energies of around 5 x 10@xmath0 ergs when beaming is accounted for @xcite .
their temporal distribution shows a bimodality @xcite which separates them into ` long ' or ` short ' grbs ( lgrb and sgrb respectively ) depending on a parameter known as t@xmath1 ; the time in which 90% of the gamma - ray fluence is detected .
nominally , long bursts have t@xmath2 seconds , and short ones have t@xmath3 seconds , but in reality the distinction is far more blurred for a significant number of cases ( eg @xcite ; @xcite ; @xcite ) .
both classes have been observed to be distributed isotropically across the sky @xcite .
the most popular theory for sgrbs is that they are caused by mergers of compact objects , such as double neutron star ( ns ns ) binaries , ns black hole ( bh ) mergers , white dwarf ( wd ) ns mergers , wd bh mergers or possibly even wd wd mergers @xcite .
lgrbs are perhaps the better understood of the two classes , and are believed to be the product of massive star collapse @xcite since , in cases where it would be observable , they are always accompanied by type ib / c supernovae @xcite .
sgrbs are in fact not necessarily short .
@xcite found that 1/3 of their sample of sgrbs exhibit extended emission ( ee ) in their light curves , and an even greater fraction were found to be extended in the batse catalogue .
ee is the term given to a rebrightening in the light curve after the initial emission spike .
it happens at early times , usually beginning at @xmath4 10 s , and typically has lower peak flux but much longer duration than the initial spike , resulting in comparable total fluences between the two @xcite .
those bursts that were believed to exhibit ee were catalogued by @xcite .
these bursts present a challenge to the standard merger scenario , since they require an injection of energy arising seconds after the trigger , then naturally switching off at later times , around 100 s after trigger in the rest frame .
one central engine with the potential to provide such an energy supply is a rapidly spinning , highly magnetized neutron star , known as a magnetar @xcite . whether by collapse or merger , this magnetar may be formed with sufficient rotational energy to avoid collapsing into a bh @xcite .
grbs require a rapidly rotating central engine with a strong , large scale magnetic field of around @xmath5 g or higher @xcite , and a magnetar with such a field and an initial spin period @xmath6 ms has enough rotational energy to power a @xmath7 erg grb .
magnetar spin down has been often discussed as a source of grbs , both long @xcite and short @xcite , in the literature , and has also been suggested as the origin of ee @xcite along with a variety of other mechanisms .
alternatives include a two - jet solution @xcite , fallback accretion @xcite , r - process heating of the accretion disc @xcite and magnetic reconnection and turbulence @xcite .
a major motivation for a common central engine can be seen in figure 1 ; when all bursts are overlaid , a striking similarity can be seen in the evolution of the bursts , both temporally and energetically .
conformity like this is highly suggestive of a shared origin .
+ this paper is structured as follows : section 2 contains the sample selection and data reduction process while section 3 details the motivation for finding a common central engine for ee grbs .
the magnetar model is introduced in section 4 , the results are discussed in section 5 , and the main conclusions are summarised in section 6 .
.selected sample of ee grbs . bracketed values for redshift , @xmath8 ,
indicate no published value was available . in these cases
the mean value of the ee sample where @xmath8 is known was used . a - upper limit . 1 - @xcite ; 2 - @xcite ; 3 - @xcite ; 4 - @xcite ; 5 - @xcite ; 6 - @xcite ; 7 - @xcite ; 8 - @xcite [ cols="<,^,^,<",options="header " , ]
the calculated spin periods for the birth of the magnetar lie comfortably within allowed parameter space ( figure 3 ) and are consistent with values predicted in the literature @xcite .
bursts that do not have a set redshift may vary on the energy scale , with an error of @xmath9 in @xmath8 roughly corresponding to an order of magnitude in the luminosity scale .
@xcite discussed the effect of varying redshift on the results for @xmath10 and @xmath11 in their work , and the argument is well illustrated by figure 9(b ) in @xcite .
the general result is that a higher @xmath8 corresponds to a lower rotation period ( ie faster spin ) and lower magnetic field .
a good example is the change in results if the sample average redshift @xmath12 is used for grb 051227 ; fitting the magnetic dipole spin - down model then gives a magnetic field @xmath13 g and a spin period of @xmath14 ms .
the light curve is also far less luminous .
the ee energy release is just @xmath15 ergs , which translates into @xmath16 ms .
figure 4 shows where the values found for @xmath10 and @xmath17 place the ee bursts relative to other sgrb and lgrb populations taken from figure 9(a ) of @xcite .
it can be seen that the ee bursts show properties that most closely resemble the unstable magnetar population of sgrbs .
since both magnetic field and spin period are very similar between these two groups , the difference must lie in some other property , perhaps mass or formation mechanism .
this key difference must prevent the ee sample bursts from collapsing into bhs , and enable , perhaps even cause , the release of ee energy .
@xcite showed that accretion discs and fallback accretion exhibit a much wider spread of behaviours when the compact objects involved in the merger have different masses . in their work , a ns
ns binary showed fairly homogeneous behaviour , whilst a ns
bh merger produced a much broader spread of fallback activity .
a magnetar can not be formed from a bh , but the same principle of unequal masses can be achieved by a system involving a ns wd merger , or , with the discovery of increasingly @xcite , possibly a more exotic ns ns system .
ee grb light curves show a remarkable uniformity when plotted alongside each other , particularly amongst the bursts where redshift is known .
this consistency in plateau luminosity and turnover times suggests ee grbs share a common progenitor mechanism which distinguishes them from ordinary sgrbs . we have fitted the magnetic dipole spin - down model of @xcite to the late - time data of the light curves of 9 grbs under the assumption that the central engine is a highly magnetized neutron star .
these fits have yielded values for the magnetic field strength and late - time spin period .
we have also performed calculations of the energy contained in the ee region of bursts in this sample . assuming this energy release is due to the spin - down of the central magnetar , and assuming a constant magnetic field
, we infer the spin period these magnetars possessed at birth .
the spin periods found are in good agreement with published values for the birth of a magnetar ( eg @xcite ; @xcite ; @xcite ) .
these results are consistent with the idea that ee grbs could be powered by a spinning - down magnetar .
bg acknowledges funding from the science and technology funding council .
the work makes use of data supplied by the uk _ swift _ science data centre at the university of leicester and the _ swift _ satellite .
_ swift _ , launched in november 2004 , is a nasa mission in partnership with the italian space agency and the uk space agency . _
swift _ is managed by nasa goddard .
penn state university controls science and flight operations from the mission operations center in university park , pennsylvania .
los alamos national laboratory provides gamma - ray imaging analysis .
we thank the anonymous referee for their swift response .
99 arnaud k.a . , 1996 , aspc , 101 , 17 barbier l. , et al . , 2005 , gcn circ , 4397 , 1 barkov m.v . , pozanenko a.s . , 2011 , mnras , 417 , 2161 barthelmy s.d . , et al . ,
2005a , ssrv , 120 , 143 belczynski k. , perna r. , bulik t. , kalogera v. , ivanova n. , lamb d.q . , 2006 , apj , 648 , 1110 berger e. , 2005 , gcn circ , 3966 , 1 berger e. , 2007 , apj , 670 , 1254 bernardini m.g .
, margutti r. , zaninoni e. , chincarini g. , 2012 , msais , 21 , 226 bloom j.s .
, frail d.a . , sari r. , 2001 , apj , 121 , 2879 briggs m.s . , et al .
, 1996 , apj , 459 , 40 bucciantini n. , metzger b.d . ,
thompson t.a .
, quataert e. , 2012 , mnras , 419 , 1537 burrows d.n . ,
et al . , 2005 , ssrv , 120 , 165 cannizzo j.k . , et al . , 2006 , gcn circ , 5904 , 1 cenko s.b . , kasliwal m. , cameron p.b . , kulkarni s.r . ,
fox d.b . , 2006 , gcn circ , 5946 , 1 covino s. , et al . , 2005 , gcn circ , 3665 , 1 chapman r. , levan a.j . , priddey r.s . , tanvir n.r . ,
wynn g.a . , king a.r . , davies m.b .
, 2007 , aspc , 372 , 415 cummings j.r . , et al . , 2009 , gcn circ , 9461 , 1 dallosso s. , stratta g. , guetta d. , covino s. , de cesare g. , stella l. , 2011 , a&a , 526 , a121 davanzo p. , fiore f. , piranomonte s. , covino s. , tagliaferri g. , chincarini g. , stella l. , 2007 , gcn circ , 7152 , 1 davanzo p. , et al . , 2009 , a&a , 498 , 711 delia v. , et al . , 2011 , gcn circ , 12578 , 1 demorest p.b . , pennucci t. , ransom s.m . , roberts m.s.e . ,
hessels j.w.t . , 2010 , nature , 467 , 1081 dessart l. , burrows a. , livne e. , ott c.d . , 2008 , apj , 673 , 43 duncan r.c .
, thompson c. , 1992 , apj , 392 , 9 evans p.a . , et al . , 2007 , a&a , 469 , 379 evans p.a . , et al . , 2009 , mnras , 397 , 1177 fan y .- z . ,
xu d. , 2006 , mnras , 372 , l19 frail d. , et al . , 2001 , apj , 562 , l55 fryer c.l . ,
woosley s.e . , herant m. , davies m.b .
, 1999 , apj , 520 , 650 galama t.j . , et al . , 1998 , nature , 395 , 670 gehrels n. , et al . , 2004 , apj , 611 , 1005 gehrels n. , et al . , 2006 , nature , 444 , 1044 graham j.f .
, et al . , 2009 , apj , 698 , 1620 kouveliotou c. , meegan c.a . ,
fishman g.j . ,
, briggs m.s . ,
koshut t.m .
, paciesas w.s . , pendleton g.n . , 1993 ,
apj , 413 , l101 lattimer j.m . , prakash m. , 2004 , sci , 304 , 536 lyons n. , obrien p.t . , zhang b. , willingale r. , troja e. , starling r.l.c . , 2010 , mnras , 402 , 705 macfadyen a.i .
, woosley s.e . , 1999 , apj , 524 , 262 mao j. , et al . , 2008 , gcn circ , 7665 , 1 mckinney j.c . , 2006 , mnras , 368 , 1561 mszros p. , 2006
, rpph , 69 , 2259 metzger b.d .
, quataert e. , thompson t.a . , 2008 , mnras , 385 , 1455 metzger b.d . , arcones a. , quataert e. , martnez - pinedo g. , 2010 , mnras , 402 , 2771 metzger b.d . , giannios t.a . ,
thompson t.a .
, bucciantini n. , quataert e. , 2011 , mnras , 413 , 2031 norris j.p .
, bonnell j.t . , 2006 , apj , 643 , 266 norris j.p . , gehrels n. , scargle j.d . , 2010 ,
apj , 717 , 411 pacznski b. , 1986 , apj , 308 , 43 pacznski b. , 1998 , apj , 494 , 45 page k.l . , et al . , 2005 , gcn circ , 3961 , 1 page k.l . , et al . , 2006 , apj , 637 , l13 parsons a.m. , et al . , 2006 , gcn circ , 5252 , 1 perley d.a . , et al .
, 2009 , apj , 696 , 1871 piran t. , bromberg o. , nakar e. , sari r. , 2012 , arxiv , arxiv:1206.0700 price p.a . , berger e. , fox d.b . , 2006 , gcn circ , 5275 , 1 prochaska j.x . , bloom j.s .
, chen h .- w .
, hansen b. , kalirai j. , rich m. , richer h. , 2005 , gcn circ , 3700 , 1 racusin j.l .
, barthelmy s.d . ,
burrows d.n . ,
chester m.m . ,
gehrels n. , krimm h.a . , palmer d.m . , sakamoto t. , 2007 , gcn circ , 6620 , 1 racusin j.l . , et al . , 2009 , gcn circ , 9666 , 1 roming p.w . , et al . , 2005 , ssrv , 120 , 95 rosswog s. , ramirez - ruiz e. , davies m.b .
, 2003 , mnras , 345 , 1077 rosswog s. , 2007 , mnras , 376 , l48 rowlinson a. , et al .
, 2010 , mnras , 409 , 531 rowlinson a. , obrien p.t . ,
metzger b.d . , tanvir n.r .
, levan a.j .
, 2013 , mnras , 608 sakamoto t. , et al . , 2007 , gcn circ , 7147 , 1 schady p. , et al . , 2006 , gcn circ , 5699 , 1 stanek k.z . , et al . , 2003 , apj , 591 , 17 thompson t.a .
, chang p. , quataert e. , 2004 , apj , 611 , 380 thompson t.a .
, 2007 , rmxac , 27 , 80 troja e. , et al . , 2009 , gcn circ , 9913 , 1 ukwatta t.n . , et al .
, 2008 , gcn circ , 7203 , 1 usov v.v . , 1992 ,
nature , 357 , 472 woosley s.e .
, 1993 , apj , 405 , 273 zhang b. , mszros p. , 2001
, apj , 552 , 35 zhang b. , yan h. , 2011 , apj , 726 , 90
|
extended emission gamma - ray bursts are a subset of the ` short ' class of burst which exhibit an early time rebrightening of gamma emission in their light curves .
this extended emission arises just after the initial emission spike , and can persist for up to hundreds of seconds after trigger . when their light curves are overlaid , our sample of 14 extended emission bursts show a remarkable uniformity in their evolution , strongly suggesting a common central engine powering the emission .
one potential central engine capable of this is a highly magnetized , rapidly rotating neutron star , known as a magnetar .
magnetars can be formed by two compact objects coallescing , a scenario which is one of the leading progenitor models for short bursts in general . assuming a magnetar is formed
, we gain a value for the magnetic field and late time spin period for 9 of the extended emission bursts by fitting the magnetic dipole spin - down model of @xcite . assuming the magnetic field is constant , and the observed energy release during extended emission is entirely due to the spin - down of this magnetar
, we then derive the spin period at birth for the sample .
we find all birth spin periods are in good agreement with those predicted for a newly born magnetar .
[ firstpage ] general gamma rays : bursts .
|
supershells are giant bubbles with a scale of hundreds of parsecs , and are believed to be created by multiple stellar winds from ob associations and supernova explosions @xcite .
supershells play an important role in galaxy evolution since they redistribute huge amounts of energy and material from the galaxy disc to the galaxy halo .
the interstellar matter swept up by supershells forms cold and dense regions in which new star formation is triggered .
the typical lifetime of supershells is about 10@xmath7 years @xcite , much longer than those of supernova remnants ( snrs ; * ? ? ?
supershells have been found with observations of 21-cm line emission originating from the cold and dense shells @xcite , in which infrared emission associated with dust is also observed . for younger supershells ,
x - ray emission was found from their inner hot gas ( t@xmath810@xmath9 k ) @xcite .
furthermore , far ultraviolet ( fuv ) emission lines such as and were also observed from their interface region of t@xmath510@xmath10 k between hot gas and cold ambient medium @xcite .
optical emission lines such as h@xmath11 were observed from their ionized shells @xcite .
hence , multi - wavelength study for supershells including fuv wavelengths may give us critical information about their environment and evolution .
gsh 006 - 15 + 7 is one of the milky way supershells discovered recently through a 21-cm line survey of the southern sky , which was compiled as the galactic all sky survey ( _ gass _ ) @xcite .
gsh 006 - 15 + 7 spans across the sky from _
= 356to _ l _ = 16 in longitude and from _ b
_ = -28to _ b _ = 2 in latitude , which is translated into the physical size of 790 @xmath12 520 pc at a distance of 1500 @xmath0 500 pc .
@xcite also discussed in detail the physical properties of the supershell .
the structure of the supershell is most obvious in the range of v@xmath13@xmath55 km s@xmath14 to v@xmath13@xmath511 km s@xmath14 , with a systemic velocity of @xmath57 km s@xmath14 , whereas it is difficult to identify the shell - like structure in the maps of the lower velocity channels of v@xmath13 @xmath15 5 km s@xmath14 , probably because of the contamination by local emission .
its dynamical age has been estimated to be @xmath515 @xmath0 5 myr , which is consistent with the ages found for similar galactic supershells with well - defined shapes .
the formation energy was estimated to be @xmath510@xmath16 ergs , with a mass of @xmath53 @xmath0 2 @xmath12 10@xmath9 m@xmath17 and an expansion velocity of 8 km s@xmath14 .
it was suggested that the energy source of the expansion of gsh 006 - 15 + 7 is likely to be associated with sgr ob association 1 and some open clusters nearby .
however , the absence of high - mass stars in these stellar groups , which are able to produce significant stellar winds , indicates the expansion was at least partially made by supernova explosions and the winds of progenitor stars .
@xcite found some evidence for the breakout of the supershell at ( _ l _ , _ b _ ) = @xmath5(10 , -25 ) based on the infrared and maps , indicating mass and energy transfer from the galactic disk to the halo . whereas the physical characteristics described above were obtained from multi - wavelength data , fuv observations
have not been reported yet .
it is well known that interstellar fuv emission is produced in the regions of a cooling hot gas @xcite , from scattering by dust grains @xcite , and in photo - dissociation regions ( pdrs ) @xcite ; of these , dust scattered fuv emission is regarded to be the dominant component @xcite .
we present in this paper the results of analysis based on archival data of fuv observations .
section 2 describes observational results of the fuv continuum image and spectrum for the region of the supershell . in section 3 , a distance to the supershell
is estimated based on three - dimensional multi - scattering simulations of dust scattering .
the results of monte carlo simulations for the pdr region of the supershell are also presented in section 3 .
a summary of the present study is given in section 4 .
the present study primarily used the fuv observations on the rectangular region spanning from ( _ l _ , _ b _ ) = ( -10 , -35 ) to ( 20 , -5 ) which included part of the supershell gsh006 - 15 + 7 .
a study of continuum emission was conducted with the data set obtained by the galaxy evolution explorer ( _ galex _ ; * ? ? ?
_ galex _ carried out an ultraviolet all sky survey in the fuv ( 13501750 ) band , but it does not provide spectral information . hence , a spectral study was proceeded with the observations from the far - ultraviolet imaging spectrograph ( _ fims _ , also known as spectroscopy of plasma evolution from astrophysical radiation or _ spear _ ) on board the korean microsatellite stsat-1 .
_ fims _ contains dual fuv imaging spectrographs , referred to as the s - band ( 9001150 ) and l - band ( 13501750 ) . _
fims _ is the main payload of stsat-1 , which was operated for a year and a half after its launch on september 27 , 2003 , in a 700 km sun - synchronous orbit .
_ fims _ was designed to have a large field of view ( 75 @xmath124 3 in l - band ) with angular resolution of @xmath55 , which is suitable for observations of diffuse targets , while its spectral resolution is rather moderate ( @xmath18/@xmath19 @xmath5 550 ) . more information on the _ fims _ mission and instruments can be found in @xcite .
we also used the dust extinction map based on the @xcite ( sfd ; * ? ? ? * ) for comparisons with fuv observations .
+ figure [ fig : fuv ] shows the ( a ) _ galex _ fuv continuum map , ( b ) sfd dust extinction map over the rectangular region spanning from ( _ l _ , _ b _ ) = ( -10 , -35 ) to ( 20 , -5 ) , and ( c ) a scatter plot for the fuv intensity against dust extinction for the region spanning from ( _ l _ , _ b _ ) = ( -5 , -35 ) to ( 15 , -15 ) , which is referred to as the lower supershell region henceforth .
contours at the systemic velocity of 7 km s@xmath14 , following the result of @xcite , are overplotted in figures [ fig : fuv](a ) and [ fig : fuv](b ) ( the parkes galactic all - sky survey ( _ gass _ ) dataset ; * ? ? ?
* ; * ? ? ?
the figure [ fig : fuv](a ) and [ fig : fuv](b ) were plotted with 05 pixel size and smoothing of the image to have 1resolution .
the _ galex _
observations avoided bright stars intentionally , leaving some regions unobserved .
most of these missing areas were filled for the region of _ b _ @xmath15 -15 , where the lower supershell region resides , with coarse - grained binning and smoothing , but some regions still remain unfilled , especially in the region of the galactic disk , as figure [ fig : fuv](a ) shows . the dust extinction map of figure [ fig : fuv](b ) shows extremely high optical depth in the region of _ b _ @xmath8 -15near the galactic disk , probably due to the foreground dust . the shell - like feature associated with gsh 006 - 15 + 7
is not seen in this optically thick region .
hence , we focus only on the lower region of _ b _ @xmath15 -15 in the present analysis . the scatter plot of figure [ fig : fuv](c ) was constructed from a pixel - by - pixel comparison between figure [ fig : fuv](a ) and [ fig : fuv](b ) for the lower supershell region .
what is notable in the fuv continuum map of figure [ fig : fuv](a ) is that the intensity is higher at the lower supershell region of _ b _ @xmath15 -15than its surroundings .
this is considered to result from dust scattering of the region since the lower supershell region has higher extinction levels than its neighbors , as shown in figure [ fig : fuv](b ) .
we note that the optical depth for the _ e(b - v ) _ value of 0.14 corresponds to @xmath20 @xmath5 1 at @xmath18 = 1565 , which is indicated by the vertical dashed line in figure [ fig : fuv](c ) . a strong correlation between the fuv intensity and _ e(b - v )
_ is seen in figure [ fig : fuv](c ) for the optically thin region of @xmath20 @xmath15 1 , except the vertical branch above the fuv intensity of 1500 cu .
this agrees well with the general notion that fuv continuum intensity correlates with _ e(b - v ) _ in an optically thin region @xcite .
the vertical branch , where data points are scattered up to @xmath5 3500 cu , are the bright pixels corresponding to the regions of _ b _ @xmath8 -18 .
being close to the galactic plane , they represent the scattered component of the high - intensity radiation fields from the galactic plane .
we also note that fuv intensity is anti - correlated with dust extinction for _ e(b - v ) _ above 0.2 , which is reasonable for an optically thick region such as r cra of the present case , as can be seen in figure [ fig : fuv](b ) . the dense cloud at ( _ l _ , _ b _ ) @xmath5 ( 0 , -18 ) in figure [ fig : fuv](b ) , corresponding to the partial extinction seen in the fuv continuum map of figure [ fig : fuv](a ) , is the r cra molecular cloud @xcite which is located at a distance of @xmath5170 pc from the sun @xcite .
figure [ fig : spec ] shows the spectrum made with _ fims _ over the rectangular region spanning from ( _ l _ , _ b _ ) = ( -10 , -35 ) to ( 20 , -5 ) of gsh 006 - 15 + 7 .
bright point sources in the region were removed from the data set when obtaining the above spectrum .
more specifically , the pixels corresponding to the bright stars in the td-1 catalog with 1565 band flux of above 10@xmath21 erg @xmath14 s@xmath14 @xmath3 were removed , together with those within a radius of 2around the stars whose intensities were more than twice as high as the median values . among various emission peaks ,
the emission line is readily noticeable in the spectrum , together with h@xmath1 fluorescent emission features .
while the molecular hydrogen emission is associated with the shell boundaries , it will be shown that the emission does not seem to be related to the shell .
as mentioned previously , the fuv continuum emission associated with gsh 006 - 15 + 7 is mainly from dust scattering .
hence , we attempted simulations of dust scattering to obtain the distance to the supershell region , by comparing the observation with a simulation for the fuv continuum , as was done previously for many individual targets @xcite .
this will produce an independent measurement of the distance that we will compare with the 1500 @xmath0 500 pc distance estimate that @xcite calculated from the gas kinematics .
for comparison with the dust scattering simulation , the observed fuv map should also have only the contribution from dust scattering .
we found that other sources of fuv emission have only a small effect on this region : ion line emission and h@xmath1 fluorescent emission are @xmath518 percent of the total fuv intensity , and the two photon continuum emission ranges from @xmath520 to @xmath590 cu , based on the southern h - alpha sky survey atlas @xcite .
we used the conversion factor of 28.4 cu for 1 rayleigh of h@xmath11 intensity to obtain the two photon continuum emission @xcite .
these contributions were removed from the map of figure [ fig : fuv](a ) and the result is shown in figure [ fig : dust](a ) with dust extinction contours at _ e(b - v ) _ of 0.07 , 0.14 and 0.21 for the lower supershell region .
the simulation code employs a heney - greenstein function for the scattering model : it describes light scattering off grains using the albedo and the phase function asymmetry factor , or the g - factor , as free parameters @xcite . in the present simulations , however ,
the albedo and g - factor are fixed at 0.4 and 0.6 , respectively , based on the theoretical estimations for carbonaceous - silicate grains @xcite , and monte carlo simulations are performed for various distributions of dust to obtain the best dust distribution model by comparing the simulated fuv intensity with the observed one .
we take a rectangular box of 280 @xmath12 280 @xmath12 700 cells as the simulation domain , corresponding to 1400 pc @xmath12 1400 pc @xmath12 3500 pc with the longest dimension along the line of sight .
the center of the map of ( _ l _ , _ b _ ) = ( 5 , -25 ) corresponds to the sightline of the central axis , and the sun is located in the center of the front face .
the 1400 pc @xmath12 1400 pc size at the rear face spans an angular size of 226 @xmath12 226 on the sky , encompassing the lower supershell region of 20@xmath12 20 .
the dust distribution of the supershell is idealized to be a single slab whose distance and thickness are to be determined in comparison with the observed fuv map .
we note that the region also has a local cloud r cra , as well as a possible contribution from the foreground , both of which must be subtracted from the sfd map to obtain the effect of dust solely associated with the supershell .
the foreground contribution is modeled as follows .
we divide the sfd map of _ l _
= -5to 15into 500 vertical strips and fit each strip with a linear function that generally follows the minimum of _ e(b - v)_. each linear fit , given as a function of latitude , represents the _ e(b - v ) _ value integrated along the sightline of the given latitude .
we distribute dust along the sightline according to the distance from the galactic plane assuming that dust distribution has an exponential decrease outwards from the galactic plane with a scale height of 100 pc in view of the previous studies @xcite . with this distribution of dust ,
88 percent of the total foreground dust is located within 500 pc from the sun , and 99 percent within 1000 pc .
the amount of dust associated with the supershell is determined by subtracting these foreground values from those of the sfd map at corresponding locations . for the region of the supershell
co - located with r cra , _
e(b - v ) _
= 0.1 , the mean value of the nearby supershell region is assigned to the supershell
. the dust layer corresponding to r cra is placed at 170 pc from the sun with a thickness of 5 pc . with
this given distribution of dust for the foreground and r cra , the distance to the dust clouds corresponding to the supershell is decided by comparing the simulation with the observation for the fuv continuum .
the source photons to be scattered by dust are difficult to model for distant targets such as gsh 006 - 15 + 7 since stars have not been completely surveyed to comparable distances .
in fact , the number of observed stars decreases rapidly when the distance from the sun is greater than @xmath5500 pc .
hence , we consider two types of source photons .
first , the photons from the bright stars of the td-1 catalogs are scattered by the foreground dust and the r cra cloud .
second , for scattering from the supershell gsh 006 - 15 + 7 , we adopted the interstellar radiation field ( isrf ) obtained from the galprop code ( http://galprop.stanford.edu/ and * ? ? ? * ) .
galprop provides energy density at a specified wavelength , for which we adopt @xmath18 = 1565 , as a function of location .
it should be noted that @xmath18 = 1565 is one of the passbands of the td-1 catalogs and is located nearly at the center of the fuv passband of _
galex_. furthermore , we confirmed that galprop gives a specific energy density at @xmath18 = 1565 , which is very similar to the mean energy density averaged over the wavelength range of 13501750 .
we assume the photons , corresponding to the specified energy density at a given location , originate from the galactic plane and propagate vertically to the supershell , which we believe is reasonable as the supershell is located in the high latitude region and there are not many nearby stars .
the scattered photons from the supershell are allowed to interact with the foreground dust as well as r cra . with these photon sources , together with the model distribution of foreground dust and r cra ,
simulations were conducted for different sets of distance and thickness of the supershell .
the distance is varied from 500 pc to 2500 pc with 100 pc steps , and the thickness from 20 pc to 200 pc with 20 pc steps , and the resulting images are compared with the observation for the region below _
b _ = -18and excluding r cra to obtain the best - fit parameters using the chi - squared minimization method .
we note that the region close to the galactic plane above _ b _
= -18is excluded because of the contamination by the foreground local dust .
figure [ fig : dust](b ) shows the simulated fuv map with 05 @xmath12 05 pixel bins for the best - fit parameters : 1300 pc for the distance to the center of the dust slab and 60 pc for the thickness of the slab .
the 1-sigma confidence range of the fit is 1300 @xmath0 800 pc for the distance to the center of the dust slab , which is in good agreement with the estimation of 1500 @xmath0 500 pc by @xcite .
we have also varied albedo and g - factor over a reasonable range , with albedo from 0.3 to 0.5 and g - factor from 0.4 to 0.7 based on @xcite , respectively , to check the dependence of the distance estimation on the changes of these optical parameters .
we obtained the distance in the range of 8002000 pc , which is within the 1-sigma range of the distance obtained with albedo = 0.4 and g - factor = 0.6 .
while the range of thickness is not well constrained in the present simulations , the best - fit value of 60 pc is similar to the @xmath553 pc estimated from the column density of n@xmath4 @xmath5 ( 2.77.7 ) @xmath12 10@xmath22 @xmath3 and the number density of n@xmath4 @xmath5 ( 1.74.8 ) atoms @xmath6 in @xcite .
figure [ fig : dust](c ) is a residual map for the best - fit parameters that shows the difference between the simulation result of figure [ fig : dust](b ) and the observation of figure [ fig : dust](a ) .
figure [ fig : dust](d ) is a scatter plot obtained from a pixel - to - pixel comparison between the observation and the simulation maps .
while the simulation and the observation are in general agreement over the entire region , we see that there are several local regions with marked differences . for example
, the simulated intensity is much higher than the observed intensity in the region around ( _ l _ , _ b _ ) = ( 10 , -16 ) in figure [ fig : dust](c ) , which is located close to the galactic plane and affected by the foreground local dust .
the large fluctuations above 2000 cu of the observed intensity in figure [ fig : dust](d ) correspond to this region .
another excess in simulated intensity is seen near ( _ l _ , _ b _ ) = ( 2 , -21 ) , which is close to r cra , implying that the discrepancy is due to the imperfect removal of the local dust effect .
high simulated intensities in the interval of 1000 to 1500 cu of the observed intensity in figure [ fig : dust](d ) correspond to this region . on the other hand , the simulated intensity is lower than the observed one in the region around ( _ l _ , _ b _ ) = ( 6 , -28 ) .
low simulated intensities in the interval of 1700 to 2400 cu of the observed intensity in figure [ fig : dust](d ) correspond to this region .
the region is affected by scattering from the foreground local dust of the photons of a bright star hd189103 located 190 pc from the sun .
hd189103 is the brightest uv star in the region of consideration shown in figure [ fig : dust ] according to the td-1 catalogs . with these three regions excluded , together with the region of _ e(b - v ) _ less than 0.07 where dust scattering effect is expected to be small ,
the correlation coefficient and the reduced chi - squared value are estimated to be 0.94 and @xmath23 = 3.18 , respectively , with 790 degrees of freedom .
the rather large value of the reduced chi - square may indicate that the foreground contamination still remains , especially in the low latitude region near the galactic plane . 3 , excluding the r cra cloud region , and the best - fit result of the pdr simulation ( red solid line ) .
the blue dash - dotted line is a model continuum emission .
[ fig : pdr],title="fig:",width=264 ] + most of the h@xmath1 fluorescent emission seen in figure [ fig : spec ] originates from the lower supershell region .
we modeled this region as a pdr with the radiative transfer simulation code cloud @xcite .
we took the lower supershell region below _
b _ = -15with a thickness of @xmath53 , excluding the r cra cloud region , whose spectrum is shown in figure [ fig : pdr ] as a black solid line with the spectral resolution of the instrument , which is 3 around 1600 .
we fit the h@xmath1 fluorescence lines only for the spectral intervals from 1562 to 1616 to avoid the effects from the ion emission lines .
the continuum level of fuv emission is shown as a blue dash - dotted line , which generally follows the minimum data points of the spectrum .
we assume i@xmath24 , a scaling factor representing the local isrf normalized by the model of @xcite , to be 2 , in view of the isrf intensity of @xmath53.5 @xmath12 10@xmath10 photons s@xmath14 @xmath3 @xmath14 given by galprop at 1300 .
the trial values and the ranges of parameters in the simulation are as follows : the density _
n@xmath4 _ = [ 1 , 3 , 10 , 30 , 100 ] @xmath6 , the temperature _ t _ = [ 10 , 30 , 100 , 300 , 1000 ] k , and the column density of molecular hydrogen _
n(h@xmath1 ) _ = 10@xmath25}$ ] @xmath3 with 10@xmath26 steps .
all 325 possible combinations of these fitting parameters were explored in our simulations .
the best - fit model is shown in figure [ fig : pdr ] as a red solid line with the parameters of _ n@xmath4 _ = 30 @xmath6 , _ t _ = 300 k , and _
n(h@xmath1 ) _ = 10@xmath27 @xmath3 .
the reduced chi - squared value is @xmath23 @xmath5 1.36 with 18 degrees of freedom .
the 1-sigma confidence ranges are as follows : _
n(h@xmath1 ) _ = 10@xmath2 @xmath3 , _ n@xmath4 _ @xmath28 10 @xmath6 , while the temperature is not constrained over the range of 101000 k. in fact , the average flux of h@xmath1 fluorescent emission was hardly affected over the given range of other parameters when temperature was increased up to @xmath55000 k. as we approximate the total hydrogen column density n@xmath4 to be the neutral atomic hydrogen column density of @xmath5 4.7 @xmath12 10@xmath29 @xmath3 suggested by @xcite because the overall contributions by n(h+ ) and n(h@xmath1 ) are expected to be minimal , the range of the molecular hydrogen column density _
n(h@xmath1 ) _ = 10@xmath2 @xmath3 , when divided by the total hydrogen column density n@xmath4 , yields the h@xmath1 abundance , n(h@xmath1)/n@xmath4 , to be 2 @xmath12 10@xmath30 2 @xmath12 10@xmath14 .
this result is reasonable when compared with the model shown in figure 31.3 of @xcite , in which the h@xmath1 abundance increases from below 10@xmath31 for the warm neutral medium ( wnm ) to above 10@xmath14 for the cold neutral medium ( cnm ) , although the transition was obtained for a slightly different value of equilibrium pressure ( p / k = 3000 @xmath6 k while p / k = 9000 @xmath6 k in the present simulation ) and occurred at n@xmath4 of @xmath58 @xmath12 10@xmath22 @xmath3 .
the intensities are given in lu ( photons s@xmath14 @xmath3 sr@xmath14 ) the data points in gray color on the bottom left denote low signal - to - noise ratios of less than 3 .
[ fig : civ],height=302 ] a emission line is prominently seen in the fuv spectrum of figure [ fig : spec ] . as a emission
is generally observed in the interface region between hot gas and cold medium at a transition temperature of t @xmath5 10@xmath10 k , it should be interesting to locate the origin of this important cooling line .
we have constructed a spectral map for the emission line , as shown in figure [ fig : civ ] , with its intensities given in line units ( photons s@xmath14 @xmath3 sr@xmath14 , henceforth lu ) integrated over the wavelength range of 15401560 .
the map is plotted with a 1pixel size , but an averaged spectrum of 4centered around each pixel was fitted to enhance the signal - to - noise ratios ( s / n ) .
we see that , while the emission peaks in the bottom region of the shell , the large angular extent of the emission region across the shell over more than ten degrees , without any resemblance to the shell - like morphology , indicates that the emission may not be related to the structure of the present supershell .
furthermore , the high intensity of 3 @xmath12 10@xmath32 lu may not be consistent with the quite evolved ( @xmath515 myrs old ) and distant nature of the shell .
for example , the peak intensity was estimated to be @xmath56000 lu at the boundary of the orion - eridanus superbubble whose age is a few myrs , much younger than gsh 006 - 15 + 7 .
@xcite also found that soft x - rays do not show any convincing features associated with the supershell .
hence , we believe the bright emission in this region is of local origin , perhaps the local hot bubble or loop i superbubble . in this regard
, we note that @xcite observed intensity above 10@xmath32 lu toward the interaction zone of the local hot bubble and loop i.
we have reported here the results of analysis based on the fuv observations of gsh 006 - 15 + 7 made with _
galex _ and _ fims_. the main findings are as follows .
fuv emission is seen to be enhanced above 2000 cu in the lower supershell region ( _ b _
@xmath15 -15 ) of the supershell region , which was demonstrated to mostly come from the dust scattering of interstellar photons of the region .
monte carlo simulation for dust scattering was performed for the lower supershell region under the theoretical albedo of 0.4 and asymmetry factor of 0.6 and the distance to gsh 006 - 15 + 7 was estimated to be 1300 @xmath0 800 pc , corresponding well with the previous estimation of 1500 @xmath0 500 pc based on kinematics .
molecular hydrogen fluorescence features are seen in the spectrum of the lower supershell region : a pdr simulation indicated an h@xmath1 column density of n(h@xmath1 ) = 10@xmath2 @xmath3 and the total hydrogen density of n@xmath4 @xmath5 30 @xmath6 . the authors appreciate the anonymous referee for the valuable comments which have significantly improved the paper .
_ fims / spear _ was funded by the ministry of science and technology ( korea ) and a grant nag5 - 5355 ( nasa ) .
it is a joint project of the korea astronomy and space science institute , the korea advanced institute of science and technology , and the university of california at berkeley ( usa ) .
min acknowledges the support by the national research foundation of korea through its grant no .
nrf 2012m1a3a4 a01056418 .
black , j. h. , & van dishoeck , e. f. 1987 , , 322 , 412 boumis , p. , dickinson , c. , meaburn , j. , et al .
2001 , , 320 , 61 choi , y .-
min , k .- w . ,
seon , k .-
i . , et al .
2013 , , 774 , 34 draine , b. t. 1978 , , 36 , 595 draine , b. t. 2003 , , 41 , 241 draine , b. t. 2011 , physics of the interstellar and intergalactic medium ( princeton , nj : princeton univ . press ) , chapter 31 edelstein , j. , et al . 2006a , , 644 , l153 edelstein , j. , et al .
2006b , , 644 , l159 ehlerov , s. , & palou , j. 2005 , , 437 , 101 ehlerov , s. , & palou , j. 2013 , , 550 , a23 frail , d. a. , goss , w. m. , & whiteoak , j. b. z. 1994 , , 437 , 781 gaustad , j. e. , mccullough , p. r. , rosing , w. , & van buren , d. 2001 , , 113 , 1326 heiles , c. 1979 , , 229 , 533 heiles , c. 1984 , , 55 , 585 heiles , c. , haffner , l. m. , & reynolds , r. j. 1999 , in asp conf .
168 , new perspectives on the interstellar medium , ed .
a. r. taylor , t. l. landecker , & g. joncas ( san francisco , ca : asp ) , 211 henyey l. g. , greenstein j. l. 1941 , , 93 , 70 hurwitz , m. , bowyer , s. , & martin , c. 1991 , , 372 , 167 jo , y .- s . ,
min , k .- w . ,
seon , k .-
i . , edelstein , j. , & han , w. 2011 , , 738 , 91 jo , y .- s . ,
min , k .- w . , lim , t .- h . , & seon , k .-
2012 , , 756 , 38 kalberla , p. m. , mcclure - griffiths n. m. , pisano , d. j. , et al .
2010 , , 521 , 17 kim , i .-
j . , min , k .- w . ,
seon , k .-
, et al . 2007 , , 665 , l139 knude , j. , & hog , e. 1998 , , 338 , 897 lee , d .- h .
, yuk , i .- s . , jin , h. , et al .
2006 , , 644 , l181 lim , t .- h . ,
min , k .- w . , & seon , k .-
2013 , , 765 , 107 mcclure - griffiths , n. m. , dickey , j. m. , gaensler , b. m. , & green , a. j. 2002 , , 578 , 176 mcclure - griffiths , n. m. , pisano , d. j. , calabretta , m. r. , ford , h. a. , lockman , f. j. , et al .
2009 , , 181 , 398 misiriotis , a. , xilouris , e. m. , papamastorakis , j. , boumis , p. , & goudis , c. d. 2006 , , 459 , 113 morrissey , p. , conrow , t. , barlow , t. a. , et al .
2007 , , 173 , 682 moskalenko , i. v. , porter , t. a. , & strong , a. a. 2006 , , 640 , l155 moss , v. a.,mcclure - griffiths n. m. , braun , r. , hill , a. s. , & madsen , g. j. 2012 , , 421 , 3159 parenago , p. p. 1945
, pa , 53 , 441 reynolds , r. j. , tufte , s. l. , haffner , l.m . , jaehnig , k. , & percival , j.w .
1998 , pasa , 15 , 14 rossano , g. 1978 , , 83 , 234 ryu , k. , min , k .- w . ,
park , j .- w .
2006 , , 644 , l185 sallmen , s. m. , korpela , e. j. , & yamashita , h. 2008 , , 681 , 1310 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 seon , k .-
i . , han , w. , nam , u .- w . ,
2006 , , 644 , l175 seon , k .-
i . , edelstein , j. , korpela , e. , et al .
2011 , , 196 , 15 suad , l. a. , caiafa , c. f. , arnal , e. m. , & cichowolski , s. 2014 , , 564 , 116s tenorio - tagle , g. & bodenheimer , p. 1988
, , 26 , 145 van dishoeck , e. f. , & black , j. h. 1986 , , 62 , 109
|
we have analyzed the archival data of fuv observations for the region of gsh 006 - 15 + 7 , a large shell - like structure discovered by @xcite from the velocity maps .
fuv emission is seen to be enhanced in the lower supershell region .
the fuv emission is considered to come mainly from the scattering of interstellar photons by dust grains .
a corresponding monte carlo simulation indicates that the distance to the supershell is 1300 @xmath0 800 pc , which is similar to the previous estimation of 1500 @xmath0 500 pc based on kinematic considerations .
the spectrum at lower galactic latitudes of the supershell exhibits molecular hydrogen fluorescence lines ; a simulation model for this candidate photodissociation region ( pdr ) yields an h@xmath1 column density of n(h@xmath1 ) = 10@xmath2 @xmath3 with a rather high total hydrogen density of n@xmath4 @xmath5 30 @xmath6 .
|
@xmath1-ray observations have long been recognized as a potential , valuable tool for supernovae research @xcite .
only @xmath1-rays provide a direct link to the ni distribution which hardly depends on details of the physics and on the numerical treatment .
different scenarios can be distinguished by line fluxes and profiles , the structure of the progenitors can be probed , and the time of the explosion can be determined .
@xmath0-rays can provide a good determination of the @xmath2 production for nearby sne ia because , nowadays , accurate distances of nearby galaxies can be obtained by @xmath3ceph .
moreover , all sky surveys by @xmath0-rays may provide an unbiased rate of sne ia . as we will discuss below
, advances in the fields of optical and ir observations and of the theory have helped to redefine the goals of sn research by @xmath0-rays .
new challenges emerged which emphasize their central role as probes for the 3-d structure of sne ia .
the results presented are based on our gamma - ray codes for spherical @xcite and arbitrary 3-d geometries @xcite .
we assume homologous expansion of density and chemical structures calculated by spherical and 3-d hydro simulations .
all nuclear decay lines of @xmath2 and @xmath4 are included .
pair production and bound - free opacities are taken into account .
* classical questions and problems : * in parts , this section is based on our previous analyses @xcite . for further discussions ,
we also want to refer to @xcite,@xcite and @xcite .
overall , @xmath1 spectra and their evolution is characterized by a turnover from a phase dominated by @xmath2 to @xmath7 lines ( fig .
the time of the explosion can be determined by the ratio between the @xmath2(0.81mev ) and the @xmath7(0.84 mev ) lines because it varies strongly with time of the explosion but it hardly depends on the model . with time
, the envelope becomes increasingly transparent .
the spectral evolution depends sensitively on the density and chemical structure of the envelope and , thus , provides a valuable tool for the discrimination of explosion models .
for example , sub - chandrasekhar mass models show as a distinguishing feature an outer layer of @xmath2 which reveals itself by high @xmath1-ray fluxes and broads already a few days after the explosion ( fig .
[ f1 ] ) . at late phases ( @xmath8100 days ) ,
the absolute line fluxes and profiles provide a direct measure of the total @xmath2 mass and its distribution . for @xmath9 models ,
the high densities close to the center ( @xmath10 ) result in the production of neutron - rich iron - group isotopes rather than @xmath2 .
the size of this central region increases with @xmath11 .
a knowledge of this property is critical to detect because systematic variations in @xmath11 can produce an offset ( up to @xmath12 , @xcite ) in the brightness decline relation @xcite , a cornerstone of modern cosmology with sne ia ( e.g. @xcite ) .
optical and near ir - spectra do not allow to distinguish isotopes .
however , the lack of central @xmath2 reveals itself in by flat - topped line profiles ( fig .
[ f2 ] ) . to detect the variation , we need resolutions between 20 to 30 which are well within reach for the upcoming integral mission .
we want to mention one problem related to the analysis of observed line fluxes during the first months after the explosion .
an advantage of @xmath0-ray compared to optical analyses is that the results are insensitive to details of the physics or numerical treatment .
e.g. the specific opacities do not depend on temperature or density .
this advantage is somewhat lost along the way when comparing the observations with theoretical predictions .
problems are caused by the time- and model - dependent line shifts and widths ( @xmath13 ) , the intrinsic response function of the instrument which is non - gaussian , and they are connected to the actual definition used to determine integrated line fluxes from synthetic spectra .
the actual value may differ by up to a factor of 2 even if based on a given synthetic spectrum ( e.g. @xcite ) .
preferable , the same machinery should be employed for both the synthetic spectra and the observations to derive accurate values or good upper limits for the @xmath1 emission .
-rays at day 1 ( left ) and 23 ( right ) based on our full 3-d mc gamma ray transport @xcite ( upper panels ) .
the diameter of the wd is normalized to 100 . in the lower panels , the @xmath0 spectra at day 23 are given as seen from various @xmath14 of 90 , 60 , 30 and 0@xmath15 ( red , blue , green , pink ) and @xmath16 . in the presence of inhomogeneities , typical fluctuations of about 10 %
can be seen .
their frequency provides a measure of the scale .
the explosion model is based on the delayed detonation model 5p0z22.20 ( [ f2 ] ) with @xmath17 assuming chemical inhomogeneities according to @xcite . ,
title="fig:",width=196 ] -rays at day 1 ( left ) and 23 ( right ) based on our full 3-d mc gamma ray transport @xcite ( upper panels ) . the diameter of the wd is normalized to 100 . in the lower panels , the @xmath0 spectra at day 23 are given as seen from various @xmath14 of 90 , 60 , 30 and 0@xmath15 ( red , blue , green , pink ) and @xmath16 .
in the presence of inhomogeneities , typical fluctuations of about 10 % can be seen .
their frequency provides a measure of the scale .
the explosion model is based on the delayed detonation model 5p0z22.20 ( [ f2 ] ) with @xmath17 assuming chemical inhomogeneities according to @xcite . ,
title="fig:",width=200 ] * new challenges : * during the last few years , observational and theoretical methods in the field of sn - research advanced significantly well beyond the point imaginable as little as 10 years ago .
back then , sne ia were discovered a few days before maximum or later and , in general , light curves were rather uncertain and the spectral coverage was poor . even for nearby sne ia , distances to the host galaxy were uncertain by 20 to 30 % .
now , accurate @xmath3-ceph .
based distances of nearby galaxies are available ( e.g. @xcite ) . combined with the establishment of the optical brightness decline relation relation @xcite , sne ia provide a unique tool for cosmology . robotic , systematic sn - search programs ( e.g. loss , c.f .
@xcite ) continuously increase the sample both at low , intermediate and high red - shifts .
often , sne ia are discovered 4 to 5 magnitudes before maximum light @xcite , almost eliminating the selection effects by galactic extinction and providing a tight handle on the rise times .
detailed observations of optical and infrared spectra and light curves allowed sophisticated analyses and test of scenarios , ruling all but out the once popular helium triggered detonations , and strongly favoring deflagration or delayed detonation @xmath9 models or , in some instances , the merger scenario ( e.g. @xcite ) .
observation of polarization in sne ia has shown that , in general , these objects are fairly spherical @xcite with the noticeable exception of very subluminous sne ia @xcite . for the first time , a direct connection with the progenitors seems to be within reach . in particular , there is mounting evidence for a connection between the properties of the progenitor , and the physics of the explosion @xcite . the recent progress may redefine the role of modern @xmath0-ray astronomy for the field of supernovae , and some of classical goals may have been rendered less compelling .
@xmath0-rays are particular valuable to measure the 3-d structure of sne ia .
a comprehensive list of new goals is beyond the format of this paper .
we want to address two of the areas , namely , @xmath0-rays as tools to reveal properties of nuclear burning front and large scale asymmetries . _
nuclear burning fronts : _ within @xmath9 models , optical and ir lcs and spectra can be reproduced by models in which a ( slow ) deflagration front turns into a detonation ( e.g. @xcite ) or , alternatively , a deflagration front is rapidly accelerating as in w7 @xcite . however , successful models require parameterized descriptions for the propagation of the burning front . for a discussion , see @xcite and references therein .
the propagation of a detonation front is well understood but the description of the deflagration and the deflagration to detonation transition ( ddt ) pose problems . on a microscopic scale , a deflagration propagates due to heat conduction by electrons .
though the laminar flame speed in sne ia is well known , the front has been found to be rayleigh - taylor ( rt ) unstable , increasing the effective speed of burning @xcite .
recently , significant progress has been made toward a better understanding of the physics of flames .
starting from static wds , hydrodynamic calculations of the deflagration fronts have been performed in 2-d @xcite , and full 3-d @xcite .
these calculations demonstrated the complicated morphology of the front .
@xcite finds that , while the expansion of the envelope becomes almost spherical , the inhomogeneous chemical structure will fill about 50 to 70% of the star ( in mass ) .
the resulting chemical inhomogeneities and their scale depends sensitively on the structure of the initial wd , i.e. progenitor and the pre - conditioning of the runaway . if a ddt occurs at densities needed to reproduce normal - bright sne ia , most of the unburned fuel will be consumed during the detonation phase and , by enlarge , the chemical inhomogeneities will be eliminated .
however , they will survive in pure deflagration models .
currently , state of the art 3-d calculations are restricted to the regime of linear instabilities , i.e. the flamelet regime .
the consistent treatment is restricted to the early part of the deflagration phase . in these calculations ,
the initial acceleration of the the deflagration front is followed by a declining rate of burning .
this leave a significant fraction of the wd unburned ( @xmath18 ) .
the resulting structures can not account for the observations of typical sne ia @xcite .
this problem is well known from spherical deflagration models with a low deflagration speed ( e.g. df1 , @xcite ) , and it triggered the suggestion of a ddt , or , alternatively , showed the need for continuously accelerating deflagration front @xcite . despite the limitations of current models
, the chemical clumps will not mix in the subsequent phase . to test for possible effects of chemical inhomogeneities produced during the deflagration
, we have remapped a 3-d chemical structure on a spherical explosion model ( fig .
the resulting line profiles show fluctuations of about 10 % .
their frequency provides a measure of the scale of the instabilities which depends on the c / o ratio of the progenitor .
weak or absent fluctuations in the @xmath0-ray spectra would be a strong indicator for a ddt .
_ large scale asymmetries : _ one of the open questions is the nature of subluminous sne ia such as sn 1991bg @xcite . among them , sn 1999by is one of the best observed sne ia .
in addition to the studies of optical light curves , detailed polarization spectra of the subluminous sn 1999by have been obtained and analyzed @xcite . whereas ` normal '
sne ia tend to show little or no polarization @xcite , this supernova was significantly polarized , up to 0.7% , indicating an overall asphericity of the photosphere of @xmath19 .
this result suggests that there may be a connection between the observed asphericity and the subluminosity in sne ia . among others ,
a possible explanations are the explosion of a rapidly rotating wd and its effect on the propagation of nuclear flames during the explosive phase of burning , or extensive burning of carbon just prior to the runaway @xcite .
polarization in the optical measures the asymmetry of the photosphere but , in general , provides little information whether it is caused by a global asymmetry in the density structure or in the excitation mechanism , i.e. the @xmath2 distribution . to test the sensitivity of @xmath0-rays ,
we have calculated the transport for the subluminous model of ellipsoidal shape ( figs .
the flux and line profiles vary as a function of inclination ( e.g. by @xmath20 at day 23 ) . in combination with optical and ir observations
, @xmath0-rays can be a key to answer the question on the nature of the asymmetry .
aldering g. , knop , p. , nugent p. 2000 , apj , 119 , 2110 ambwani k. , sutherland p.g .
1988 , apj 325 , 820 burrows a. , shankar a. , van riper k.a .
1991 , apj , 379 , l7 chan k.w . ,
lingenfelter r.e .
1991 , 368 , 515 clayton d. , colgate s.a . , fishman g.j .
1969 , apj 155 , 75 domnguez i. , hflich p. , straniero o. 2001 , apj , in press ( aug .
10 ) filippenko a. v. et al .
1992 , aj , 104 , 1543 hamuy , m. , phillips , m. m. , maza , j. , suntzeff , n. b. , schommer , r. a. , & avils , r. , 1995 , aj , 109 , 1 hflich p. , mller e. , khokhlov a. 1992 , a&a 259 , 549 hflich , p. , & khokhlov , a. 1996 , apj 457 , 500 hflich p. , wheeler c.j . , khokhlov a. 1998 , apj 492 , 228 hflich , p. , wheeler , j. c. , & thielemann , f. k 1998 , apj , 495 , 617 hflich p. , gerardy c. , fesen r. , sakai s. 2001 , apj , submitted hflich p. , lee e. 2001 , apj , in preparation howell , a. , hflich , p. , wang , l. , & wheeler , j. c. 2001 , apj 556 , 302 iwamoto , k. , brachwitz , f. , nomoto , k. , kishimoto , n. , umeda , h. , hix , w. r. , & thielemann , f. k. 1999 , apj s , 125 , 439 khokhlov , a. 2001 , apj , in press & astro - ph/0008463 khokhlov , a. 1991 , apj , 245 , 114 kumagai k. , nomoto k. 1997 , in `` thermonuclear supernovae '' , ed .
ruiz - lapuente et al .
, nato asi ser .
c. 486 , 515 leibundgut , b. et al .
1993 , aj , 105 , 301 lisewski , a. m. ; hillebrandt , w. ; woosley , s. e. ; niemeyer , j. c. ; kerstein , a. r. 2000 , apj 537 , 405l mller e. , hflich p. , khokhlov a. 1991 248l , 7 nomoto k. , sugimoto s. , & neo s. 1976 , apss , 39 , l37 nomoto , k. , thielemann , f. -k . , & yokoi , k. 1984 , apj , 286 , 644 perlmutter s. et al .
1999 , apj 517 , 565 pinto p. , eastman r.g . ,
rogers t. 2001 , apj 551 , 231 riess a. et al .
1999 , aj 118 , 2675 saha a. et al .
1997 , apj 486 , 1 , schmidt b. p. et al .
1998 , apj 507 , 46 treffers et al .
1997 , iau - circ .
7627 wang , l. , howell , a. , hflich , p. , wheeler j.c .
2001 , apj 550 , 1030 wang , l. , wheeler , j. c. , & hflich , p. 1997 ,
apj , 476 , l27
|
we present @xmath0-ray spectra for a set of type ia supernovae models .
our study is based on a detailed monte carlo transport scheme for both spherical and full 3-d geometries .
classical and new challenges of the @xmath0 ray astronomy are addressed .
we find that @xmath0-rays are very suitable to reveal the structure of the envelope and , thus , they allow to probe properties of the nuclear burning front and the progenitor , namely its central density and global asphericities .
the potential problems are discussed for the quantitative comparison between theoretical and observed line fluxes during the first few months after the explosion .
|
physics beyond the standard model ( sm ) is required by the following experimental observations : * neutrino oscillations : @xmath18 ( mass difference)@xmath19 needed to understand atmospheric and solar neutrino observations ; * cmb anisotropy ( @xmath20 ) : requires inflation which can not be realized in the sm ; * non - baryonic dark matter ( @xmath21 ) : sm has no plausible candidate ; * baryon asymmetry ( @xmath22 ) : not possible to achieve in the sm . recall that at the renormalizable level the sm possesses a global @xmath0 symmetry .
if the symmetry is gauged , the anomaly cancellation requires the existence of three right handed neutrinos .
an important question therefore is the symmetry breaking scale of @xmath0 .
note that this scale is not fixed by the evolution of the three sm gauge couplings .
remarkably , we will be able to determine the @xmath3 by implementing inflation . with @xmath3 well below the planck scale the seesaw mechanism enables us to realize light neutrino masses in the desired range .
furthermore , it will turn out that leptogenesis is a natural outcome after inflation is over .
the introduction of a gauge @xmath0 symmetry broken at a scale well below the planck scale exacerbates the well known gauge hierarchy problem .
there are at least four potential hierarchy problems one could consider : * @xmath23 ; * @xmath24 ( required by neutrino oscillations ) ; * @xmath25 ( where @xmath26 denotes the inflaton mass ) ; * @xmath27 , where @xmath28 denotes the axion decay constant .
supersymmetry ( susy ) can certainly help here , especially if the susy breaking scale in the observable sector is of order tev .
thus , it seems that a good starting point , instead of sm@xmath29 , could be mssm@xmath29 . the @xmath13 ` matter ' parity associated with the mssm has two important consequences .
it eliminates rapid ( dimension four ) proton decay , and it delivers a respectable cold dark matter candidate in the form of lsp . however , planck scale suppressed dimension five proton decay is still present and one simple solution is to embed @xmath13 in a @xmath30 symmetry .
it turns out that the r symmetry also plays an essential role in realizing a compelling inflationary scenario and in the resolution of the mssm @xmath12 problem .
finally it seems natural to extend the above discussion to larger groups , especially to @xmath31 and its various subgroups .
in this section , we review a class of supersymmetric hybrid inflation models @xcite where inflation can be linked to the breaking of @xmath0 . we compute the allowed range of the dimensionless coupling in the superpotential and the dependence of the spectral index on this coupling , in the presence of canonical supergravity ( sugra ) corrections .
the simplest supersymmetric hybrid inflation model @xcite is realized by the renormalizable superpotential @xcite @xmath32 where @xmath33 denote a conjugate pair of superfields transforming as nontrivial representations of some gauge group @xmath34 , @xmath35 is a gauge singlet superfield , and @xmath36 @xmath37 is a dimensionless coupling . a suitable @xmath11 r - symmetry , under which @xmath38 and @xmath35 transform the same way , ensures the uniqueness of this superpotential at the renormalizable level @xcite . in the absence of supersymmetry
breaking , the potential energy minimum corresponds to non - zero vacuum expectation values ( vevs ) @xmath39 in the scalar right handed neutrino components @xmath40 for @xmath41 and @xmath42 , while the vev of @xmath35 is zero .
( we use the same notation for superfields and their scalar components . )
thus , @xmath34 is broken to some subgroup @xmath43 which , in many interesting models , coincides with the mssm gauge group . in order to realize inflation , the scalar fields
@xmath41 , @xmath42 , @xmath35 must be displayed from their present minima . for @xmath44 ,
the @xmath41 , @xmath42 vevs both vanish so that the gauge symmetry is restored , and the tree level potential energy density @xmath45 dominates the universe , as in the originally proposed hybrid inflation scenario @xcite . with supersymmetry
thus broken , there are radiative corrections from the @xmath46 supermultiplets that provide logarithmic corrections to the potential which drives inflation . in one loop approximation
the inflationary effective potential is given by @xcite @xmath47\ , , \end{aligned}\ ] ] where @xmath48 , @xmath49 is the dimensionality of the @xmath41 , @xmath42 representations , and @xmath50 is a renormalization mass scale . the scalar spectral index @xmath51 is given by @xcite @xmath52 where @xmath53 gev is the reduced planck mass ( @xmath54 ) .
we are going to use units @xmath55 , although sometimes we will write @xmath56 explicitly .
the primes denote derivatives with respect to the normalized real scalar field @xmath57 . for relevant values of the parameters ( @xmath58 ) ,
the slow roll conditions ( @xmath59 , @xmath60 ) are violated only ` infinitesimally ' close to the critical point at @xmath61 ( @xmath62 ) @xcite .
so inflation continues practically until this point is reached , where it abruptly ends .
the number of @xmath63-folds after the comoving scale @xmath64 has crossed the horizon is given by @xmath65 where @xmath66 is the value of the field at the comoving scale @xmath64 and @xmath67 is the value of the field at the end of inflation . using eqs .
( [ loop ] , [ efold1 ] ) , we obtain @xmath68 here , the subscript 0 implies that the values correspond to the comoving scale @xmath69 , with @xmath70 mpc@xmath71 .
@xmath72 is the number of @xmath63-folds+@xmath73 , where @xmath9 is the reheat temperature and @xmath74 is the false vacuum energy density . ] and @xmath75 with @xmath76 the amplitude of the curvature perturbation @xmath77 is given by . ]
@xmath78 using eqs .
( [ loop ] , [ m_kap ] , [ perturb ] ) , @xmath77 is found to be @xcite @xmath79 up to now , we ignored supergravity ( sugra ) corrections to the potential . more often than not , sugra corrections tend to derail an otherwise succesful inflationary scenario by giving rise to scalar ( mass)@xmath19 terms of order @xmath80 , where @xmath43 denotes the hubble constant .
remarkably , it turns out that for a canonical sugra potential ( with minimal khler potential @xmath81 ) , the problematic ( mass)@xmath19 term cancels out for the superpotential @xmath38 in eq .
( [ super ] ) @xcite .
this property also persists when non - renormalizable terms that are permitted by the @xmath30 symmetry are included in the superpotential .
is expanded as @xmath82 , and only the @xmath83 term in @xmath84 generates a ( mass)@xmath19 for @xmath35 , which would spoil inflation for @xmath85 @xcite . from the requirement @xmath86 ,
one obtains an upper bound on @xmath87 ( @xmath88 for @xmath89 , @xmath90 for @xmath91 ) @xcite . we should note that , since the superpotential is linear in the inflaton , the presence of other fields with planck scale vevs also induce an inflaton mass of order @xmath43 . some ways to suppress the inflaton mass are discussed in @xcite . ] the scalar potential is given by @xmath92+v_d\,,\ ] ] with @xmath93 where the sum extends over all fields @xmath94 , and @xmath95 is the minimal khler potential .
the d - term @xmath96 vanishes in the d - flat direction @xmath97 . from eq .
( [ sugra ] ) , with a minimal khler potential one contribution to the inflationary potential is given by @xcite @xmath98\,.\ ] ] there are additional contributions to the potential arising from the soft susy breaking terms . in @xmath99 sugra
these include the universal scalar masses equal to @xmath100 ( @xmath101 tev ) , the gravitino mass . however , their effect on the inflationary scenario is negligible , as discussed below .
the more important term is the @xmath102 term @xmath103 .
for convenience , we write this as @xmath104 , where @xmath105 . the effective potential is approximately given by eq .
( [ loop ] ) plus the leading sugra correction @xmath106 and the @xmath102 term : @xmath107 + a\,m_{3/2 } \kappa m^2|s|\ , .
\end{aligned}\ ] ] we perform our numerical calculations using this potential , taking @xmath108=1 tev .
it is , however , instructive to discuss small and large @xmath36 limits of eq .
( [ potential ] ) . for @xmath109 , @xmath110 , and eq .
( [ potential ] ) becomes @xmath111\ ] ] to a good approximation .
comparing the derivatives of the radiative and sugra corrections one sees that the radiative term dominates for @xmath112 . from @xmath113 , @xmath114 for the one - loop effective potential , so that sugra effects are negligible only for @xmath115 .
( for @xmath116 , this essentially agrees with @xcite ) .
@xmath77 is found from eq .
( [ v1 ] ) to be @xmath117 in the absence of the sugra correction , the gauge symmetry breaking scale @xmath118 is given by eq .
( [ quad ] ) . for @xmath109 , @xmath119 and @xmath120 .
@xmath77 in this case turns out to be proportional to @xmath121@xcite . using the wmap best fit @xmath122 @xcite
, @xmath118 approaches the value @xmath123 gev .
the presence of the sugra term leads to larger values of @xmath124 and hence larger values of @xmath118 for @xmath125 .
for @xmath126 , @xmath127 where @xmath128 is the value of the field at @xmath129 , i.e. @xmath130 .
( note that due to the flatness of the potential the last 55 or so e - folds occur with @xmath131 close to @xmath118 . ) from eqs .
( [ perturb ] , [ potential ] ) , as @xmath132 @xmath133 the denominator of eq .
( [ spert ] ) contains the radiative , sugra and the @xmath102 terms respectively . comparing them
, we see that the radiative term can be ignored for @xmath134 . there is also a soft mass term @xmath135 in the potential , corresponding to an additional term @xmath136 in the denominator .
we have omitted this term , since it is insignificant for @xmath137 . for a positive @xmath102 term ( @xmath138 ) , the maximum value of @xmath77 as a function of @xmath118 is found to be @xmath139 setting @xmath140 , we find a lower bound on @xmath36 ( @xmath141 ) . for larger values of @xmath36 , there are two separate solutions of @xmath118 for a given @xmath36 .
the solution with larger @xmath118 is not valid if the symmetry breaking pattern produces cosmic strings .
for example , strings are produced when @xmath142 break @xmath143 to @xmath144 matter parity , but not when @xmath142 are @xmath145 doublets . for @xmath146 , there are again two solutions , but for the solution with a lower value of @xmath118 , the slope changes sign as the inflaton rolls for @xmath134 and the inflaton gets trapped in a false vacuum .
note that the @xmath102 term depends on @xmath147 , so it should be checked whether @xmath147 changes significantly during inflation .
numerically , we find that it does not , except for a range of @xmath36 around @xmath148 @xcite . for this range , if the initial value of the @xmath35 field is greater than @xmath118 by at least a factor of two or so , the @xmath102 term and the slope become negative even if they were initially positive , before inflation can suitably end .
however , larger values of the @xmath102 term , or the mass term coming from a non - minimal khler potential ( or from a hidden sector vev ) would drive the value of @xmath118 in that region up , allowing the slope to stay positive ( see @xcite for the effect of varying the @xmath102 term and the mass ) . vs. @xmath36 , for susy hybrid inflation with @xmath116 ( solid ) , and for shifted hybrid inflation ( dot - dashed for @xmath149 , dotted for @xmath150 gev ) .
light grey portions of the curves are for @xmath146 , where only the segments that do not overlap with the solutions for @xmath138 are shown .
the grey segments denote the range of @xmath36 for which the change in @xmath147 is significant . ] the dependence of @xmath118 on @xmath36 is shown in fig .
[ fig:21 ] .
note that with inflation linked to the breaking of mssm@xmath151 , @xmath118 corresponds to the @xmath0 breaking scale , which is not fixed by the evolution of the three sm gauge couplings .
the amplitude of the curvature perturbation ( or , equivalently , @xmath20 ) determines this scale to be close to the susy gut scale , suggesting that @xmath0 could be embedded in @xmath31 or its subgroups .
for example , @xmath118 can be determined in flipped @xmath152 from the renormalization group evolution of the @xmath153 and @xmath154 gauge couplings .
the values are remarkably consistent with the ones fixed from @xmath20 considerations @xcite . here
, some remarks concerning the allowed range of @xmath36 is in order . as discussed above , a lower bound on @xmath36
is obtained from the inflationary dynamics and the amplitude of the curvature perturbation .
an upper bound on @xmath36 is obtained from the value of the spectral index , which we discuss next .
the gravitino constraint provides a more stringent upper bound ( @xmath155 ) , as discussed in the next section .
if cosmic strings form ( as would be the case for @xmath116 ) , the range of @xmath36 is also restricted by the limits on the cosmic string contribution to @xmath77 , however most of the range may still be allowed @xcite . in the absence of sugra corrections , the scalar spectral index @xmath51 for @xmath109
is given by @xcite @xmath156 while it approaches unity for small @xmath36 .
when the sugra correction is taken into account , from eq .
( [ v1 ] ) , @xmath157 and it exceeds unity for @xmath158 . for @xmath119 , @xmath159 where @xmath160 . using eq .
( [ nq ] ) , one finds that the spectral index @xmath51 exceeds unity for @xmath161 .
the dependence of @xmath51 on @xmath36 is displayed in fig .
[ fig:22 ] .
@xmath162 is small and the tensor to scalar ratio @xmath16 is negligible , as shown in fig .
[ fig : rdn ] .
the experimental data disfavor @xmath51 values in excess of unity on smaller scales ( say @xmath163 mpc@xmath71 ) , which leads to @xmath164 for @xmath165 . are allowed in models where dissipative effects are significant .
such effects become important for large values of @xmath36 , provided the inflaton also has strong couplings to matter fields @xcite . ]
thus , even for the largest allowed @xmath36 the vacuum energy density during inflation is considerably smaller than the symmetry breaking scale , and the tensor to scalar ratio @xmath166 . vs. @xmath36 , for susy hybrid inflation with @xmath116 ( solid ) , and for shifted hybrid inflation with @xmath149 ( dot - dashed ) .
the grey segments denote the range of @xmath36 for which the change in @xmath147 is significant . ] vs. @xmath36 ( top ) and tensor to scalar ratio @xmath16 vs. @xmath36 ( bottom ) , for susy hybrid inflation with @xmath116 ( solid ) , and for shifted hybrid inflation with @xmath149 ( dot - dashed ) . ] note that the initial wmap analysis suggests a running spectral index , with @xmath167 disfavored at the @xmath168 level @xcite . on the other hand
, an analysis including the constraints from the sloan digital sky survey ( sdss ) finds no evidence for running , with @xmath169 and @xmath170 @xcite . clearly , more data is necessary to resolve this important issue .
modifications of the models discussed here , generally involving two stages of inflation , have been proposed in @xcite and elsewhere to generate a much more significant variation of @xmath51 with @xmath171 .
the inflationary scenario based on the superpotential @xmath38 in eq .
( [ super ] ) has the characteristic feature that the end of inflation essentially coincides with the gauge symmetry breaking .
thus , modifications should be made to @xmath38 if the breaking of @xmath34 to @xmath43 leads to the appearance of topological defects such as monopoles or domain walls .
for instance , the breaking of @xmath172 @xcite to the mssm by fields belonging to @xmath173 , @xmath174 produces magnetic monopoles that carry two quanta of dirac magnetic charge @xcite . as shown in @xcite , one simple resolution of the topological defects problem is achieved by supplementing @xmath38 with a non - renormalizable term : @xmath175 where @xmath176 is comparable to the susy gut scale @xmath177 gev and @xmath178 is an effective cutoff scale .
the dimensionless coefficient of the non - renormalizable term is absorbed in @xmath179 .
the presence of the non - renormalizable term enables an inflationary trajectory along which the gauge symmetry is broken .
thus , in this ` shifted ' hybrid inflation model the topological defects are inflated away .
the inflationary potential is similar to eq .
( [ potential ] ) @xcite : @xmath180+a\,m_{3/2}\kappa v^2|s|{}\ , .
\end{aligned}\ ] ] here @xmath181 where @xmath182 , @xmath183 , and @xmath184 is replaced by @xmath185 in the expression for @xmath186 .. this term , however , does not have a significant effect for the range of parameters where @xmath187 gev . ]
just like in the previous section , for @xmath188 the sugra correction is negligible , and one obtains eq .
( [ quad ] ) with @xmath49 replaced by 2 , eq .
( [ m_kap ] ) with @xmath49 replaced by 4 , and @xmath118 replaced by @xmath189 in both equations .
the slow roll parameters ( and therefore @xmath51 , @xmath190 , and @xmath16 ) are identical to the susy hybrid inflation model with @xmath191 for @xmath192 ( fig .
[ fig:22 ] ) .
the vev @xmath193 at the susy minimum is given by @xcite @xmath194 and is @xmath195 gev depending on @xmath36 and @xmath179 .
the system follows the inflationary trajectory for @xmath196 @xcite , which is satisfied for @xmath137 if the effective cutoff scale @xmath149 . for lower values of @xmath179
, the inflationary trajectory is followed only for higher values of @xmath36 , and @xmath118 is lower for a given @xmath36 ( fig .
[ fig:21 ] ) .
a variation on these inflationary scenarios is obtained by imposing a @xmath197 symmetry on the superpotential , so that only even powers of the combination @xmath198 are allowed @xcite : @xmath199 where the dimensionless parameter @xmath36 is absorbed in @xmath176 .
the resulting scalar potential possesses two ( symmetric ) valleys of local minima which are suitable for inflation and along which the gut symmetry is broken .
the inclination of these valleys is already non - zero at the classical level and the end of inflation is smooth , in contrast to inflation based on the superpotential @xmath38 ( eq . ( [ super ] ) ) .
an important consequence is that , as in the case of shifted hybrid inflation , potential problems associated with topological defects are avoided .
the common vev at the susy minimum @xmath200 . for @xmath201 ,
the inflationary potential is given by @xmath202\,,\ ] ] where the last term arises from the canonical sugra correction .
the soft terms in this case do not have a significant effect on the inflationary dynamics .
if we set @xmath118 equal to the susy gut scale @xmath203 , we get @xmath204 gev and @xmath205 gev .
( note that , if we express eq .
( [ super3 ] ) in terms of the coupling parameter @xmath36 , this value corresponds to @xmath206 . )
the value of the field @xmath207 is @xmath208 gev at the end of inflation ( corresponding to @xmath209 ) and is @xmath210 gev at @xmath129 . in the absence of the sugra correction ( which is small for @xmath211 gev ) ,
@xmath212 , @xmath213 and the spectral index is given by @xcite @xmath214 a value which coincides with the prediction of some d - brane inflation models @xcite .
this may not be surprising since , in the absence of sugra correction , the potential @xmath215 ( eq . ( [ v3 ] ) ) has a form familiar from d - brane inflation .
the sugra correction raises @xmath216 from 0.97 to 1.0 for @xmath217 gev , and above unity for @xmath218 gev ( fig .
[ fig:24 ] ) . as a function of the gauge symmetry breaking scale @xmath118 for smooth hybrid inflation ( dashed line without sugra correction , solid line with sugra correction ) . ]
an important constraint on susy hybrid inflation models arises from considering the reheat temperature @xmath219 after inflation , taking into account the gravitino problem which requires that @xmath220@xmath221 gev @xcite .
this constraint on @xmath9 depends on the susy breaking mechanism and the gravitino mass @xmath100 . for gravity mediated susy breaking models with unstable gravitinos of mass @xmath2221 tev , @xmath223@xmath224 gev @xcite , while @xmath225 gev for stable gravitinos @xcite . in gauge mediated models the reheat temperature is generally more severely constrained , although @xmath226@xmath227 gev is possible for @xmath228100 gev @xcite . finally , the anomaly mediated symmetry breaking ( amsb ) scenario may allow gravitino masses much heavier than a tev , thus accommodating a reheat temperature as high as @xmath221 gev @xcite .
after the end of inflation in the models discussed in section [ chap2 ] , the fields fall toward the susy vacuum and perform damped oscillations about it .
the vevs of @xmath42 , @xmath41 along their right handed neutrino components @xmath229 , @xmath230 break the gauge symmetry .
the oscillating system , which we collectively denote as @xmath231 , consists of the two complex scalar fields @xmath232 ( where @xmath233 , @xmath234 are the deviations of @xmath229 , @xmath230 from @xmath118 ) and @xmath35 , with equal mass @xmath26 .
we assume here that the inflaton @xmath231 decays predominantly into right handed neutrino superfields @xmath235 , via the superpotential coupling @xmath236 or @xmath237 , where @xmath238 are family indices ( see below for a different scenario connected to the resolution of the mssm @xmath12 problem ) .
their subsequent out of equilibrium decay to lepton and higgs superfields generates lepton asymmetry , which is then partially converted into the observed baryon asymmetry by sphaleron effects @xcite .
the right handed neutrinos , as shown below , can be heavy compared to the reheat temperature @xmath9 . note that unlike thermal leptogenesis , there is then no washout factor since lepton number violating 2-body scatterings mediated by right handed neutrinos are out of equilibrium as long as the lightest right handed neutrino mass @xmath239 @xcite .
more precisely , the washout factor is proportional to @xmath240 where @xmath241 @xcite , and can be neglected for @xmath242 . without this assumption , the constraints to generate sufficient lepton asymmetry would be more stringent .
guts typically relate the dirac neutrino masses to that of the quarks or charged leptons .
it is therefore reasonable to assume the dirac masses are hierarchical .
the low - energy neutrino data indicates that the right handed neutrinos in this case will also be hierarchical in general . as discussed in ref .
@xcite , setting the dirac masses strictly equal to the up - type quark masses and fitting to the neutrino oscillation parameters generally yields strongly hierarchical right handed neutrino masses ( @xmath243 ) , with @xmath244 gev .
the lepton asymmetry in this case is too small by several orders of magnitude .
however , it is plausible that there are large radiative corrections to the first two family dirac masses , so that @xmath245 remains heavy compared to @xmath9 .
a reasonable mass pattern is therefore @xmath246 , which can result from either the dimensionless couplings @xmath247 or additional symmetries ( see e.g. @xcite ) .
the dominant contribution to the lepton asymmetry is still from the decays with @xmath248 in the loop , as long as the first two family right handed neutrinos are not quasi degenerate . under these assumptions ,
the lepton asymmetry is given by @xcite @xmath249 where @xmath250 denotes the mass of the heaviest right handed neutrino the inflaton can decay into .
the decay rate @xmath251 @xcite , and the reheat temperature @xmath9 is given by @xmath252 from the experimental value of the baryon to photon ratio @xmath253 @xcite , the required lepton asymmetry is found to be @xmath254 @xcite . using this value , along with eqs .
( [ nls ] , [ reheat ] ) , we can express @xmath9 in terms of the symmetry breaking scale @xmath118 and the inflaton mass @xmath26 : @xmath255 here @xmath26 is given by @xmath256 , @xmath257 and @xmath258 respectively for hybrid , shifted hybrid and smooth hybrid inflation .
the value of @xmath26 is shown in figs .
[ minf ] and [ minf2 ] .
we show the lower bound on @xmath9 calculated using this equation ( taking @xmath259 ev ) in figs .
[ ktr ] , [ sm_mtr ] . vs. @xmath36 , for susy hybrid inflation with @xmath116 ( solid ) , and for shifted hybrid inflation ( dot - dashed for @xmath149 , dotted for @xmath150 gev ) .
the grey segments denote the range of @xmath36 for which the change in @xmath147 is significant . ] vs. the symmetry breaking scale @xmath118 for smooth hybrid inflation .
] vs. @xmath36 , for susy hybrid inflation with @xmath116 ( solid ) and for shifted hybrid inflation ( dot - dashed for @xmath149 , dotted for @xmath150 gev ) .
the segments in the top left part of the figure correspond to the bounds in the presence of a @xmath260 coupling .
the grey segments denote the range of @xmath36 for which the change in @xmath147 is significant . ] vs. the symmetry breaking scale @xmath118 for smooth hybrid inflation . ] eq .
( [ reheat ] ) also yields the result that the heaviest right handed neutrino the inflaton can decay into is about 400 ( 6 ) times heavier than @xmath9 , for hybrid inflation with @xmath261 ( @xmath262 ) . for shifted hybrid inflation ,
this ratio does not depend on @xmath36 as strongly and is @xmath263 @xcite .
this is consistent with ignoring washout effects as long as the lightest right handed neutrino mass @xmath245 is also @xmath264 .
both the gravitino constraint and the constraint @xmath239 favor smaller values of @xmath36 for hybrid inflation , with @xmath265 gev for @xmath91 .
similarly , the gravitino constraint favors @xmath36 values as small as the inflationary trajectory allows for shifted hybrid inflation , and @xmath266 gev for @xmath149 .
smooth hybrid inflation is relatively disfavored since @xmath267 and @xmath268 for @xmath269 gev ( @xmath270 gev ) .
so far we have not addressed the @xmath12 problem and the relationship to @xmath9 in the present context .
the mssm @xmath12 problem can naturally be resolved in susy hybrid inflation models in the presence of the term @xmath271 in the superpotential , where @xmath272 contains the two higgs doublets @xcite .
( the ` bare ' term @xmath273 is not allowed by the @xmath11 r - symmetry . ) after inflation the vev of @xmath35 generates a @xmath12 term with @xmath274 , where @xmath275 is required for the proper vacuum .
the inflaton in this case predominantly decays into higgses ( and higgsinos ) with @xmath276 . as a consequence the presence of this term significantly increases the reheat temperature @xmath9 .
following @xcite , we calculate @xmath9 for the best case scenario @xmath277 .
we find a lower bound on @xmath9 of @xmath278 gev in hybrid inflation , see fig .
@xmath267 gev for shifted hybrid inflation with @xmath149 for shifted hybrid inflation .
some scalars belonging to the inflaton sector acquire negative ( mass)@xmath19 if @xmath279 is smaller .
@xmath280 corresponds to @xmath281 . ] , and @xmath282 gev for smooth hybrid inflation .
an alternative resolution of the @xmath12 problem in susy hybrid inflation involves a peccei - quinn ( pq ) symmetry @xmath283 @xcite .
the lower bounds on @xmath9 are summarized in table 1 . even for hybrid and shifted models , there is some tension between the gravitino constraint and the reheat temperature required to generate sufficient lepton asymmetry , particularly for gravity mediated susy breaking models , and if hadronic decays of gravitinos are not suppressed .
however , we should note that having quasi degenerate neutrinos would increase the lepton asymmetry per neutrino decay @xmath59 @xcite and thus allow lower values of @xmath9 corresponding to lighter right handed neutrinos .
provided that the neutrino mass splittings are comparable to their decay widths , @xmath59 can be as large as @xmath284 @xcite .
the lepton asymmetry in this case is of order @xmath285 where @xmath286 gev for @xmath91 , and sufficient lepton asymmetry can be generated with @xmath9 close to the electroweak scale .
l@r@r & without @xmath260 & with @xmath260 + susy hybrid inflation & @xmath287 & @xmath288 + shifted hybrid inflation & @xmath289 & @xmath290 + smooth hybrid inflation & @xmath290 & @xmath291 + [ ttable ]
* mssm@xmath292 ( matter parity ) , when extended to mssm@xmath293 helps us to implement inflation and leptogenesis ; natural extensions include @xmath31 and its various subgroups ; * the @xmath0 breaking scale is constrained from @xmath20 to be of order the susy gut scale ; * the spectral index is predicted to be @xmath294 , with a negligible tensor to scalar ratio and @xmath17 . for smooth hybrid inflation @xmath15 ;
* @xmath30 contains @xmath13 matter parity , eliminates planck suppressed dimension five proton decay , and may play an essential role in resolving the mssm @xmath12 problem .
this work is partially supported by the us department of energy under contract number de - fg02 - 91er40626 .
r. jeannerot and m. postma , jhep * 0505 * , 071 ( 2005 ) , hep - ph/0503146 .
b. kyae and q. shafi , hep - ph/0510105 .
m. bastero - gil and a. berera , phys .
rev . * d71 * , 063515 ( 2005 ) , hep - ph/0411144 .
u. seljak _ et al .
_ , phys .
rev . * d71 * , 103515 ( 2005 ) , astro - ph/0407372 . ,
* * , ( ) , .
|
a gauged @xmath0 symmetry predicts three right handed handed neutrinos and its spontaneous breaking automatically yields the seesaw mechanism . in a supersymmetric setting this breaking can be nicely linked with inflation to yield @xmath1 proportional to @xmath2 , where @xmath3 @xmath4 denote the @xmath5 breaking ( planck ) scale .
thus @xmath3 is estimated to be of order @xmath6 gev , and the heaviest right handed neutrino mass is less than or of order @xmath7 gev .
a second right handed neutrino turns out to have a mass of order @xmath8 , where @xmath9 ( @xmath10 gev ) denotes the reheat temperature .
a @xmath11 r symmetry plays an essential role in implementing inflation and leptogenesis , resolving the mssm @xmath12 problem and eliminating dimension five nucleon decay .
an unbroken @xmath13 subgroup plays the role of matter parity .
the scalar spectral index @xmath14 for the simplest models , while in smooth hybrid inflation @xmath15 .
the tensor to scalar ratio @xmath16 is negligible , and @xmath17 .
|
the statistical relation between recorded crime and economic and demographic variables , such as unemployment and the number of young males , is not only of fundamental importance @xcite , but becomes increasingly important in forecasting tools for policy and planning decisions @xcite . as recorded crime and its predictor variables
are generally integrated of order one , two kinds of models exist .
there are short - run models specified in first differences and cointegrated models specified in levels . in a short - run model , crime has a strong stochastic dynamic of its own , eventually drifting away uncontrollably from the relation with its predictors .
conversely , in a cointegrated model , crime fluctuates with constant variability around the relation with its predictors , thereby preserving its importance in the long run . as a consequence ,
cointegrated models are to be favored over short - run models when making long term forecasts . theoretically , given a set of predictor variables
, one can not argue conclusively whether a cointegrated model applies .
empirically , however , a time series analysis indicates whether a cointegrated model is supported by the data or not @xcite .
for example , no cointegration was found between crime and unemployment in england and wales @xcite , while a cointegrated relation was established between robbery rate and divorce rate in the united states @xcite .
the objective of this work is to go beyond a time series analysis for a fixed set of predictor variables .
rather , we consider all subsets of a set of potential predictor variables and search for cointegrated models .
the kind of crime under investigation is violent crime in the netherlands , composed of robbery , assault , homicide , rape and extortion .
the set of potential predictor variables consists of unemployment , the number of males aged 15 - 25 years of western background , the number of males aged 15 - 25 years of non - western background , the number of divorces and deflated consumption .
a manual inspection of all subsets is not feasible because the number of subsets , and thereby the number of candidate models , grows exponentially with the number of potential predictor variables .
even with five potential predictor variables and allowing for a potential error correction parameter , a constant or a constant and a trend , nearly two hundred candidate models exist . to investigate these models
, we developed a modelgenerator that automatically builds and estimates them and subsequently checks for cointegration .
the cointegration check consists of the first step of the engle - granger ( eg ) procedure @xcite .
the generator discards models that are not cointegrated and then estimates the cointegrated models using nls regression .
finally , the generator orders the cointegrated models using the information criterions aic @xcite and bic @xcite .
a necessary condition for the existence of a cointegrated relation is that all variables involved are integrated of the same non - vanishing order .
adf tests @xcite indicate that recorded violent crime and its potential predictor variables do not reject one or more null hypotheses of a unit root up to the 0.10 level .
these tests were performed on annual data ranging from 1978 to 2006 and the detailed results are listed in table [ adfkpss ] .
we also performed kpss tests @xcite that reverse the null hypothesis ( a unit root ) and the alternative hypothesis ( stationarity ) of the adf tests .
the kpss results are included in table [ adfkpss ] .
the probabilities of @xmath0 having a unit root and of @xmath0 being stationary are all rather small ( see table [ adfkpss ] for the symbolic notation of the variables ) .
further analysis reveals that @xmath0 is more likely to be integrated of order two than of order one .
however , based on the joint picture emerging from table [ adfkpss ] , we conclude that all variables ( including @xmath0 ) are integrated of order one .
having met the necessary condition , the following cointegrated model may exist @xmath1 here , @xmath2 denotes discrete time measured in years , @xmath3 is a non - empty subset of the integer set @xmath4 , @xmath5 is the coefficient connecting @xmath6 to @xmath7 and @xmath8 is gaussian white noise . the operator @xmath9 is the backshift operator retarding a variable with one year and the autoregressive noise @xmath10 is stationary and causal with correlation length @xmath11 measured in years .
model ( [ boxjenkins ] ) may be augmented with a constant @xmath12 or a constant and a trend @xmath13 .
.results of adf tests and kpss tests on recorded violent crime and its potential predictor variables .
the lag length of the adf tests is selected with bic .
the numbers in the adf columns are the probabilities of the variable having a unit root .
the subcolumns ` 0 ' , ` c ' and ` ct ' correspond to a random walk around , respectively , zero , a constant and a constant and a trend .
the probabilities are based on finite - sample distributions of the df statistic @xcite of ref .
the kpss tests are performed with the bartlett kernel and the data - based newey - west bandwidth selection procedure @xcite .
the numbers in the kpss subcolumns ` c ' and ` ct ' are the probabilities of the variable being stationary around , respectively , a constant and a constant and a trend .
these probabilities are based on the asymptotic test statistics of ref .
[ cols="<,<,^,^,^,^,^ " , ] [ shortrunest ] the predictive power of short - run models i and iii is compared to that of cointegrated model ii by in - sample forecasts .
the outcome of the comparison is not known beforehand : @xmath14 may be a good predictor of @xmath7 and the short - run models can outperform the cointegrated model when making short term forecasts .
we make a long ( short ) term forecast , estimating the models using data from 1978 up to 1995 ( 2000 ) and forecasting them from 1996 ( 2001 ) to 2006 .
the forecasts are based on a monte carlo simulation of @xmath15 repetitions ( including coefficient uncertainty ) and consist of mean forecasts and error bounds .
the results for short - run model i and the cointegrated model are indicated in fig .
[ forecasts ] .
panel ( a ) holds the long term forecasts and indicates that the cointegrated model outperforms the short - run model in the sense that its mean predictions are closer to the realizations and that its error bounds are smaller . in panel ( b ) , holding the short term forecasts , the mean forecasts of the models are very similar , but the error bounds of the short - run model are larger .
the forecasts of short - run model iii are similar to that of short - run model i. in panel ( a ) , short - run model iii would produce mean forecasts a bit worse than that of short - run model i. in panel ( b ) , it would produce mean forecasts a bit closer to the realizations than that of short - run model i and the cointegrated model .
the error bounds of short - run model iii are comparable to that of short - run model i.
we have developed a modelgenerator that searches for cointegrated models among candidate models specified in levels .
the generator employs the first step of the eg procedure to decide whether a model is cointegrated or not .
subsequently , it estimates the cointegrated models with nls regression and checks the residuals with the bg lm test .
finally , it orders models according to both aic and bic .
the generator is applied to recorded violent crime and five potential predictor variables : unemployment , the number of males aged 15 - 25 years of western background , the number of males aged 15 - 25 years of non - western background , the number of divorces and deflated consumption .
the pattern emerging from the ordered list of cointegrated models is that the number of males aged 15 - 25 years of non - western background and deflated consumption are the key predictors of recorded violent crime .
based on an analysis of vec models , a plausibility argument and the statistical significance of estimated parameters , we select a cointegrated model out of the ordered list .
the selected model holds , in addition to the key predictors , the number of males aged 15 - 25 years of western background .
with the cointegration check switched off , the generator is used to find short - run models specified in first differences .
the key predictors of the short - run models are the number of males aged 15 - 25 years of non - western background and the number of divorces . based on a plausibility argument and the statistical significance of estimated parameters ,
two short - run models are selected from the ordered list .
in addition to the key predictors , one of the short - run models holds deflated consumption and the other unemployment . in -
sample forecasts reveal that the cointegrated model outperforms the two short - run models in both mean predictions ( long term ) and error bounds ( long term and short term ) .
( solid black lines ) and short - run model i of sec . [ shortrunmodels ] ( solid gray lines ) .
the dashed black ( gray ) lines constitute error bounds of two standard deviations above and below the mean forecasts of the cointegrated ( short - run ) model . in panel
( a ) , the models are estimated based on realizations from 1978 up to 1995 and forecasted from 1996 up to 2006 . in panel
( b ) , the models are estimated based on realizations from 1978 up to 2000 and forecasted from 2001 up to 2006.,title="fig:",width=340 ] ( solid black lines ) and short - run model i of sec .
[ shortrunmodels ] ( solid gray lines ) .
the dashed black ( gray ) lines constitute error bounds of two standard deviations above and below the mean forecasts of the cointegrated ( short - run ) model . in panel
( a ) , the models are estimated based on realizations from 1978 up to 1995 and forecasted from 1996 up to 2006 . in panel
( b ) , the models are estimated based on realizations from 1978 up to 2000 and forecasted from 2001 up to 2006.,title="fig:",width=340 ] c. hale and d. sabbagh , journal of research in crime and delinquency * 28 * , 400 ( 1991 ) . c. l. britt , american journal of economics and sociology * 53 * , 99 ( 1994 ) .
d. f. greenberg , journal of quantitative criminology * 17 * , 291 ( 2001 ) .
r. harries , international journal of forecasting * 19 * , 557 ( 2003 ) .
d. deadman , international journal of forecasting * 19 * , 567 ( 2003 ) .
r. f. engle and c. w. j. granger , econometrica * 55 * , 251 ( 1987 ) .
s. johansen , journal of economic dynamics and control * 12 * , 231 ( 1988 ) .
h. akaike , in _
2nd international symposium on information theory _ , edited by b. n. petrov and f. csaki ( akademia kiado , 1973 ) .
g. schwarz , annals of statistics * 6 * , 461 ( 1978 ) .
s. e. said and d. a. dickey , biometrika * 71 * , 599 ( 1984 ) .
d. kwiatkowski , p. c. b. phillips , p. schmidt and y. shin , journal of econometrics * 54 * , 159 ( 1992 ) .
d. a. dickey and w. a. fuller , journal of the american statistical association * 74 * , 427 ( 1979 ) .
j. g. mackinnon , journal of applied econometrics * 11 * , 601 ( 1996 ) .
w. k. newey and k. d. west , review of economic studies * 61 * , 631 ( 1994 ) .
d. cantor and k. c. land , american sociological review * 50 * , 317 ( 1985 ) .
g. e. p. box and g. m. jenkins , _ time series analysis : forecasting and control _
( holden - day , 1976 ) .
t. hirschi and m. gottfredson , american journal of sociology * 89 * , 552 ( 1983 ) .
d. j. steffensmeier , e. a. allan , m. d. harer and c. streifel , american journal of sociology * 94 * , 803 ( 1989 ) .
t. s. breusch , australian economic papers * 17 * , 334 ( 1978 ) . l. g. godfrey , econometrica * 46 * , 1293 ( 1978 ) . c. m. hurvich and c. -l .
tsai , biometrika * 76 * , 297 ( 1989 ) . j. g. mackinnon , a. a. haug and l. michelis , journal of applied econometrics * 14 * , 563 ( 1999 ) .
|
a modelgenerator is developed that searches for cointegrated models among a potentially large group of candidate models .
the generator employs the first step of the engle - granger procedure and orders cointegrated models according to the information criterions aic and bic .
assisted by the generator , a cointegrated relation is established between recorded violent crime in the netherlands , the number of males aged 15 - 25 years ( split into western and non - western background ) and deflated consumption . in - sample forecasts reveal that the cointegrated model outperforms the best short - run models .
|
among the various subjects concerning the physics of polymers , the behavior of heteropolymers containing hydrophobic and hydrophilic components in solvents are of particular interest since they have obvious importance in biology @xcite .
it is well known for instance that in a polar solvent , these heteropolymers prefer conformations where the hydrophilic components are in contact with the polar solvent , whereas hydrophobic components avoid contacts with the solvent . the behavior of heteropolymers in the presence of an interface separating two selective solvents , one favorable to the hydrophobic components and the other to the hydrophilic components , is less obvious , and has been much studied recently . in the pioneering work of garel , huse , leibler and orland @xcite ,
a model was proposed and studied via imry - ma arguments , an analysis of the replica hamiltonian and numerics : it was found that a chain with a symmetric distribution in hydrophobic / hydrophilic components is always localized around the interface at any temperature ( in the thermodynamic limit ) , whereas a chain with a dissymmetric distribution in hydrophobic / hydrophilic components presents a phase transition separating a localized phase at low temperatures from a delocalized phase into the most favorable solvent at high temperatures .
experimentally , the presence of copolymers was found to stabilize the interface between the two immiscible solvents @xcite , since the localization of the heteropolymers at the interface reduces the surface tension . by now , the predictions of reference @xcite have been confirmed in the physics community by various approaches including molecular dynamics simulations @xcite , monte carlo studies @xcite , variational methods for the replica hamiltonian @xcite @xcite , and exact bounds for the free - energy @xcite .
mathematicians have also been interested in this model , but the exact results obtained up to now are still far from a complete explicit solution .
the localization at all temperatures for the symmetric case was proven in references @xcite @xcite by various criteria of localization on the paths measure . in the dissymmetric case ,
the existence of a transition line in temperature vs dissymmetry plane was proven in reference @xcite where some bounds for the free - energy were obtained in the asymptotic regimes of low and high temperatures .
finally , the relations between various localization criteria concerning either free - energy or path properties have been studied in reference @xcite . in this paper
, we propose a new approach to study the model of garel et al .
@xcite , based on a disorder - dependent real space renormalization procedure . since this type of approach is not usual in the context of the thermodynamics of classical disordered systems , it seems useful to recall how these methods have been successfully applied in other domains .
disordered - dependent real space renormalization approaches have first been introduced in the field of disordered quantum spin chains @xcite @xcite @xcite @xcite @xcite .
more recently they have been used to study random walkers in 1d disordered environments @xcite @xcite , reaction - diffusion in 1d disordered environments @xcite and non - equilibrium dynamics of disordered classical spin chains @xcite @xcite .
a similar method has also been introduced independently to study the coarsening of the pure one - dimensional @xmath2 model at zero temperature @xcite . in all these fields , these methods , which are not exact from first principles " , have however been remarkably successful in reproducing the exact results that were already known ( see references @xcite @xcite ) , and in producing a lot of novel exact results for exponents and scaling functions for a large variety of physical quantities .
these results apply in the large renormalization scale regime , corresponding respectively to low - temperature in the field of disordered quantum spin chains , and to large time behaviors in the other fields concerning dynamics .
these new predictions have moreover been confirmed numerically whenever they have been tested @xcite @xcite @xcite . for the random heteropolymer problem , the disorder - dependent real space renormalization procedure that we propose in this paper
is defined to select the important " configurations for the thermodynamics of a chain with a given realization of the disordered sequence in hydrophobic / hydrophilic components at a given temperature .
we expect this renormalization approach to become accurate in the large renormalization scale regime , corresponding here to the region of high temperatures .
since the localized phase extends up to a critical temperature @xmath3 that goes to infinity as the dissymmetry parameter of the distribution in hydrophobic / hydrophilic components goes to zero , our approach allows to study the localized phases of symmetric chains and slightly dissymmetric chains .
the paper is organized as follows . in the remainder of the introduction , we recall the model of garel et al .
@xcite in section [ model ] , and give a summary of the main results of the present paper in section [ summary ] .
in section [ physical ] , we recall the physics of the heteropolymer problem at @xmath4 and at high temperatures , in order to motivate the introduction of a disorder - dependent real - space renormalization procedure in section [ rg ] . the results of the renormalization approach are given for symmetric and dissymmetric chains in sections [ symmetric ] and [ biased ] respectively . in section [ finite ]
, we characterize the finite - size properties of the problem by considering chains of finite ( large ) length . in section [ dynamics ] , we discuss how the renormalization procedure also describes the non - equilibrium dynamics of a random chain at high temperature starting at @xmath5 from a zero - temperature initial condition .
finally section [ conclusion ] contains the conclusions . in the appendix [ appendiximryma ] , we compare some results of the renormalization approach with the corresponding results obtained by garel and orland via an extension of the usual imry - ma argument @xcite .
we consider the model introduced by garel et al .
@xcite with some slightly different notations : a polymer chain consists of @xmath0 monomers indexed by @xmath6 , where each monomer @xmath7 carries a quenched random charge @xmath8 .
the @xmath9 are independent identical random variables drawn with some probability distribution @xmath10 .
an interface at @xmath11 separates a solvent in the domain @xmath12 favorable to positive charges @xmath13 , from a solvent in the domain @xmath14 favorable to negative charges @xmath15 .
more precisely , the energy of a configuration @xmath16 of the heteropolymer characterized by a realization @xmath17 reads @xmath18 and the partition function reads @xmath19 in a continuum version of this model , the monomer index @xmath7 becomes a continuous variable @xmath20 , and the @xmath21 coordinates play no role so that the partition function becomes a path integral over one - dimensional brownian trajectories @xmath22
@xcite @xmath23 in this paper , we propose a disorder - dependent real - space renormalization procedure , that should give valid results in the universal regime of large scales , where the details of the microscopic charge distribution @xmath10 are not important , as long as the distribution @xmath10 satisfy the hypothesis of the central limit theorem .
the important parameters of @xmath10 will thus be the mean value and the variance @xmath24 to study the biased case @xmath25 , we will always choose the convention @xmath26 and introduce the parameter @xmath27 defined as the non - vanishing root of the equation @xcite @xmath28 for instance , in the case of a gaussian distribution considered in references @xcite @xcite @xcite @xcite @xmath29 the parameter @xmath27 is simply the ratio between the mean value @xmath30 and the variance @xmath31
@xcite @xmath32 for the case of binary distribution considered in references @xcite @xcite @xcite @xcite where the charge @xmath33 takes a positive value @xmath34 with probability @xmath35 and a negative value @xmath36 with probability @xmath37 then @xmath38 and @xmath27 is the solution of @xmath39 for instance in the case @xmath40 , we have @xmath41 an important property of the parameter @xmath27 is that it becomes universal and coincides with the gaussian expression in the limit of small dissymmetry @xmath42 @xcite @xmath43 concerning the spatial behavior of the polymer , the only important parameter will be the diffusion coefficient @xmath44 characterizing the large distance behavior of the free chain @xmath45 we stress that in this paper , we do not take into account excluded volume effects because they do not have any crucial effect for the problem of the localization of the heteropolymer at the interface @xcite @xcite @xcite .
the quantities of interest to characterize the localized phase of heteropolymers around the interface are on one hand the free - energy @xmath46 per monomer that governs the thermodynamic , and on the other hand the statistical properties of the spatial conformations of the chain . in particular , the important
characteristic lengths are the typical lengths @xmath47 of blobs in the @xmath48 solvents as well as the typical distances @xmath49 between the chain and the interface . at a more refined level
, one may also consider the probability distributions @xmath50 of the lengths of @xmath48 blobs , the density @xmath51 of polymer at a distance @xmath52 from the interface in the @xmath48 solvents , and even correlation functions as for instance the solvent - solvent correlation @xmath53 .
finally , we will also be interested in the finite size effects of the problem . before introducing our approach and proceeding with the calculations of these quantities of interest
, we now summarize the main results that will be derived from the real - space renormalization approach in the remainder of the paper .
for the case of symmetric chains ( @xmath54 ) , the renormalization approach yields the following behaviors at high temperature @xmath55 in agreement with the imry - ma argument results @xcite ( see also section [ hight ] for the discussion of the presence of the logarithmic correction ) .
in addition , we obtain that the rescaled length of blobs @xmath56 is distributed with the law @xmath57 the polymer density takes the scaling form @xmath58 where the rescaled distance to the interface @xmath59 is distributed with the following law @xmath60 that decays exponentially at large distance @xmath61 finally , to characterize the finite size effects , we will compute the distribution of the delocalization temperature @xmath62 over the ensemble of random cyclic finite chains of ( large ) length @xmath0 .
the result is that the rescaled variable @xmath63 is distributed with the law @xmath64 e^{- ( 2m+1)^2 g^2 } \opsimeq_{g \to \infty } \frac{4}{\sqrt{\pi } } g^2 e^{- g^2 } \qquad .\end{aligned}\ ] ] for the case of dissymmetric chains ( @xmath26 ) , in the limit @xmath66 , the renormalization approach yields that the delocalization transition takes place at the critical temperature @xmath67 in agreement with the scaling obtained previously by other methods @xcite
@xcite @xcite @xcite . in addition
, we obtain that the transition is of infinite order , with the following essential singularity for the free energy the free - energy @xmath46 @xmath68 \qquad .\end{aligned}\ ] ] for the statistical properties of the heteropolymer chain , we find that the typical length @xmath69 of blobs in the preferred solvent diverges with an essential singularity at the transition , whereas the typical length @xmath70 in the other solvent diverges algebraically @xmath71 \\ & & l^{-}_{blob}(t ) \opsimeq_{t \to t_c^- } \frac{\sigma}{q_0 ^ 2 } \frac { \ln \frac{4 \sigma}{q_0 } } { \left(1-\frac{t}{t_c } \right ) } \end{aligned}\ ] ] the rescaled length of blobs in the preferred solvent , defined as @xmath72 } \end{aligned}\ ] ] is distributed with the exponential law @xmath73 near the transition .
the polymer density @xmath74 takes the scaling form @xmath75 where the rescaled distance to the interface @xmath76 } \qquad \end{aligned}\ ] ] is distributed with the following scaling distribution that decays exponentially at large distance @xmath77 finally , to characterize the finite size effects , we will compute the distribution of the delocalization temperature @xmath62 over the ensemble of random cyclic finite chains of ( large ) length @xmath0 .
the result is that the random variable @xmath78 is distributed with the law @xmath79 in particular , the typical value for the delocalization temperature presents a correction of order @xmath80 with respect the critical temperature @xmath3 @xmath81
in this section , we recall the physics of the heteropolymer problem at @xmath4 and the imry - ma picture of reference @xcite valid at high temperature
. they will be useful to motivate the introduction of the real - space renormalization procedure in the next section . at @xmath4 , for a given realization @xmath17 , each monomer @xmath7 will be in its preferred solvent according to @xmath82 the ground states have thus for energy @xmath83 the corresponding configurations of the chain are all the random walks that cross the interface each time there is a change of signs in the realization @xmath17 .
more precisely , it is convenient to decompose the realization @xmath17 into groups @xmath84 of consecutive charges of the same sign , each group @xmath84 containing a number @xmath85 of monomers , and carrying an absolute charge @xmath86 with @xmath87 then the ground states of the chain can be decomposed into blobs @xmath88 containing @xmath89 monomers .
the chain is thus localized around the interface with a typical distance of order @xmath90 . in this section ,
we recall the imry - ma arguments @xcite of garel et al .
@xcite , for the symmetric and dissymmetric cases respectively .
they are expected to be valid at high temperature .
the imry - ma argument exposed in @xcite for the symmetric case ( @xmath54 ) can be summarized as follows .
assuming that the chain is localized around the interface with typical blobs of length @xmath91 in each solvent , the typical energy gain per blob scales as @xmath92 , whereas the reduction of entropy per blob scales as @xmath93 ( in @xcite , only powers of @xmath91 were considered for the symmetric case , and thus @xmath93 was replaced by @xmath94 , but since this @xmath93 dependence plays a crucial role in the dissymmetric case , we prefer to keep this @xmath93 dependence everywhere in this paper ) . neglecting prefactors of order 1
, the optimization of the free - energy per monomer @xmath95 with respect to @xmath91 leads to @xmath96 at high temperature , the scaling of the typical blob length @xmath97 and of the free - energy @xmath46 are therefore given by : @xmath98 and thus the heteropolymer remains localized at any finite temperature .
the imry - ma argument exposed in @xcite for the dissymmetric case is actually much more subtle than in the symmetric case , because to describe correctly the blobs in the @xmath99 solvent , it is necessary to consider the rare events " where the sum of random variables @xmath8 of positive mean @xmath100 happens to be negative enough in order to make more favorable for the heteropolymer to make an excursion in the @xmath99 solvent rather than to stay in the otherwise preferred @xmath101 solvent .
more precisely , the imry - ma argument of reference @xcite is as follows in our notations : at high temperature , the heteropolymer is expected to be mostly in the preferred @xmath101 solvent , except when a blob of length @xmath102 in the @xmath99 solvent becomes energetically favorable , with a blob energy @xmath103 that is sufficient " . the probability to have @xmath104
is given by the gaussian @xmath105 thus the typical spacing @xmath106 between two such events behaves as the inverse of this small probability @xmath107 as a consequence , in a configuration with blobs of order @xmath108 with @xmath109 , the energy gain @xmath110 due to one excursion in the @xmath99 solvent behaves as @xmath111 whereas the loss of entropy due to this excursion is of order @xmath112 .
the free - energy difference per monomer between this localized state and the delocalized state in the preferred solvent can then be estimated as ( neglecting prefactors of order 1 ) : @xmath113 the optimization with respect to @xmath102 leads to @xmath114 and @xmath115 thus both the energy gain @xmath110 and the entropy cost have the same @xmath116 dependence ! as a consequence ,
the free - energy difference factorizes into @xmath117 where the parameter @xmath118 thus represents the critical temperature between the localized phase @xmath119 and the delocalized phase @xmath120 .
contrary to the symmetric case , the behavior of the free - energy with respect to the temperature remains unknown , since the typical blob length @xmath106 depends upon the temperature t in a way that is not determined by the present argument .
still , the relations ( [ relationlmlp ] ) and ( [ relationflp ] ) between the typical blobs lengths @xmath106 and @xmath102 and the free - energy @xmath46 represent non - trivial tests for any theory constructed to describe the heteropolymer problem .
to go beyond the imry - ma arguments of previous section , one needs to consider probability distributions of blob lengths and blob energies . in appendix
[ appendiximryma ] , we reproduce an extension of the usual imry - ma argument which was introduced by garel and orland @xcite , and which shows how to get the asymptotic behavior of the probability distribution @xmath121 of the blob length @xmath91 in the limit of small @xmath91 . here , to study the probability distributions of blob lengths and blob energies , we will construct via a real - space renormalization procedure the optimal " imry - ma domain structure associated with a given heteropolymer at a given temperature .
we will then argue that the configurations corresponding to this optimal imry - ma domain structure dominate in the partition function asymptotically at high temperature .
at @xmath4 , we have seen in section [ ground ] that the chain is decomposed into blobs @xmath88 containing @xmath122 consecutive monomers of charges of the same sign , and carrying an absolute charge @xmath86 . as @xmath1 grows from @xmath4 , we consider the configurations of the chain obtained from the ground states structure by iteratively flipping the blobs of smallest absolute charge @xmath123 .
when we flip the blob @xmath124 surrounded by the two neighboring blobs @xmath125 and @xmath126 , we obtain a new blob of absolute charge @xmath127 and length @xmath91 given by ( see figure [ figrg ] ) @xmath128 1.0 cm = 10.0 true cm 1.0 cm the renormalization procedure corresponding to the rules ( [ rgrules ] ) has already been extensively studied in the context of the random transverse field ising chain @xcite , where the role of the absolute charges @xmath129 was played by the logarithm of the random couplings and random fields ( @xmath130 and @xmath131 ) , and in the context of random walks in disordered 1d environments @xcite where the role of the absolute charges @xmath129 was played by the energy barriers @xmath132 . as explained in details in reference @xcite , the renormalization procedure constructs iteratively the large scale extrema statistics of random walks , where only barriers bigger than a given scale @xmath133 are kept . here
the corresponding random walk is simply given by the sum of the quenched charges @xmath134 as a function of the monomer index @xmath7 , and the renormalization procedure gives the best " blob structure , given the constraint that only blobs of charges bigger than a given scale @xmath133 are kept . to establish a relation between the renormalization scale @xmath133 and the temperature @xmath1
, we now have to determine the conditions under which the flip of the blob @xmath135 is favorable .
the cost in energy of the flip of the blob @xmath135 is simply @xmath136 whereas the corresponding gain in entropy reads @xmath137 \qquad , \end{aligned}\ ] ] where @xmath138 represents the number of 1d random walks of @xmath91 steps going from @xmath11 to @xmath11 in the presence of an absorbing boundary at @xmath139 . as a consequence ,
the free - energy difference corresponding to this flip reads @xmath140 the optimal imry - ma domain structure corresponding to a given realization of the disorder at a given temperature is thus constructed as follows : we iteratively flip the blobs of smallest absolute charge @xmath127 , starting from the ground - states blob structure , and continue as long as these flippings produce a decrease of the free - energy ( @xmath141 ) .
we stop the procedure when the next flipping would correspond to a raise of free - energy ( @xmath142 ) .
we will define @xmath143 as the renormalization scale @xmath133 where we have to stop the renormalization .
we now present a dynamical interpretation of the renormalization procedure . the renormalization procedure defined above to construct the optimal blob structure at temperature @xmath1
can be given a perhaps more direct physical meaning if one considers the dynamics at high temperature @xmath1 for @xmath144 starting from a zero - temperature initial condition at @xmath5 , i.e. after a quench to @xmath4 for @xmath145 .
indeed , the time necessary to flip a blob of absolute charge @xmath127 follows an arrhenius law @xmath146 ( where @xmath147 is a microscopic time scale ) .
the dynamics thus corresponds to the iterative flipping of the blobs of smallest absolute charge remaining in the chain described by the renormalization procedure , where the renormalization scale @xmath133 now corresponds to time via @xmath148 as in references @xcite @xcite .
the dynamics takes place up to time @xmath149 where equilibrium at temperature @xmath1 is reached @xmath150 let us define the effective partition function @xmath151 as the sum over all the configurations that correspond to the disorder and temperature dependent optimal imry - ma domain structure constructed via the renormalization procedure defined above .
the real partition function @xmath152 can be decomposed into @xmath153 the effective thermodynamics based on @xmath151 is of course very approximate at low temperatures .
however , at high temperatures , the renormalization scale @xmath154 will be large and the probability distributions of absolute charges of the blobs will become infinitely broad @xcite .
more precisely in the symmetric case , the difference between the typical value @xmath155 of the absolute charges and the minimum value @xmath156 will also be large and of order @xmath133 : @xmath157 . as a consequence
, we expect that all the configurations that do not correspond approximatively to the optimal imry - ma domain structure will be highly suppressed in the partition function , in comparison with the optimal configurations " .
the validity of the effective thermodynamics is thus based on the idea that , for a given heteropolymer at a given high temperature , the large scale imry - ma domain structure is essentially _
unique_. as explained in the introduction , in the other fields where disordered - dependent real space renormalization approaches have been used , they have been able to reproduce the non - trivial exact results obtained previously via other methods ( see references @xcite @xcite ) .
these agreements with the mc coy - wu exact results in the field of disordered quantum spin chains @xcite , and with the kesten - golosov distribution in the field of 1d random walks in random media @xcite have provided the best evidence for the asymptotic exactness of the renormalization approach in these fields , and give confidence in all its other novel predictions .
here for the heteropolymer problem , we do not have at our disposal so precise exact analytical results derived via other methods to test " unambiguously the asymptotic exactness of the predictions of the renormalization approach . as a consequence ,
we hope that in the future , the explicit results of the renormalization approach presented in this paper will be tested precisely by numerical studies
. it would be also very interesting to test more directly the idea of the dominance of the optimal imry - ma domain structure in the partition function , by comparing , sample by sample , the typical configurations of the heteropolymer at equilibrium at temperature @xmath1 with the optimal blob structure obtained by the numerical implementation of the renormalization procedure up to scale @xmath143 .
the remainder of the paper is devoted to the detailed study of the properties of the effective thermodynamics . to proceed now with the concrete calculations of the renormalization approach , it is again more convenient to discuss separately the symmetric case @xmath54 and the dissymmetric case @xmath158 .
it was shown in reference @xcite that with the above renormalization procedure ( [ rgrules ] ) , the probability distribution @xmath159 that a blob at renormalization scale @xmath133 has absolute charge @xmath127 and length @xmath91 flows towards a fixed point distribution @xmath160 in the rescaled variables @xmath161 with @xmath162 defined in ( [ defq ] ) , and that this fixed point distribution @xmath160 reads in laplace transform with respect to @xmath163 @xmath164 in particular , the distribution of the rescaled absolute charge @xmath165 is a simple exponential @xmath166 and the distribution of the rescaled length @xmath163 of a blob reads @xmath167 we also introduce the notation @xmath168 for the distribution of the bonds which are about to be decimated ( i.e. having @xmath169 ) @xmath170 e^{- \frac{1}{\lambda } ( m+\frac{1}{2})^2 } \qquad .\end{aligned}\ ] ] we now examine the condition @xmath171 , where the free - energy difference @xmath172 due to a blob flip was given in equation ( [ flipcondition ] ) . for large @xmath91 , the number @xmath138 of 1d random walks of @xmath91 steps going from @xmath11 to @xmath11 in the presence of an absorbing boundary at @xmath139 behaves as @xmath173 where @xmath174 is a constant , and where @xmath175 characterizes the entropy of the free random walk ( for instance @xmath176 for the 1d lattice ) . in the effective thermodynamics at large renormalization scale @xmath133
, the free - energy difference of a blob flip then reads @xmath177 using the rescaled variables @xmath178 , we obtain the average @xmath179 where @xmath180 is a pure numerical constant that can be obtained from the distributions ( [ pdel ] ) and ( [ pdeci ] ) as @xmath181 the scale @xmath143 where the renormalization procedure has to be stopped is the solution of @xmath182 and @xmath183 . for large @xmath133 , and
thus for large temperature , we find that @xmath143 is implicitly defined as the inverse of the function @xmath184 given by @xmath185 as a consequence , for any arbitrary large temperature @xmath1 , we find that the renormalization procedure has to be stopped at the _ finite _ large renormalization scale @xmath143 .
the polymer is thus always localized around the interface , with blobs of typical length @xmath186 behaving at high temperature as @xmath187 in this regime , the typical distance @xmath188 to the interface behaves as @xmath189 these scaling behaviors are in agreement with the imry - ma argument of reference @xcite ( see section [ hight ] , where the logarithmic correction is found to be present in the imry - ma argument ) and with the exact free energy bounds of reference @xcite .
the replica variational approaches of references @xcite @xcite give the same scaling behaviors without the logarithmic correction .
we now study more precisely various physical quantities for which the present renormalization approach yields explicit predictions .
the energy of a blob is simply given by minus its absolute charge @xmath190 , and thus the energy per monomer @xmath191 of the chain in the thermodynamic limit can be obtained through a decomposition into blobs indexed by @xmath88 as @xmath192 with @xmath143 defined in ( [ gammaeqt ] ) . using the fixed point solution of equations ( [ rescaling ] , [ solu ] ) , we obtain at large temperature @xmath193 the entropy associated with the @xmath52 direction for a blob of length @xmath91 is given by @xmath194 , where @xmath138 behaves as in equation ( [ mlasymp ] ) in the large @xmath91 regime .
as a consequence , the entropy per monomer @xmath195 is given in the thermodynamic limit by @xmath196 } { \sum_{\alpha } l_{\alpha } } = \frac{\int_{q , l } [ l \ln \mu + \ln \kappa -\frac{3}{2 } \ln l ] p_{\gamma_{eq}(t)}(q , l ) }
{ \int_{q , l } l p_{\gamma_{eq}(t)}(q , l ) } \qquad .\end{aligned}\ ] ] using again the fixed point solution ( [ rescaling ] , [ solu ] ) , we have at large temperature @xmath197 finally , we obtain the free - energy per monomer @xmath46 of the chain at large temperature @xmath198 the dominant behavior depending on the disorder parameter @xmath162 is in agreement with the imry - ma argument of reference @xcite as explained in section [ hight ] . in the renormalization approach ,
the lengths of the different blobs are independent identical random variables , and the rescaled blob length defined as @xmath199 is asymptotically distributed at high temperature with the distribution @xmath200 given in equation ( [ pdel ] ) . in particular , its asymptotic behaviors for large and small @xmath163 are respectively given by @xmath201 @xmath202 as noted by garel and orland @xcite , the small @xmath163 behavior found in equation ( [ pdelzero ] ) for the distribution @xmath203 can be obtained via an extension of the usual imry - ma argument , as explained in the appendix ( [ appendiximryma ] ) .
we now study the density @xmath204 of polymer at a distance @xmath52 from the interface as follows .
we first need to introduce the probability @xmath205 that a given point of the chain belongs at renormalization scale @xmath133 to a blob of length @xmath91 and is at distances @xmath206 from the ends of the blob .
it reads @xmath207 and is normalized to @xmath208 .
we then need the probability @xmath209 to be at height @xmath52 at time " @xmath210 for a brownian motion of diffusion coefficient @xmath44 starting at @xmath11 at @xmath211 and finishing at @xmath11 at @xmath212 in the presence of a reflexive boundary at @xmath11 @xmath213 it is normalized to @xmath214 .
the probability @xmath204 that a given point of the polymer is at a distance @xmath52 from the interface ( normalized to @xmath215 ) can be now expressed within the renormalization picture as @xmath216 introducing again the rescaled variable @xmath217 , we thus obtain the scaling form @xmath218 where the rescaled distance @xmath219 to the interface , defined as @xmath220 is distributed with the following scaling function @xmath221 where @xmath200 is the fixed point distribution of equation ( [ pdel ] ) . deriving with respect to @xmath219 yields @xmath222 \end{aligned}\ ] ] in terms of the bessel function @xmath223 . in particular , the asymptotic behaviors at @xmath224 and @xmath225
are given respectively by @xmath226 and @xmath227 we may also compute the moments @xmath228 in particular , the mean value of the square distance to the interface is given at high temperature by @xmath229 to compare our approach with the replica gaussian variational description of reference @xcite , we now compute the probability @xmath230 that a given monomer of a given polymer chain is at distance @xmath52 from the interface in a configuration ( a ) and at a distance @xmath231 from the interface in a configuration ( b ) ( note that in the effective thermodynamics , we have @xmath232 ) .
using the same notations as in the previous section , the joint distribution of @xmath52 and @xmath231 can be obtained as @xmath233 using again the rescaled variable @xmath217 , we obtain the scaling form @xmath234 the scaling function @xmath235 reads in terms of the fixed point distribution @xmath200 of equation ( [ pdel ] ) @xmath236 it is interesting to note that it is actually a function of the single variable @xmath237 . the probability that a given monomer is at the same @xmath52-coordinate in the two configurations ( a ) and ( b ) thus behaves at high temperature as @xmath238 as compared to the decay as @xmath239 found in reference @xcite for the replica symmetric solution . within the renormalization picture ,
the correlation function @xmath53 of the solvents seen by two monomers at distance @xmath240 along the polymer chain turns out to correspond exactly to the spin correlation function @xmath241 of the random field ising model in the glauber dynamics starting from a random initial condition computed in references @xcite @xcite .
we thus only quote the result @xmath242 \nonumber \\ & & = \sum_{n=-\infty } ^{\infty } \frac{48 ( 2n+1)^2\pi^2 + 32 \sigma \frac{\vert
s - s ' \vert}{\gamma^2 } } { ( 2n+1)^4\pi^4 } e^{-(2n+1)^2\pi^2 \sigma
\frac{\vert s - s ' \vert}{\gamma^2 } } \qquad , \end{aligned}\ ] ] where @xmath243 . in particular ,
the correlation length @xmath244 associated with this correlation function is given by the decay of the dominant exponential at large separation @xmath245 @xmath246
in the biased case , it is necessary to introduce two probability distributions @xmath247 and @xmath248 representing the probabilities for a blob at scale @xmath133 in the @xmath249 ) solvent to have absolute charge @xmath127 and length @xmath91 .
it was shown in reference @xcite that with the renormalization procedure ( [ rgrules ] ) , these probability distributions flow at large @xmath133 towards @xmath250 } \mp \delta \nonumber \\
u_{\gamma}^{\pm}(p ) & = & \frac{\delta(p ) } { \sinh{[\gamma \delta(p ) ] } } e^{\mp \delta \gamma } \nonumber \\
\delta(p ) & = & \sqrt{\delta^2+\frac{p}{\sigma } } \qquad , \nonumber \end{aligned}\ ] ] where @xmath162 and @xmath27 have been defined in equations ( [ defq ] ) and ( [ defdelta ] ) in terms of the initial charge distribution @xmath10
. the solutions ( [ solu - biased ] ) are valid in the scaling regime of large @xmath133 , small @xmath27 and small @xmath251 with @xmath252 fixed and @xmath253 fixed @xcite .
in particular , the distributions of the absolute charge @xmath254 are simple exponentials : @xmath255 whereas the distributions of the length of blobs have the form of an infinite series of exponentials @xcite @xmath256 } \mp \delta \sinh{[\gamma \delta(p ) ] } } \right ) \nonumber \\ & & = \frac{\sigma}{\gamma^2 } \sum_{n=0}^{\infty } j_n^{\pm}(\gamma=\delta \gamma ) e^{- \frac{\sigma l}{\gamma^2 } s_n^{\pm}(\gamma=\delta \gamma ) } \qquad , \end{aligned}\ ] ] where the functions @xmath257 and @xmath258 are defined in terms of the roots @xmath259 ( @xmath260 ) of the equation @xmath261 for @xmath262 , the root @xmath263 does not exist , but is replaced by the positive root @xmath264 of the equation @xmath265 . in terms of these roots , we have @xmath266 except for the @xmath101 @xmath267 term in the domain @xmath268 for which @xmath269 the mean lengths of the blobs in the domains @xmath48 have simple expressions @xmath270 but the mean values of @xmath271 are unfortunately much more complicated : @xmath272 where @xmath273 is the euler constant .
using again equation ( [ mlasymp ] ) , the averaged free - energy difference corresponding to the flip of a blob out of the domain @xmath48 reads @xmath274 which yields to a quite complicated expression for arbitrary @xmath275 .
in the following , we consider the case of large @xmath276 , where the distribution @xmath277 is dominated by the @xmath267 term @xcite @xmath278 leading to @xmath279 in this regime , the probabilities to have blobs with absolute charge @xmath280 in the @xmath48 domain read respectively ( [ distriq ] ) @xmath281 and thus the renormalization consists asymptotically in flipping only blobs out of the @xmath99 domain .
we thus now concentrate on the free - energy difference corresponding to the flip of a @xmath99 blob , which can be estimated as follows @xmath282 \nonumber \\ & & \simeq 2 \gamma - \frac{3}{2 } t \left [ 2\delta \gamma + \ln \frac{1}{4 \delta^2 \sigma}-c_{\rm euler } -1\right ] \qquad , \end{aligned}\ ] ] so that finally , in the regime @xmath42 , @xmath283 , with @xmath284 fixed , and large temperatures , we have @xmath285 since the renormalization procedure has to be performed as long as @xmath286 , we obtain a transition at the temperature @xmath287 where @xmath27 is the parameter characterizing the distribution of quenched charges ( [ defdelta ] ) . for @xmath119 ,
the renormalization procedure has to be stopped at scale @xmath143 given by @xmath288 at the temperature @xmath289 , the renormalization scale @xmath290 diverges : this corresponds to the delocalization transition found previously in references @xcite @xcite @xcite @xcite @xcite @xcite . for the case where the initial distribution of charges is gaussian ( [ gauss ] )
, we get @xmath291 in agreement with the scaling obtained previously @xcite @xcite @xcite and with the bounds established in reference @xcite , which read in terms of our notations @xmath292 we note moreover that the value of the transition temperature obtained here coincides with the critical temperature obtained by stepanow et al .
@xcite and is lower than the critical temperature obtained by trovato et al .
@xcite , since their results read respectively in our notations : @xmath293 for the case where the initial distribution of charges is a binary distribution , we obtain that @xmath3 is the solution of the equation ( [ deltabinary ] ) @xmath294 in particular for the case @xmath40 ( [ binarypm ] ) , we obtain @xmath295 which is consistent with the bounds given in reference @xcite reading in our notations : @xmath296 we may also compare with the bounds obtained in reference @xcite concerning the binary case with @xmath297 , @xmath298 and @xmath299 where @xmath300 for which we have : @xmath301 our renormalization approach yields thus for this case the following behavior for the critical line @xmath302 at high temperature @xmath303 which satisfies the bounds established in reference @xcite @xmath304 following the computations done for the symmetric case in section [ thermosym ] , we obtain that the energy per monomer @xmath191 of the chain in the thermodynamic limit can be obtained as @xmath305 } { \int_{q , l } l [ p^+_{\gamma_{eq}(t)}(q , l)+p^-_{\gamma_{eq}(t)}(q , l ) ] } \qquad , \end{aligned}\ ] ] with @xmath143 defined in equation ( [ geqd ] ) . using the fixed point solution of equations ( [ solu - biased ] ) , we have at large temperature @xmath306 - \frac{2 \delta^2 \sigma \gamma_{eq}(t ) } { \sinh^2 [ \delta \gamma_{eq}(t ) ] } \nonumber \\ & & \simeq - 2 \delta \sigma -8 \sigma \delta^2 \gamma_{eq}(t )
e^{-2 \delta \gamma_{eq}(t)}+ \ldots\end{aligned}\ ] ] since in the limit of small dissymmetry @xmath42 , we have the relation ( [ univ ] ) , the energy per monomer at the transition @xmath3 is simply @xmath307 and coincides of course with the energy per monomer when the heteropolymer is delocalized in the solvent @xmath101 for @xmath120 .
surprisingly however , we find that the critical behavior near @xmath3 is governed by the following essential singularity : @xmath308 \qquad .\end{aligned}\ ] ] the entropy per monomer @xmath195 is given in the thermodynamic limit by the generalization of equation ( [ ssymm ] ) @xmath309 [ p^+_{\gamma_{eq}(t)}(q , l)+p^-_{\gamma_{eq}(t)}(q , l ) ] } { \int_{q , l } l [ p^+_{\gamma_{eq}(t)}(q , l)+p^-_{\gamma_{eq}(t)}(q , l ) ] } \nonumber \\ & & = \ln \mu- \frac{3}{2 } \frac { \overline{\ln l^+ } + \overline{\ln l^- } } { \overline { l^+ } + \overline { l^- } } - 2 \frac{\ln \kappa } { \overline { l^+ } + \overline { l^- } } \qquad , \end{aligned}\ ] ] where the averages have to be computed with the probability distributions @xmath310 given in equations ( [ solu - biased ] ) .
considering as before the regime @xmath42 , @xmath283 , with @xmath311 fixed where we can use equation ( [ asymp+ ] ) , we get @xmath312 \qquad .\end{aligned}\ ] ] we finally give the expression of the free - energy per monomer @xmath313\end{aligned}\ ] ] which again presents an essential singularity near @xmath3 : the delocalization transition is thus found to be of infinite order .
the well known _
pure _ models presenting infinite order transitions are the two - dimensional xy model @xcite and the 1d ising or potts model with inverse square long range interactions @xcite @xcite .
the important property of these systems is that the interaction between two defects ( two vortices in the xy model or two kinks in the spin chain ) is logarithmic at larges distances .
the usual estimation of the free - energy for a pair of defects , where both the energy and the entropy behave as the logarithm of the size @xmath0 of the system , leads to a factorization of @xmath314 as in the imry - ma argument leading to equation ( [ relationflp ] ) for the heteropolymer problem .
it would be interesting to discuss in more details the similarities / differences between these systems , but this goes beyond the scope of the present paper . in the localized phase @xmath119
, we find that the blobs in the domain @xmath315 and @xmath316 have respectively the typical lengths ( [ lengthpm ] ) @xmath317 as @xmath1 approaches @xmath318 , @xmath69 thus diverges with an essential singularity as @xmath319 \qquad , \end{aligned}\ ] ] whereas @xmath70 also diverges but only algebraically as @xmath320 we note that these behaviors are compatible with the relation ( [ relationlmlp ] ) between the typical lengths @xmath321 and @xmath322 obtained in reference @xcite via an imry - ma argument , since we have obtained @xmath323 our results ( [ freedyss],[lplust ] ) also satisfy the relation ( [ relationflp ] ) between the free - energy and the typical blob length @xmath324 in the preferred solvent obtained again in reference @xcite via an imry - ma argument , since we have @xmath325 as a direct consequence of equations ( [ lplust ] , [ lmoinst ] ) , we obtain that the typical distances @xmath326 to the interface in the domain @xmath315 and @xmath316 diverge respectively as @xmath327 as @xmath328 \\ & & z^-(t ) \sim \sqrt { d l^{- blob}(t ) } \opsimeq_{t \to t_c^- } = \frac { \sqrt { d \sigma}}{q_0 } \left [ \frac { \ln \frac{4 \sigma}{q_0 } } { \left(1-\frac{t}{t_c } \right ) } \right]^{1/2 } \qquad .\end{aligned}\ ] ] as in the symmetric case , the lengths @xmath329 of the blobs are independent random variables distributed respectively with the probability distributions @xmath330 .
in particular , near @xmath3 , the rescaled variable @xmath331 } \end{aligned}\ ] ] is distributed with a simple exponential law @xmath73 ( see equation [ asymp+ ] ) .
following the calculations done previously in the symmetric case , we obtain that the probabilities @xmath51 that a given point of the polymer is at a distance @xmath52 from the interface in the @xmath48 domain ( with the normalization @xmath332 ) reads @xmath333 in the regime @xmath284 studied above where @xmath334 has the asymptotic simple form of equation ( [ asymp+ ] ) , we get the following scaling form @xmath335 where the rescaled distance @xmath219 to the interface and the scaling function @xmath60 read @xmath336 } \qquad , \end{aligned}\ ] ] and @xmath337 deriving with respect to @xmath219 yields @xmath338 in particular , the asymptotic behaviors at @xmath224 and @xmath225 are given respectively by @xmath339 and @xmath340 within the renormalization picture , the correlation function @xmath53 of the solvents seen by two monomers at distance @xmath240 along the polymer chain containing a dissymmetric distribution in hydrophilic / hydrophobic components corresponds exactly to the spin correlation function @xmath241 of the random field ising model in the glauber dynamics starting from a random initial condition in the presence of an external field , which is computed in reference @xcite . here
we do not give the full result , but only the correlation length @xmath244 which , not surprisingly , is simply given in the regime @xmath284 that we consider by the typical behavior of @xmath69 obtained in equation ( [ lpmblobs ] ) @xmath341\qquad .\end{aligned}\ ] ]
up to now we have always considered the thermodynamic limit of heteropolymers of infinite length @xmath342 . in this section ,
we study the finite - size properties of the localization at the interface for the case of periodic boundary conditions , i.e. we consider cyclic chains @xmath343 of finite ( but large ) size @xmath0 , with positions @xmath344 .
we can apply the renormalization procedure as before , but the procedure will now stop at some finite renormalization scale @xmath345 where the chain delocalizes from the interface .
this renormalization scale depends upon the realization of the quenched charges and corresponds to some temperature @xmath62 via the relation @xmath346 given in equations ( [ gammaeqt ] ) and ( [ geqd ] ) respectively for the symmetric case and the biased case . in the following ,
we characterize the distribution of @xmath62 for the ensemble of finite chains of size @xmath0 , and the probability distribution of the number of blobs for the chains of length @xmath0 that are still localized at a given temperature . as shown in references
@xcite,@xcite , it is possible to follow the renormalization procedure for finite size systems .
here we deal with periodic boundary conditions , and to avoid problems with the translation invariance along the chain , it is convenient to mark a point of the chain , called the origin " in the following , and to give an orientation to the chain . for @xmath347 , we define @xmath348 as the probability that in the chain of length @xmath0 at renormalization scale @xmath133 , the origin belongs to a blob @xmath125 in the domain @xmath48 , and that there are exactly @xmath349 other blobs in the chain of absolute charges and lengths given by the sequence @xmath350 .
we also need to introduce the probability @xmath351 that the chain at scale @xmath133 is already delocalized in the @xmath48 solvent with absolute charge @xmath127 .
the normalization of these probabilities read @xmath352 note that in the last two terms , the absolute charge @xmath127 is a random variable of the domain @xmath353 , contrary to the other terms where by definition of the renormalization rule we have @xmath354 .
the renormalization equations for these probabilities read for @xmath355 @xmath356 and @xmath357 as already obtained in references @xcite @xcite for the case of fixed boundary conditions , the above renormalization equations concerning periodic boundary conditions are solved by a quasi - factorized form for @xmath355 @xmath358 where @xmath310 are the bulk distributions given in equation ( [ solu - biased ] ) , satisfying @xmath359 note that in the measure ( [ solu - measure ] ) , the first bond plays a special role as it is defined as the bond containing the origin .
the equation ( [ eqn1 ] ) now reads @xmath360 using now the explicit expressions ( [ solu - biased ] ) , we get in laplace transform with respect to @xmath0 @xmath361 } e^{-q ( \delta(p ) \coth{[\gamma \delta(p ) ] } \mp \delta ) } \right ) \end{aligned}\ ] ] where @xmath362 . the equations ( [ eqn1ex ] ) for the probabilities @xmath363 that a chain of length @xmath0 is already delocalized at scale @xmath133 with an absolute charge @xmath127 in the domain @xmath48 can be integrated with respect to @xmath133 to yield @xmath364}{\delta(p)}-\frac{\gamma}{\sinh^2[\gamma \delta(p ) ] } \right ) e^{-q ( \delta(p ) \coth[\gamma \delta(p ) ] \mp \delta ) } \qquad .\end{aligned}\ ] ] we may check that @xmath365 vanishes as it should for @xmath366 . in the limit @xmath283 , the chain is expected to be always delocalized , with a charge corresponding to the sum @xmath367 of l independent identical random variables . indeed , using ( [ univ ] ) , we find @xmath368 = \frac { 1 } { 2 \sqrt{2 \sigma l } } e^{- \frac{(q \mp l q_0)^2}{4 \sigma l } } \qquad , \end{aligned}\ ] ] as expected from the central limit theorem . for arbitrary @xmath133
, we can obtain the total probabilities @xmath369 that the chain is already delocalized at scale @xmath133 in the domain @xmath48 as @xmath370 } \mp \delta \sinh{[\gamma \delta(p ) ] } ) } \left ( \frac{\cosh[\gamma \delta(p)]}{\delta(p ) } -\frac{\gamma}{\sinh[\gamma \delta(p ) ] } \right ) \right ] \qquad .\end{aligned}\ ] ] in particular for the symmetric case @xmath371 , we have @xmath372 the scaling function @xmath373 represents the total probability for a symmetric chain of length @xmath0 to be delocalized at temperature @xmath1 corresponding to the renormalization scale @xmath243 defined in equation ( [ gammaeqt ] ) .
it reads @xmath374 the probability that the polymer chain delocalizes from the interface between @xmath133 and @xmath375 is given by @xmath376 ) $ ] . to characterize the distribution of the delocalization temperature @xmath62 for the ensemble of the symmetric finite chains of length @xmath0 , it is thus convenient to define the rescaled variable @xmath377 where the function @xmath143 has been defined in equation ( [ gammaeqt ] ) .
the final result is that @xmath378 is distributed with the law : @xmath379 e^{- ( 2m+1)^2 g^2 } \qquad .\end{aligned}\ ] ] in the biased case , the probability that a chain of length @xmath0 delocalizes from the interface between @xmath133 and @xmath375 in the domain @xmath48 is given by @xmath380 ( \delta(p ) \cosh{[\gamma \delta(p ) ] } \mp \delta \sinh[\gamma \delta(p ) ] ) } \right ) \right ] \qquad .\end{aligned}\ ] ] in the regime @xmath311 considered before ( see equation ( [ asymp+ ] ) ) , the inverse laplace transform is dominated by a single pole contribution @xmath381 where @xmath382 has been introduced in ( [ asymp+ ] ) .
we now use the correspondence between temperature and renormalization scale given by @xmath243 ( [ geqd ] ) . to characterize the distribution of the delocalization temperature @xmath62 for the ensemble of dissymmetric finite chains of large length @xmath0
, it is thus convenient to define the random variable @xmath383 the final result is that @xmath384 is distributed with the law @xmath79 in particular , the typical value for the delocalization temperature corresponds to @xmath385 and thus presents a correction in @xmath386 with respect to the critical temperature @xmath3 @xmath387 the probability that a chain is still localized around the interface at scale @xmath133 and is in a state with @xmath388 blobs read for @xmath355 @xmath389 the laplace transform of the generating function of the probabilities @xmath390 reads @xmath391 \frac{z } { 1- z p^+_{\gamma}(p ) p^-_{\gamma}(p ) } \nonumber \\ & & = \frac{z \sinh [ \gamma \delta(p ) ] ( \delta^2 \sigma \sinh
[ \gamma \delta(p ) ] + p \gamma \delta(p ) \cosh [ \gamma \delta(p ) ] ) } { ( \delta^2 \sigma ( 1-z ) + p
( \cosh^2 [ \gamma \delta(p)]-z ) ) ( \delta^2 \sigma + p \cosh^2 [ \gamma \delta(p ) ] ) } \ .\end{aligned}\ ] ] in particular , the average number of blobs reads @xmath392
\frac{1 } { ( 1- p^+_{\gamma}(p ) p^-_{\gamma}(p))^2 } \right ) \nonumber \\ = & & { \cal l}^{-1}_{p \to
l } \left ( 2 \frac{\delta^2 \sigma + p \gamma \delta(p ) \coth [ \gamma \delta(p ) ] } { p^2 \sinh^2 [ \gamma \delta(p ) ] } \right ) \nonumber \\ = & & \frac{2 \sigma l \delta^2 } { \sinh^2(\delta \gamma ) } + \sum_{n=1}^{\infty } e^{- \sigma l ( \delta^2+\frac{n^2 \pi^2}{\gamma^2 } ) } 4 \frac{\sigma l n^2 \pi^2}{\gamma^2 ( n^2 \pi^2+\delta^2 \gamma^2 ) } \left ( 1+\frac{2 \delta^2 \gamma^2}{n^2
\pi^2+\delta^2 \gamma^2}-2 \frac{\sigma l n^2 \pi^2}{\gamma^2 } \right ) \ .\end{aligned}\ ] ] the first term is proportional to @xmath0 and corresponds of course to the ratio @xmath393 that could have been anticipated from the study of section [ biased ] concerning the thermodynamic limit @xmath342 .
the other terms , which represent the finite - size corrections to this dominant contribution , decay exponentially with @xmath0 .
we may also compute the second moment @xmath394 \frac{1+p^+_{\gamma}(p ) p^-_{\gamma}(p ) } { ( 1- p^+_{\gamma}(p ) p^-_{\gamma}(p))^3 } \right ) \nonumber \\ & & = { \cal l}^{-1}_{p \to
l } \left ( 4 \frac{(2 \delta^2 \sigma + p
( \cosh^2 [ \gamma \delta(p)]+1 ) ) ( \delta^2 \sigma + p \gamma \delta(p ) \coth [ \gamma \delta(p ) ] ) } { p^3 \sinh^4 [ \gamma \delta(p ) ] } \right ) \nonumber \\ & & = \frac{4 \sigma^2 l^2 \delta^4 } { \sinh^4(\delta \gamma ) } + \frac{2 \sigma l \delta^2 } { \sinh^4(\delta \gamma ) } \left ( \cosh(2 \delta \gamma ) - 4 \delta \gamma \coth ( \delta \gamma)+3 \right ) + \sum_{n=1}^{\infty } e^{- \sigma l ( \delta^2+\frac{n^2 \pi^2}{\gamma^2 } ) } ( \ldots ) \qquad .\end{aligned}\ ] ] the dominant behavior of the variance of the number of blobs is thus given by @xmath395 in the symmetric case @xmath371 , the generating function may be more explicitly computed as @xmath396 } { \sqrt { p \sigma } ( \cosh^2 [ \gamma \delta(p ) ] -z ) } \right ) \nonumber \\ & & = \sum_{n=-\infty}^{+\infty } \left ( e^{-\frac{\sigma l}{\gamma^2 } ( \alpha+n\pi)^2 } - e^{-\frac{\sigma l}{\gamma^2 } ( \frac{\pi}{2}+n\pi)^2 } \right ) \qquad , \end{aligned}\ ] ] where @xmath397 for @xmath398 .
in particular , the average and the variance of the number of blobs read : @xmath399 where @xmath243 , and where the dots represent terms that are exponentially small in @xmath400 .
as explained in section [ dynamicsinter ] , the dynamics at temperature @xmath1 for @xmath144 starting from a zero - temperature initial condition at @xmath5 , i.e. after a quench to @xmath4 for @xmath145 , can be described by the renormalization procedure , where the renormalization scale @xmath133 now corresponds to time via @xmath401 as in references @xcite @xcite .
the dynamics takes place up to time @xmath149 where equilibrium at temperature @xmath1 is reached : @xmath402 in the symmetric case , we thus have the scaling @xmath403 whereas in the biased case @xmath404 all the quantities computed before for the equilibrium at temperature @xmath1 as functions of @xmath143 have the same expressions for the dynamics as functions of @xmath405 for large time @xmath406 .
for instance , in the symmetric case , the typical length of blobs behaves as @xmath407 whereas in the biased case , the typical lengths of blobs in the domains @xmath315 and @xmath316 behave respectively as @xmath408
in this paper , we have proposed a new approach based on a disorder - dependent renormalization procedure to study the localization of random heteropolymers at the interface between two selective solvents within the model of garel et al . @xcite .
the renormalization procedure has been defined to construct an effective thermodynamics , where one only considers the heteropolymer configurations that correspond to the optimal imry - ma domain structure . at high temperatures , where the distribution of absolute charges of blobs generated by the renormalization procedure becomes infinitely broad
, the effective thermodynamics is expected to become accurate and to give asymptotic exact results . with the renormalization approach
, we have recovered that a chain with a symmetric distribution in hydrophilic / hydrophobic components is localized at the interface at any finite temperature in the thermodynamic limit , whereas a dissymmetry in hydrophilic / hydrophobic components leads to a delocalization phase transition , in agreement with previous studies .
in addition , for both cases , we have given explicit expressions for the high temperature behaviors of various physical quantities characterizing the localized phase , including in particular the free - energy per monomer , the distribution of the blob lengths in each solvent and the polymer density in the direction perpendicular to the interface .
for the case of a small dissymmetry in hydrophobic / hydrophilic components , where the delocalization transition takes place at high temperature , the renormalization approach allows to study the critical behaviors near the transition : we have found that the free energy presents an essential singularity at the transition , making all of its derivatives continuous at @xmath3 ( infinite order transition ) , that the typical length of blobs in the preferred solvent diverges with an essential singularity , whereas the typical length of blobs in the other solvent diverges algebraically .
we have then studied the finite - size properties of the problem by considering cyclic finite ( large ) chains . in particular , we have given the probability distribution of the delocalization temperature for the ensemble of chains of fixed finite length @xmath0 , and the distribution of the numbers of blobs in the chain still localized at some temperature .
finally , we have briefly discussed the non - equilibrium dynamics at temperature @xmath1 starting from a zero - temperature initial condition . in conclusion ,
the disorder - dependent renormalization approach yields explicit predictions for the high temperature behaviors of various physical quantities , that would certainly be interesting to test precisely by numerical studies .
in particular , the predictions concerning the critical behaviors near the delocalization transition are certainly surprising , and disagree with the monte carlo studies of reference @xcite where power laws were found , and with the replica gaussian variational approach of reference @xcite where the transition was found to be of second order . it would thus be particularly worthy to test numerically more precisely the order of the transition .
we hope that the predictions of the renormalization approach concerning the finite - size properties of the problem presented in section [ finite ] will make easier the comparison with numerical studies .
finally , we would like to stress that the renormalization approach yields not only predictions for averages over the quenched disorder , but also precise predictions sample by sample .
for instance , for a given realization of the quenched charges , one may compare the typical configurations of the heteropolymer at equilibrium at temperature @xmath1 with the blobs structure obtained by the numerical implementation of the renormalization procedure up to scale @xmath143 .
for the case of cyclic chains , one may also test numerically sample by sample the renormalization prediction for the delocalization temperature .
it is a pleasure to thank t. garel , h. orland , and e. guitter for fruitful discussions , and t. garel for his useful remarks on the manuscript .
in this appendix , we recall the determination via a direct imry - ma argument @xcite of the asymptotic behavior of the probability distribution of blob lengths in the limit of small lengths , and compare it with the renormalization result . as noted by garel and orland @xcite , one may extend the usual imry - ma argument for the heteropolymer exposed in reference @xcite to obtain not only the scaling of the typical length of imry - ma domains , but also some information on the probability distribution of the length of domains .
the idea is as follows @xcite . on one hand ,
the energy associated to a blob of length @xmath91 is the sum of @xmath91 independent random variables and is thus typically given by @xmath409 where @xmath410 is a random variable of order 1 , that has at least to be positive since the system tries to use the favorable fluctuations of the disorder to lower its energy . according to @xcite , the probability distribution for large @xmath410 is simply given by the gaussian of the central limit theorem : @xmath411 the behavior of the distribution @xmath412 for small @xmath410 seems however not simple to determine a priori , because the small values of @xmath410 are certainly suppressed with respect to the gaussian distribution .
indeed , if an imry - ma domain corresponds to a small value of @xmath410 , it is very likely that the heteropolymer will prefer an organization into other imry - ma domains .
it seems however difficult to write precisely the consequences of this effect on the distribution @xmath412 . on the other hand ,
the entropy associated to a blob of length @xmath91 behaves as @xmath413 ( in @xcite , only powers of @xmath91 were considered for the symmetric case and thus @xmath93 was replaced by @xmath94 , but here for a better comparison with the renormalization approach , it is more convenient to keep this @xmath93 dependence ) .
the minimization of the free energy per monomer @xmath414 with respect to @xmath91 leads to @xmath415 the typical blob length @xmath416 corresponding to @xmath417 thus satisfies @xmath418 since within the rg approach @xmath419 , the above equation corresponds to equation ( [ gammaeqt ] ) with the choice @xmath420 i.e. a numerical factor of order 1 . now introducing the rescaled length of imry - ma domains
@xmath421 we find that it is simply related to the random variable @xmath410 through @xmath422 the behavior ( [ ularge ] ) of the distribution @xmath412 at large @xmath410 thus corresponds to the following behavior of the distribution @xmath203 at small @xmath163 : @xmath423 where @xmath424 is a constant of order 1 .
this asymptotic behavior corresponds exactly to the behavior ( [ pdelzero ] ) found via the renormalization approach .
concerning the behavior of @xmath412 for small @xmath410 , we explained above why it is difficult to determine it a priori . using the relation ( [ ulambda ] ) ,
the asymptotic behavior ( [ pdelinfty ] ) found via the rg approach @xmath425 would correspond to @xmath426 i.e. a very strong suppression of small values of @xmath410 for the energies of imry - ma domains .
t. garel , h. orland and e. pitard , `` protein folding and heteropolymers '' , in _ spin glasses and random fields _ , a.p .
young ( ed . ) , world scientific , singapore ( 1997 ) p. 387 - 443 ; t.e .
creighton ( ed . ) , _ protein folding _
freeman , new york ( 1992 ) .
bray , b. derrida and c. godreche , europhys .
* 27 * 175 ( 1994 ) ; a.d .
rutenberg and a.j .
bray phys .
e * 50 * 1900 ( 1994 ) ; a.j .
bray and b. derrida phys .
e * 51 * r1633 ( 1995 ) ; b. derrida , c. godreche and i. yekutieli europhys .
* 12 * 385 ( 1990 ) and phys .
a * 44 * 6241 ( 1991 ) .
young and h. rieger , phys .
b * 53 * , 8486 ( 1996 ) ; a.p .
young , phys .
b * 56 * , 11691 ( 1997 ) ; f. igloi and h. rieger , phys . rev .
b * 57 * , 11404 ( 1998 ) ; f. igloi , r. juhasz and h. rieger , cond - mat/9811369 .
|
to study the localization of random heteropolymers at an interface separating two selective solvents within the model of garel , huse , leibler and orland , europhys .
lett . * 8 * 9 ( 1989 ) , we propose a disorder - dependent real space renormalization approach .
this approach allows to recover that a chain with a symmetric distribution in hydrophobic / hydrophilic components is localized at any temperature in the thermodynamic limit , whereas a dissymmetric distribution in hydrophobic / hydrophilic components leads to a delocalization phase transition .
it yields in addition explicit expressions for thermodynamic quantities as well as a very detailed description of the statistical properties of the heteropolymer conformations in the high temperature limit .
in particular , scaling distributions are given for the lengths of the blobs in each solvent , for the polymer density , and for some correlation functions .
in the case of a small dissymmetry in hydrophobic / hydrophilic components , the renormalization approach yields explicit expressions for the delocalization transition temperature and for the critical behaviors of various quantities : in particular , the free energy presents an essential singularity at the transition ( the transition is thus of infinite order ) , the typical length of blobs in the preferred solvent diverges with an essential singularity , whereas the typical length of blobs in the other solvent diverges algebraically .
finite - size properties are also characterized in details in both cases .
in particular , we give the probability distribution of the delocalization temperature for the ensemble of random chains of finite ( large ) length @xmath0 , and the distribution of the numbers of blobs for the chains that are still localized at a given temperature . finally , we discuss the non - equilibrium dynamics at temperature @xmath1 starting from a zero - temperature initial condition . epsf 1.5 cm -0.5 cm
|
relativistic solitons are self trapped , finite size , electromagnetic waves of relativistic intensity @xmath0 that propagate without diffraction spreading .
many different physical effects play a role in the formation of relativistic solitons : dispersion effects due to the finite particle inertia , nonlinearities due to relativistic mass increase , as well as ponderomotive effects which force the plasma density redistribution .
the theoretical investigation of relativistic solitons in electron ion plasmas is a relatively old problem in plasma physics , which has been treated by many authors in the past , and has recently gained new attention in the literature @xcite .
the analyses have been performed mainly in the framework of the 1d relativistic fluid approximation , in which solitons are described by the solutions of a set of coupled nonlinear differential equations for the electrostatic and electromagnetic potentials with suitable boundary conditions .
relativistic solitons have been seen in multi - dimensional particle in cell ( pic ) as well as in fluid simulations of laser pulse interaction with the plasma @xcite .
these solitons consist of electron density depressions and intense electromagnetic field concentrations with a larger amplitude and a lower frequency than those of the laser pulse .
since a significant portion of the overall electromagnetic energy is trapped in the form of solitons , solitary waves can play an important role in the laser plasma interaction .
here , the theory of 1d circularly polarized solitons is presented within the relativistic hydrodynamic approximation used to describe both the electron and ion components @xcite .
the plasma is assumed to be cold with zero ion and electron temperature .
length , time , velocity , momentum , vector and scalar potential , and density are normalized over @xmath1 , @xmath2 , @xmath3 , @xmath4 , @xmath5 , and @xmath6 , respectively , being @xmath7 the electron plasma frequency , @xmath8 the rest mass with @xmath9 , and @xmath6 the unperturbed electron ( and ion ) density . in the coulomb gauge ,
the maxwell s equations for the vector and scalar potentials , @xmath10 and @xmath11 , and the hydrodynamic equations ( continuity , and momentum balance ) for the densities @xmath12 and the canonical momentum @xmath13 of electrons and ions can be written as @xmath14 where @xmath13 , and @xmath15 are related to the kinetic momentum @xmath16 by @xmath17 , and @xmath18 with the parameter @xmath19 , and @xmath20 is the fluid velocity . for the 1d case in which @xmath21 , the vector potential is assumed of the form ( circular polarized wave ) @xmath22 } \ , \exp { ( -i\omega t + i k x ) } \label{aperp}\ ] ] with @xmath23 , while all the other quantities , @xmath24 , @xmath25 , @xmath26 , and @xmath27 , are assumed to depend only on the variable @xmath28 . in the present case , the relations @xmath29 and @xmath30 are satisfied , where the symbol @xmath31 refers to the direction of the laser pulse propagation .
imposing as boundary condition at the point @xmath32 , @xmath33 , @xmath34 , @xmath35 , and @xmath36 ( plasma at rest ) , the longitudinal component of the kinetic momentum , the energy , and the density of each species can be expressed as a function of the potentials as @xmath37 , @xmath38 , @xmath39 , where @xmath40 , @xmath41^{1/2 } $ ] , @xmath42 and @xmath43 .
then , for an electron ion plasma the following closed system of equations for the potentials is obtained @xmath44 where @xmath45 , @xmath46 , and @xmath47 .
the system of equations ( [ eq : phi],[eq : a ] ) describes coupled langmuir and circularly polarized transverse electromagnetic waves .
note that the r.h.s .
of is proportional to @xmath48 and thus represents the balance between the striction nonlinearity due to perturbation of the density and the relativistic nonlinearity due to mass . describes the related phase evolution of the e. m. amplitude . for @xmath49 , the solution of
is simply @xmath50 , so that the e. m. field defined in ( [ aperp ] ) has the simple form @xmath51 $ ] , being @xmath52 . for @xmath53 ,
the phase evolution is non trivial .
the system ( [ eq : phi]-[eq : a ] ) can be put in hamiltonian form , and has a first integral @xmath54 where the symbol @xmath55 denotes derivative with respect to @xmath56 .
the value of the constant @xmath57 is determined by the boundary condition . for the case of a pure transverse electromagnetic wave ( i.e. , @xmath58 and @xmath59 ) , from the dispersion equation
is obtained @xmath60 @xmath61 being the plasma frequency modified by the relativistic effects .
corresponds to the akhiezer - polovin result @xcite with the ion motion taken into account .
so far , explicit reference to an electron ion plasma has been made .
however , putting @xmath62 , eqs .
( [ eq : phi]-[eq : h ] ) are valid also in a electron positron plasma @xcite .
the particular class of localized solutions of equations ( [ eq : phi],[eq : a ] ) is investigated , satisfying @xmath63 , @xmath64 , @xmath65 , and @xmath66 at @xmath67 . for sake of simplicity ,
the case @xmath68 is considered , so that @xmath69 , and @xmath70 ( see e.g. , @xcite ) .
note that here @xmath71 is the group velocity . within the quasineutral approximation
the main features of the localized solutions can be identified in a simple manner , assuming @xmath72 , i.e. , putting @xmath73 @xcite .
then , the electrostatic potential can be written in terms of the vector potential amplitude , @xmath74 , being @xmath75 , and the equation for the vector potential in the quasineutral approximation becomes @xcite @xmath76^{1/2}}. \label{eq : qua - neu}\ ] ] the above equation has the following first integral @xmath77 , and the solution for @xmath78 is reduced to quadrature .
then , in the limit @xmath79 , the following simple expressions hold : @xmath80 , @xmath81 , and @xmath82 . from the analysis of the topology of the phase space @xmath83 of the hamiltonian @xmath84 , different kind of solitary solutions
are found varying the propagation speed , and are summarized in .
@llllll propagation & solution & potential @xmath78 & potential @xmath11 & density + speed & kind & at @xmath85 & profile & profile + @xmath86 & bright & @xmath87 , @xmath87 & positive & accumulation + @xmath88 & shock wave & @xmath87 , @xmath89 & positive & compression + @xmath90 & dark ( gray ) & @xmath91 , @xmath91 & negative & evacuation + @xmath92 & shock wave & @xmath91 , @xmath87 & negative & rarefaction + @xmath93 & dark ( black ) & @xmath91 , @xmath94 & negative & evacuation + solitary solutions are found for @xmath95 , with dark solitons satisfying the dispersion relation ( [ disprel ] ) . in both cases
, the frequency spectrum corresponds to the evanescent region in the @xmath96 space , i.e. , to the region @xmath97 .
when the ion dynamics is taken into account , bright solitons occurs for propagation speed larger than a threshold value scaling with @xmath98 , and dark solitons occur below this threshold value . standing bright solitons
are found only when the ion dynamics is neglected .
solutions in the form of shock - waves are found at critical propagation speeds .
these shock waves ( occurring in the cold plasma limit ) are collisionless , and similar to those found at ion - acoustic speed @xcite .
the characteristic velocities introduced in depend on the potential amplitude and have the following expressions in the low amplitude limit and for @xmath79 : @xmath99 , @xmath100 , @xmath101 . in the same limit ,
bright solitons are described by @xmath102 , with @xmath103 , and @xmath104 , while dark solitons by @xmath105 , with @xmath106 , and @xmath107 @xcite . at large amplitude ,
quasineutrality is violated @xcite , and the set of coupled equations ( [ eq : phi],[eq : a ] ) has to be solved numerically .
however , the above investigation on the existence of the different kind of solutions has a general validity . in the case of large amplitude
bright solitons , a wide class of solutions can be found . here
, the analysis is restricted to the class of solutions with single humped @xmath24 profiles , and @xmath78 profiles characterized by the number of nodes @xmath108 . for fixed @xmath71
, the eigenspectrum of the bright solitons can been determined solving numerically eqs .
( [ eq : phi]-[eq : a ] ) with suitable boundary conditions @xcite .
the eigenspectrum relevant for low @xmath108 values is shown in , both for the case in which the ions are fixed , i.e. , at @xmath109 , and for the case in which the ion dynamics is taken into account . .
cases ( a ) and ( b ) refer to @xmath109 ( no ion dynamics ) and @xmath110 ( with ion dynamics ) , respectively . both the regions corresponding to the continuos spectrum of single peaked solitons and the discrete spectrum for @xmath111 are shown .
the dashed curve in frame ( b ) delimits a region of no solutions .
, title="fig:",width=201 ] .
cases ( a ) and ( b ) refer to @xmath109 ( no ion dynamics ) and @xmath110 ( with ion dynamics ) , respectively . both the regions corresponding to the continuos spectrum of single peaked solitons and the discrete spectrum for @xmath111 are shown . the dashed curve in frame ( b ) delimits a region of no solutions . , title="fig:",width=201 ] when the ion dynamics is neglected ( i.e. , @xmath109 ) , the bright soliton eigenspectrum has the following features .
single - peaked solitons ( @xmath112 ) admit a continuous spectrum for any velocity value @xcite .
in particular , standing stable solitons ( @xmath113 ) are found with frequency in the range @xmath114 , and @xmath115 , as it has been shown analytically in @xcite .
multi peaked solitons @xmath116 have a discrete spectrum for velocities larger than a small critical value , and at fixed @xmath71 , the frequency @xmath117 decreases with increasing @xmath108 .
the peak value of the potentials @xmath78 and @xmath11 is found at the minimum velocity , and increases with @xmath108 @xcite .
quite different results are found when the ion dynamics is taken into account ( i.e. , @xmath118 ) , even at a finite propagation speed .
note that at high frequencies ( and velocities ) , the ion dynamics does not play any role .
humped solutions are found for velocities larger than @xmath119 @xcite , with both a continuous spectrum @xcite and a discrete spectrum @xcite . for @xmath120 , only a discrete spectrum
is found @xcite . at low @xmath108 values ,
the frequency is non monotonous as a function of @xmath71 , and at @xmath121 the branch ends since the soliton breaks .
the structure of the solution for @xmath122 and a velocity value close to breaking is shown in , where the potential waveforms are plotted .
it is found that the ions pile at the soliton center , while the electrons at its edges , giving rise to very peaked density distributions , and breaking occurs because the ion density diverges in the center . at the same time , the ion and electron velocities tend to @xmath123 at the center of the soliton and at its edges , respectively , so that in these regions the particles move with almost the group velocity of the soliton .
the ion velocity profile shows a cusp in the center , which is the signature of the nonlinear wavebreaking .
the soliton breaking provides a novel mechanism for the ion acceleration in the high intensity laser pulse interaction with plasmas @xcite .
solitons at large @xmath108 number have been investigated numerically and analytically in different approximations @xcite . and
velocities close to breaking.,title="fig:",width=144 ] and velocities close to breaking.,title="fig:",width=144 ] and velocities close to breaking.,title="fig:",width=144 ] as already pointed out , dark solitons are found only when the ion dynamics is taken into account and at very low velocities ( @xmath124 ) .
the solutions of the full nonlinear system for @xmath53 are obtained numerically , as in the case of bright solitons @xcite . the obtained spectrum and the corresponding @xmath125 value
are shown in as a function of the velocity . at fixed @xmath71 ,
the spectrum is continuous at low @xmath125 , and it becomes discrete above a threshold value , which increases with @xmath71 .
the discrete spectrum is made up by a large number of separate branches , each of them characterized by different densities and velocities profiles .
the ion density profile show always a dip , while at large @xmath91 the electron density tends to peak in the region inside the soliton . at a critical amplitude value ( dependent on @xmath71 ) breaking of the solutions
occurs , due to the peaking of the electron density .
solitary solutions are found up to @xmath126 , and @xmath127 ( for @xmath110 ) . above these values ,
no solutions have been found for the chosen parameters .
both black and gray solitons are found for @xmath71 smaller and larger than @xmath128 , while a shock wave is found at @xmath129 .
shows the waveforms of the potentials for dark solitons ( black and gray ) and the shock wave .
( b ) as a function of the group velocity @xmath71 .
the symbols gs and bs indicate the regions corresponding to the continuous spectrum of gray and black dark solitons , respectively .
the curve corresponding to shock waves is also plotted.,title="fig:",width=201 ] ( b ) as a function of the group velocity @xmath71 .
the symbols gs and bs indicate the regions corresponding to the continuous spectrum of gray and black dark solitons , respectively .
the curve corresponding to shock waves is also plotted.,title="fig:",width=201 ] , and @xmath130 , @xmath131 , and @xmath132 , respectively.,title="fig:",width=153 ] , and @xmath130 , @xmath131 , and @xmath132 , respectively.,title="fig:",width=153 ] , and @xmath130 , @xmath131 , and @xmath132 , respectively.,title="fig:",width=153 ] the results described in the previous section have been derived in the cold plasma approximation .
however , it can be easily argued that solitons moving at `` low '' velocity can be affected by pressure terms .
the theoretical investigation of temperature effects has been performed within the quasineutral approximation for standing solitons in a fully relativistic treatment @xcite , and for the case of moving solitons in the weakly relativistic approximation @xcite .
the results are summarized in @xcite , which refer to small finite amplitude solitons and non relativistic temperatures .
@lllll propagation&soliton&potential @xmath11&density & species + speed&variety&at
the center&at the center&involved + @xmath133&bright&positive&evacuation&i@xmath134e + @xmath135&bright&negative&evacuation&i@xmath134e + @xmath136&dark&negative&evacuation&i@xmath134e + @xmath137&bright&positive&accumulation&i@xmath134e + @xmath138&bright&positive & evacuation&only e + as expected , temperature effects play a crucial role in the range @xmath139 , being @xmath140 the ion acoustic speed . note that both inertia and temperature terms drive the process of soliton formation in a warm plasma .
in particular , the formation of single - peaked solitons is observed at zero and very low propagation speed , lower than the ion acoustic speed .
the relevant electron and ion density profiles show a dip in the center of the soliton . in the low temperature limit
, density depletion in the center of standing solitons may become almost total @xcite .
relativistic solitons has been investigated with the help of fluid simulations and 2d , 3d pic simulations for laser plasma interaction @xcite .
the mechanism of soliton generation by a high intensity laser pulse propagating in a underdense plasma can be descibed as follows .
the laser pulse interacting with the plasma loses its energy , as it generates wake fields behind itself .
since the process is adiabatic the ratio between the e. m. energy density and the frequency is conserved , so that frequency downshift occurs for the wake field .
moreover , since the group velocity of the laser pulse decreases with the frequency , the low frequency part of the e. m. radiation of the pulse propagates with very low velocity .
when the frequency becomes lower than the langmuir frequency , the e. m. energy becomes trapped inside the related density cavity , thus forming a slowly moving solitary structure .
note that in 2d and 3d it is found that solitary waves can merge , thus they are not strictly speaking solitons .
it has been observed that a large part of the laser pulse energy can be transformed into solitons @xcite .
it can then be argued that solitary waves may play an important role in the laser plasma interaction .
when pic simulations are performed in inhomogeneous plasmas @xcite , it is found that low frequency solitary waves generated by superintense laser pulses are accelerated along the density gradient towards the lowest density values ( e.g. , the plasma vacuum interface ) where they radiate their energy in the form of low frequency electromagnetic bursts . this process can be a signature of soliton formation .
pic simulations show that the time of the soliton formation is much shorter than the ion response time , so that ions can be assumed at rest during this process . for approximately @xmath141 oscillation periods of the e. m. field inside the soliton ( which is of the order of @xmath142 )
, the ions can be assumed as fixed . for longer times , the ponderomotive force starts to dig a hole in the ion density , and the parameters of the solitons change .
what was a soliton on the electron timescale is no more a soliton on the ion timescale .
nevertheless , a low frequency e. m. wave packet remains well confined inside a slowly expanding plasma cavity .
this e. m. entity has been called a postsoliton @xcite . in addition , 1d fluid simulations in `` slightly '' overdense plasmas ( i.e. , for plasma densities @xmath143 times the critical density ) have shown that penetration of relativistically intense laser radiation occurs by soliton like structures moving into the plasma @xcite .
many theoretical problems are still open even for 1d solitons , as , e.g. , the temperature effects at any propagation speed and arbitrary wave amplitude , the soliton dynamics in the presence of plasma inhomogeneity , the properties of linearly polarized solitons , and last but not least the stability of such structures ( e. g. , see discussions in ref .
@xcite ) . in conclusion ,
we point out that the first macroscopic evidence of soliton formation in the interaction of an intense ( @xmath144 w/@xmath145 ) laser pulse with an underdense plasma has been reported in @xcite . long - lived , macroscopic bubble
like structures have been detected through the deflection that the associated electric charge separation causes in a proton probe beam @xcite .
these structures are interpreted as the remnants of a cloud of relativistic solitons generated in the plasma by the ultraintense laser pulse .
|
relativistic solitons are self - trapped , finite size , electromagnetic waves of relativistic intensity that propagate without diffraction spreading . they have been predicted theoretically within the relativistic fluid approximation , and have been observed in multi
dimensional particle in cell simulations of laser pulse interaction with the plasma .
solitons were observed in the laser irradiated plasmas with the proton imaging technique as well .
this paper reviews many theoretical results on relativistic solitons in electron ion plasmas .
|
the compaction of dna in biological cells and viruses presents us with rather unconventional problems in the physics of soft matter . in the case of phages or viruses
, the dna may be so close - packed that the thermodynamics is no longer extensive because it is dominated by the energy arising from regions of tight bending ( odijk 1998 ) .
thus we are dealing with a defect rather than a bulk system . confronting the mathematical physics of the dna configurations as
such is a formidable undertaking .
still , the condensed dna globule ought to have relatively minor fluctuations in the density ( ubbink & odijk 1996 ) so that simplified analyses are possible .
a general maxim in theoretical biology is not to theorize before one has assimilated pertinent data and the same would seem to apply to the field of dna compaction also .
in setting up the computations outlined here , i have relied heavily on what we have learned from a number of recent experiments .
the spoollike structure of dna in t7 and t4 bacteriophages has been established in a quantitative fashion ( cerritelli et al 1997 ; olson et al 2001 ) .
for the phi29 phage we now have a fairly good comprehension of the connector motor ( simpson et al 2000 ) and the typical forces involved in dna packaging as a function of length reeled in ( smith et al 2001 ) .
cyclization experiments ( jary & sikorav 1999 ) of dna in condensed globules allow us to estimate the local friction the dna coil experiences in a congested environment .
provisional theories for dna compacted in phages were already proposed a couple of decades ago ( reimer & bloomfield 1978 ; gabashvili et al 1991 ; gabashvili & grosberg 1992 ) .
nevertheless , the notion that the tight curvature of the dna sheath surrounding a hole determines the interaxial spacing , was stressed only recently ( odijk 1998 ) .
theoretical activity in this area is currently gaining momentum but will be discussed in the last section . the experimental literature on dna packaging in phages and viruses is vast and complicated and , owing to space limitations , can not be dealt with here ( for recent reviews of the experimental situation , see grimes et al ( 2002 ) and jardine & anderson ( 2003 ) ) . here
i wish to emphasize several theoretical issues that have not been focussed on before .
a continuum approximation of the curvature energy of dna condensed within a capsid is derived in terms of the inhomogeneous dna density in the limit of small fluctuations .
a further useful approximation for close packed dna is to let the density be constant .
it turns out that it is then possible to derive general nanothermodynamic relations for quantities like the osmotic pressure , force of insertion , etc . owing to the very high degree of packing in phages ,
we are obliged to reinvestigate the classical cell model for rodlike polyelectrolytes .
the interpretation of the cyclization experiments of linear globular dna leads to the ( surprising ? ) conclusion that simple reptation seems to be valid though i note that regions of tight curvature are often absent in globules .
the friction of dna in highly congested phages is discussed in terms of current thinking on sliding friction at nanoscales . in the last section
, i confront these new computations as well as previous theoretical work with the experiments mentioned above .
the persistence length of dna is generally of the order of magnitude of the typical size of bacteriophages .
this makes the statistical treatment of dna viewed as a confined wormlike chain with self - interaction very difficult .
fortunately , there is one limit where we can make considerable headway , viz . when the dna is close - packed at moderate to high densities owing to bending stresses .
such a condensed state may exhibit so - called frustration because the packing - hexagonal for instance - can not be perfect ( earnshaw & harrison 1977 ; pereira & williams 2000 ) , but this effect will be neglected here .
let us then consider a configuration of a dna molecule enclosed within a capsid of volume @xmath0 and defined by the radius vector @xmath1 as a function of the contour distance @xmath2 from one end .
the bending energy @xmath3 in a hookean approximation is written in terms of the local radius of curvature @xmath4 of the dna curve at @xmath2 ( yamakawa 1971 ) @xmath5 here , the bending force constant equals @xmath6 where @xmath7 is the dna persistence length , @xmath8 is boltzmann s constant and @xmath9 is the temperature .
we wish to introduce the normalized density of dna segments at position @xmath10 expressed in terms of a delta function @xmath11 @xmath12 the contour length @xmath13 of the dna consists of @xmath14 `` segments '' of length @xmath15 , the distance between elementary charges viewed along the dna helical axis .
we next insert the density into eq ( 1 ) and average over all configurations ( denoted by @xmath16 ) which is possible because @xmath1 is a unique nonintersecting curve .
@xmath17 we have also introduced a close - packing approximation @xmath18 which is certainly accurate when the dna undulations away from the average are small .
i note that eq ( 3 ) pertains to both purely mechanical and statistical mechanical types of theories . even in the former ,
we still need to average over all configurations , for the capsid has a complicated shape so there must be a host of dna conformations of virtually identical energies .
the bending energy given by eq ( 3 ) was used recently ( odijk and slok 2003 ) to set up a simple density functional theory of packing and will be commented on below .
next , we consider a more severe approximation by letting the density be constant . it is conveniently expressed in terms of the area @xmath19 of the unit cell of the packaged dna ( @xmath20 ) which we often take to be hexagonal .
the supposition of uniform density allows us to deduce expressions of some generality .
the bending energy is rewritten as @xmath21 where the infinitesimal volume @xmath22 .
the volumes @xmath23 and @xmath0 are the inner and outer volumes , respectively , enclosing the packed dna ( see fig 1 ) .
we assume there is only one hole though it is stressed that general shapes of the capsid and hole are allowed .
now it can be argued by scaling and other arguments ( odijk 1986 ; selinger & bruinsma 1991 ; kamien et al 1992 ) that the twist and splay contributions to the free energy of a tightly packed liquid crystal are small compared with the bending energy .
thus we interpret @xmath24 as an energy density of pure bending which depends on the director @xmath25 but not on @xmath19 , the size of the unit cell .
moreover , because there should be no hairpins at close packing , the dna mesophase is splayless at constant density ( de gennes 1977 ) .
there is also a free energy of the dna interacting with itself .
we omit minor curvature and surface contributions ( ubbink & odijk 1995 ) and express this energy as an extensive form proportional to the contour length @xmath13 @xmath26 this pertains to a hypothetically straightened dna lattice .
accordingly , we now want to minimize the total free energy @xmath27 subject to a volume constraint @xmath28 we minimize @xmath29 with respect to the director @xmath30 ( implicit in @xmath24 ) and to the cell size @xmath19 , keeping the outer volume @xmath0 fixed but also the chemical potentials of the small ions of the buffer in the reservoir containing the phage .
@xmath32 these expressions yield @xmath33 in terms of an optimized inverse square radius of curvature @xmath34 and an integral over the surface of the inner hole ( see fig 2 ) .
a variation in @xmath19 induces a change in the vector @xmath35 at point @xmath36 , pointing normal to the inner surface in the direction of the hole .
the variation of the inner volume can be similarly written with the help of eq ( 7 ) @xmath37 hence , it is expedient to define a measure of the hole size given by @xmath38 noting that @xmath39 is simply the osmotic pressure @xmath40 , we finally attain the nanothermodynamic relation @xmath41 the usual thermodynamics does not hold : if the hole is small , we have @xmath42 and @xmath43 and so the osmotic pressure is not a purely intensive quantity .
next , the average pressure @xmath44 on the capsid wall exerted by the dna is computed by considering a virtual change in @xmath0 . concomitantly ,
the director configuration @xmath25 , the inner volume @xmath45 and cell size @xmath19 also change in such a way that the free energy remains a minimum ( eq 8) and the constraint ( eq 7 ) is obeyed .
the resulting analysis is similar to that given above , leading to @xmath46 the typical size @xmath47 of the capsid is defined via its surface similarly to eq ( 11 ) .
i note that this pressure is not at all identical to the osmotic pressure as one might have naively surmised .
it is also interesting to study a quasistatic force @xmath48 ( in the absence of dissipative losses ; kindt et al 2001 ) upon increasing the length of dna in the capsid .
we require local equilibrium ( eqs ( 14 ) ) as @xmath30 , @xmath19 and @xmath49 adjust to the changing degree of packing subject to eq ( 13 ) .
it is straightforward to show that the `` pressure '' @xmath50 of insertion is given by @xmath51 again , it would have been difficult to anticipate this expression .
we have derived eqs ( 12 - 14 ) for a phage of general shape at the expense of introducing the capsid and hole sizes as unconventional averages .
we have argued that we merely need the free energy per unit length @xmath52 of a straightened dna lattice in order to evaluate pertinent quantities of the dna enclosed in a capsid . in this section
we assume such a lattice is perfectly hexagonal without defects .
actually , the dna chains also undulate about a reference configuration but this effect is temporarily disregarded . the lattice model for rodlike polyelectrolytes
is , of course , well known ( oosawa 1971 ) , but the usual cell model which disposes of the hexagonal symmetry , breaks down at high degrees of packing when the polyion cylinders are close .
a new approximation is introduced here , though some general features of the donnan equilibrium need to be discussed first . for now , the lattice is free from simple electrolyte and consists of parallel cylindrical polyions together with attendant counterions dissolved in the water in the intervening void . in some model or approximation one
introduces , there is generally an interface or line of symmetry where the negative electrostatic potential is a maximum and is set equal to zero .
the electric field vanishes so the osmotic pressure @xmath40 and chemical potential of the ions @xmath53 reduce to ( israelachvili 1985 ) @xmath54 where @xmath55 is the concentration of counterions at the interface .
we restrict ourselves to monovalent ions bearing an elementary charge @xmath56 which are ideal save for their coulombic interaction mediated by a supposedly uniform permittivity @xmath57 .
next , the lattice is brought into thermodynamic contact with a large reservoir containing monovalent salt of concentration @xmath58 .
the activity coefficients of the small ions are set equal to unity so the resulting donnan equilibrium , due to the equality of the chemical potentials of the small ions in the respective phases , yields @xmath59 the polyelectrolyte suspension contains positive and negative ions arising from the simple salt , whose densities must be set equal in view of electroneutrality @xmath60 .
therefore , upon eliminating @xmath61 and @xmath62 , the osmotic pressure is simply ( oosawa 1971 ) @xmath63 in the so - called cell model ( developed by fuoss , katchalsky , oosawa and others , see oosawa 1971 ) , the counterion concentration @xmath55 at the interface in the salt - free case is given by @xmath64 this is valid at low and intermediate concentrations . for dna
, we have a segment length @xmath65 nm and the bjerrum length @xmath66 nm at room temperature implying that a fraction of counterions appears to be associated with the dna rods .
this is often termed counterion condensation ( manning 1969 ) .
we now address the poisson - boltzmann equation in the lattice at high concentrations ( in the salt - free case ) .
the ion distribution @xmath67 is connected to the ( negative ) electric potential @xmath68 via a boltzmann distribution @xmath69 which is inserted in the poisson equation @xmath70 this procedure is not entirely rigorous although it is quite accurate for monovalent ions ( the difficulty is that the two densities in eqs ( 19 ) and ( 20 ) are very similar though not identical ; see fixman 1979 ) .
let us consider the case when the surfaces of the polyion cylinders are separated by only a short distance @xmath71 which is much smaller than the cylinder radius @xmath72 ( see fig 3 ) .
hansen et al ( 2001 ) then still insist on using a cell model in which the effect of the rods surrounding a test rod is mimicked by a cylindrical boundary .
this seems a rather severe approximation .
an alternative is to view the hole enclosed within the triangle in fig 3 as a separate entity which is feasible provided the electrostatic screening is sufficiently high .
solving the poisson - boltzmann equation in this geometry is still difficult , so we replace the hole by three sections involving thin layers of electrolyte , and one effective hollow cylinder whose radius @xmath73 is appropriately chosen .
i consider the latter first .
the boundary condition for the electric field at its surface is @xmath74 with @xmath75 the radial coordinate .
the negative charge density @xmath76 with @xmath77 the number of charges per unit area to be chosen below .
we have to solve eq ( 19 ) and ( 20 ) so it is convenient to introduce the dimensionless variables @xmath78 and @xmath79 where the screening length @xmath80 is analogous to the debye length @xmath81 the counterion density @xmath55 now pertains to the centreline of the cylinder . there
, the potential is set equal to zero and the electric field is also zero so that eqs ( 15 ) hold .
it is then quadrature to solve the poisson - boltzmann equation : @xmath82 the concentration @xmath55 is determined from eqs ( 21 ) and ( 22 ) .
@xmath83 hence , the osmotic pressure is conveniently expressed as @xmath84 at small @xmath85 , this reduces to a condition of electroneutrality as it should , since no counterions are condensed in this limit . to progress , we need to set the variables @xmath77 and @xmath73 .
the size of the thin boundary layers discussed below is so small that they may be neglected in the following argumentation .
the triangular region in fig 3 bears half an electron charge per length @xmath15 along the major axis .
the surface density of the effective cylinder is thus @xmath86 implying that @xmath87 is independent of the interaxial spacing @xmath88 ( @xmath89 for dna ) .
the radius @xmath73 is chosen such that the volumes of enclosed liquid are the same in the respective compartments .
@xmath90 we rewrite eq ( 25 ) by introducing the volume fraction @xmath91 of dna ( the segment density of pure dna @xmath92 ; the area of a hexagonal unit cell is @xmath93 ) .
@xmath94 the density @xmath95 is identical to the uniform dna density discussed in the previous section .
obviously , our choice of cylinder parameters is not at all rigorous so the numerical coefficient in eq ( 27 ) is tentative .
the scaling of parameters in eq ( 25 ) is convincing , however , because the correct limit given by eq ( 18 ) is attained at low densities ( @xmath96 i.e. @xmath97 ) .
furthermore , the screening length @xmath80 is indeed small .
for instance , at a separation @xmath71 very close to zero , we have @xmath98 from eq ( 26 ) and since @xmath99 we get @xmath100 from eq ( 24 ) .
hence , the respective triangular sections ( fig 3 ) in the entire lattice basically act as independent units as far as the electrostatics is concerned .
the thin boundary layers are treated in a derjaguin approximation taking advantage of the relatively large radius ( @xmath101 ) . of course
, discrete charge effects play a role now , so our analysis is not quantitatively accurate . for two flat plates of charge density @xmath102 equal to that of dna
, we have @xmath103 the scaled density @xmath104 , the analogue of @xmath85 in eq ( 24 ) , is much smaller than unity so the counterions are essentially uncondensed .
the strategy of derjaguin is to superpose normal forces exerted by opposing infinitesimal platelets on the two nearby charged objects ( israelachvili 1985 ) .
we therefore focus on two particular cylinders in fig 3 and introduce the plane defined by their centre axes .
two interacting platelets are located at a height @xmath105 from this plane and the distance @xmath106 between them is @xmath107 with @xmath108 . hence , for relatively large @xmath72 , we have @xmath109 . the normal force between the two platelets per unit area is from eq ( 28 ) @xmath110 and the total force between the two cylinders of length @xmath111 is hence given by @xmath112 this can be integrated to give @xmath113 , the free energy of interaction between the two cylinders .
if there are @xmath114 rods in the actual hexagonal lattice , the total free energy is @xmath115 for there are six cylinders surrounding a test rod though we have to avoid doublecounting . by definition
the osmotic pressure is @xmath116 upon using the density for dna , @xmath102 and eq ( 30 ) . at small separations
the density @xmath55 in the middle of the hollow cylinder ( eq ( 27 ) ) is the maximum density within the triangle in fig .
there is another @xmath117 defined midway between the surfaces of two nearby cylinders . in accordance with eq
( 28 ) , this is given by @xmath118 so the screening length @xmath119 from eq ( 22 ) is quite small , significantly smaller , in fact , than the transverse scale @xmath120 of the `` surface of interaction '' evident in the integrand of eq ( 30 ) .
this justifies the derjaguin approximation but also the splitting up of the original triangular region into a hollow cylinder and three thin boundary layers .
we now place the lattice in a reservoir containing monovalent electrolyte of concentration @xmath58 .
a donnan equilibrium is established as described by eqs ( 16 ) and ( 17 ) with @xmath55 given by eq ( 27 ) .
as the phage connector forces the dna genome into a capsid , the dna coil is bound to twist in view of its helical backbone . since the other end
probably also becomes constrained at some stage , we expect a twist - to - writhe transition to occur in dealing with the dna dynamics . at first , the dna density is statistically isotropic more or less , although the density decreases towards the central region owing to the curvature stress ( odijk & slok 2003 ) .
as the concentration is enhanced , the excluded - volume effect , at some point , causes the dna globule to become ( probably ) nematic in the outer confines of the capsid while the inner region remains isotropic ( i disregard the chirality of the dna interaction for simplicity ) . as more dna is piled up into the phage , the entire globule could become nematic in a spoollike configuration . at the next stage , a hexagonal ( or hexatic ) phase appears , again , in the outer region in coexistence with a nematic inner phase
ultimately , at high degrees of packing , the entire structure becomes hexagonal apart from faults caused by frustration .
this scenario is tentative of course .
an a priori theory of globular organization is formidable as noted earlier .
one problem is how the dna loops are forced into ordered phases though we expect rippling transitions to occur ( odijk 1996 ) .
i here present two highly simplified notions of the dna dynamics . at low degrees of congestion , the usual reptation dynamics ( de gennes 1979 )
may be thought to apply , where a test chain slithers through a tubelike environment .
gabashvili and grosberg ( 1992 ) already considered reptation during the ejection of dna from phages .
the friction the dna experiences as it is transported through the connector , was also accounted for . here
, the reptation of dna is discussed in the context of cyclization in order to gauge the accuracy of such a picture . at high degrees of congestion ,
the osmotic stress builds up because of the strong curvature stress .
simple reptation no longer seems likely then so i introduce a model based on sliding friction .
_ cyclization within globules _
the dna is compacted into a globule by external forces ( e.g. by adding inert polymer or multivalent ions , by confinement within a capsid in the initial stages ) . as argued by ubbink and odijk ( 1995 , 1996 )
, it is often possible to view the globular shape as being determined by balancing bending against surface forces while keeping the volume @xmath121 fixed . for an ideal toroid defined by the two radii @xmath122 and @xmath123 shown in fig 4 ,
the surface area is @xmath124 and the volume @xmath125 is also given in terms of the ( constant ) area @xmath19 of the unit cell .
the two ends would never come into contact if the shape of the toroid were to remain the same , for the free energy of change in volume is prohibitive .
we thus investigate fluctuations in the globular shape .
i desregard free energy contributions arising from crossovers , knots and other defects ( park et al 1998 ; arsuaga et al 2002 ; hud & downing 2001 ; nelson 2002 ) .
the curvature stress is low if @xmath126 ( ubbink & odijk 1995 ; ubbink & odijk 1996 ) . under these conditions the bending energy ( eq ( 1 ) )
is approximated by @xmath127 there is a surface tension @xmath77 scaled by @xmath128 whatever the mechanism of compaction .
the surface free energy is then @xmath129 since @xmath121 is constrained , we eliminate one of the radii so as to express the total free energy in units of @xmath128 as @xmath130 minimization of @xmath131 with respect to @xmath123 yields an optimum radius of the toroid @xmath132 therefore , the fluctuations in globular shape are governed by a gaussian distribution @xmath133 this is derived by a straightforward taylor expansion of eq ( 34 ) within the boltzmann factor @xmath134 .
we next need to assess how the winding number @xmath114 changes in response to alterations in shape .
this is simply @xmath135 so a fluctuation in @xmath114 is given by @xmath136 .
the volume constraint implies @xmath137 .
if the two ends are to meet , we require @xmath138 , leading to an inequality for the second moment ( via eq ( 36 ) ) .
we finally attain the condition @xmath139 which , remarkably , is independent of the surface tension .
accordingly , by this mechanism , cyclization can only occur if the dna is quite long .
a reptation time for cyclization is now readily written down by dimensional reasoning .
it must be proportional to the friction exerted by the fluid ( water presumably , with a viscosity @xmath140 , since the globule is not very compact ) , i.e. proportional to the contour length @xmath13 . for cyclization ,
one of the dna ends has to diffuse a distance , typically about @xmath141 at least , with respect to the other end .
@xmath142 we do not address the dynamics of the two ends as they approach each other within a region of size @xmath143 . _ sliding friction in highly compacted states _
let us first consider a classic example of coulomb friction ( persson 2000 ) .
a block of some material rests on a flat surface ( fig 5 ) .
owing to gravity , for instance , the block exerts a pressure @xmath144 on a section of the surface of area @xmath145 .
if we now exert a force @xmath146 on the block , tangential to the surface , the blocks fails to move at first .
it is only at a certain @xmath146 given by @xmath147 that the block is set in motion .
the friction constant @xmath148 is well defined , at least within certain bounds on the velocity thus induced . in a highly congested dna globule compressed by bending stresses
, we may also discern a pressure @xmath40 acting on a surface area @xmath149 where @xmath150 is the length of the perimeter of the unit cell assumed to be hexagonal .
hence , the minimum force needed to set the dna in motion should be @xmath151 at very high degrees of packing , the first term in eq ( 12 ) may be neglected .
the minimum force increases fast upon further packaging of dna in a phage ( increasing @xmath13 , decreasing @xmath88 and @xmath152 ) .
it has been assumed that the dna slides homogeneously through its own tubelike environment .
_ ordered phases in the bulk and in phages _ it is important to summarize several results for liquid - crystalline dna in the bulk that are relevant to dna packaged within phages .
the literature on dna liquid crystals in vitro is vast and has been reviewed elsewhere ( livolant & leforestier 1996 ) .
durand et al ( 1992 ) performed careful x - ray diffraction experiments on slowly evaporating suspensions of 50 nm long dna rods .
the sequence of phase transitions was argued to be : isotropic - cholesteric - 2d hexagonal - 3d hexagonal - orthorhombic .
the transitions are first order except for the one between the two hexagonal phases . in the 2d hexagonal phase ,
the rods are free to move longitudinally but this freedom is gradually frozen out as the dna concentration increases , so the 2d-3d transition is continuous .
the solution only becomes orthorhombic at a very high concentration of 670
g / l ( the interaxial spacing @xmath153 nm in the coexisting 3d hexagonal phase ) .
the helical repeat actually decreases quite dramatically from 3.5 nm to 3.0 nm within the sequence of positionally ordered phases , whereas the b conformation remains surprisingly intact with a constant rise @xmath154 nm .
the buffer contained 0.25 m electrolyte ( either ammonium acetate or sodium chloride ) so these data are pertinent to the packing of dna under physiological conditions .
the ionic - strength dependence of the cholesteric - hexagonal transition for dna of length almost 50 nm was monitored by kassapidou et al ( 1998 ) , the hexagonal spacings at coexistence ranging from 3.7 nm at 0.10 m nacl to 3.0 nm at l.65 m. the orientational fluctuations measured by neutron scattering were close to 10 degrees .
the first - order transition is visible , in fact , in a test tube with a clear meniscus discernible between the two phases and the hexagonal phase exhibits the characteristic fanlike texture under a polarizing microscope ( j.r.c .
van der maarel , private communication , 1998 ) .
on the other hand , strey et al ( 2000 ) do not seem to witness the x - ray diffraction characteristic of hexagonal order , for concentrated calf - thymus dna in 0.5 m nacl , that has been wet - spun .
one potential drawback is that the long dna is quite polydisperse which could prevent a phase transition applicable to very long monodisperse dna .
anyway , the azimuthal scattering has sixfold symmetry so that the phase is probably hexatic .
if so , the decay in the orientational order should be algebraic ( nelson 2002 ) with an exponent depending on the conditions ( like the ionic strength ) .
this ought to be gleaned from the scattering curves ( lyuksyutov et al 1992 ) .
spool models have been adopted as models of dna packaged in bacteriophages ( harrison 1973 ) .
x - ray diffraction patterns from phages have often been interpreted in terms of hexagonal packing of the dna ( north & rich 1961 ; earnshaw & harrison 1977 ; stroud et al 1981 ; ront et al 1988 ; booy et al 1991 ; cerritelli et al 1997 ; bhella et al 2000 ) .
cerritelli et al ( 1997 ) have adduced quite convincing evidence for the validity of the spool form for well - aligned bacteriophage t7 .
earnshaw and harrison ( 1977 ) pointed out that the inner region of the dna spool should be more disordered .
a theory of the nonuniform density of the dna inside phages was set up recently ( odijk & slok 2003 ) in which a cholesteric - hexagonal transition could be exhibited even under conditions of rather tight packing ( note the spacings quoted above ) .
physicists have analyzed the x - ray scattering by nematic liquid crystals of poly-@xmath155-benzyl glutamate in terms of density fluctuations ( ao et al 1991 ) .
these are coupled to director fluctuations by so - called continuity laws ( de gennes 1977 , selinger & bruinsma 1991 , kamien et al 1992 ) .
the intensity spectrum as a function of the scattering vector has a characteristic bowtie shape .
there is a typical length scale associated with these fluctuations which is determined by a trade - off of bending versus osmotic compressibility ( de gennes 1977 ) .
this turns out to be similar in size to @xmath152 the radius of the inner hole of the spool inside phages ( bulk hexagonal dna here being compared to phage dna ) .
thus , it would be of some interest to reinvestigate the scattering by the inner region of the dna spool from this point of view ( odijk & slok 2003 ) .
the osmotic pressure of dna solutions has also been readdressed recently ( strey et al 1999 , raspaud et al 2000 ) . at very high concentrations where the phase is hexagonal or hexatic ( but
i shall assume it to be hexagonal for definiteness ) , there are significant deviations from the oosawa theory ( eqs ( 17 ) and ( 18 ) ) .
the approximate theory in terms of a hollow cylinder and thin boundary layers goes some way towards explaining the high values of @xmath40 ( strey et al 1999 ) .
for instance , at a dna concentration of 1.2 monom , eq ( 27 ) predicts @xmath156 n / m@xmath157 in good agreement with experiment ( eq ( 31 ) does not apply since the separation is @xmath158 and not @xmath159 ) . at a concentration of 1.8 monom , eqs ( 27 ) and ( 31 )
yield osmotic pressures @xmath160 n / m@xmath157 and @xmath161 n / m@xmath157 respectively , compared with the experimental @xmath162 n / m@xmath157 ( strey et al 1999 ) . till now
, we have neglected configurational entropy .
the free energy of a wormlike chain trapped in a tube of diameter less than the persistence length @xmath7 was formulated long ago ( odijk 1983 ) and has been discussed many times since . for dna confined within a hexagonal lattice , we may write on the basis of scaling arguments ( odijk 1983 , selinger & bruinsma 1991 ) @xmath163 @xmath164 unfortunately , the coefficient @xmath165 is not at all known accurately for a hexagonal phase . setting the coefficient @xmath166 and using eqs ( 7 ) , ( 12 ) and ( 14 ) lead to a semiquantitative explanation of the fairly sharp rise in the loading force as an elastic worm becomes
densely packed in a cavity ( see the simulations of kindt et al ( 2001 ) and of marenduzzo and micheletti ( 2003 ) .
the latter authors also find that there is a section of chain aligned within the cylindrical hole of the spool .
arsuaga et al ( 2002 ) put forward evidence for an inner region misaligned with the well - ordered outer spool under conditions of intermediate packaging but the effects of realistic electrostatics on these computations need to be awaited . ) for comparison , the osmotic pressure of the entropic interaction is @xmath167 n / m@xmath157 at 1.8 monom , an order of magnitude smaller than the pressures quoted above ( @xmath168 nm , @xmath169 nm ) .
nevertheless , the nonlinear coupling between electrostatics and undulatory entropy probably needs looking into ( see the work of de vries ( 1997 ) on undulating charged surfaces in the salt - free poisson - boltzmann approximation ) . _ loading forces and inner radii _ at high degrees of packing , the terms containing @xmath170 in eqs ( 12 ) , ( 13 ) and ( 14 ) are negligible so the relevant thermodynamic parameters reduce to @xmath171 and @xmath172 .
the hole inside a spoollike region may have caps belonging to the wall of the capsid , a fact disregarded in section 2 , but this effect gives only corrections of @xmath173 .
we thus regain eq ( 10 ) of odijk ( 1998 ) valid for a specific model , which itself is in good agreement with the dna spacings measured inside bacteriophage t7 by cerritelli at al ( 1997 ) , odijk ( 1998 ) employed the simple oosawa model ( eqs ( 17 ) and ( 18 ) ) . in recent papers , tzlil
et al ( 2002 ) and purohit et al ( 2003 ) have used a different form : @xmath174 $ ] with @xmath175 @xmath176nm@xmath177 , decay length @xmath178 nm and @xmath179 nm stemming from the osmotic stress measurements of rau & parsegian ( 1992 ) on dna greatly perturbed by trivalent cobalt hexammine ions . why this interaction , attractive in part , should generally prevail inside phages
is not clear .
anyway , the spacings predicted by eqs ( 7 ) and ( 12 ) upon inserting this form , are virtually identical to those in table 1 of tzlil et at ( 2002 ) .
the inner radius @xmath152 is now 2.81 nm when the entire genome is packaged .
the supposition of uniform density has been released in a recent analysis based on eq .
( 3 ) ( odijk & slok 2003 ) . in a density functional approach , the gradient terms are small because they exert themselves on a scale @xmath80 or @xmath88 only .
the curvature stress is balanced against the chemical potential of inhomogeneous bulk dna .
the hole size is given by an expression similar to eq ( 12 ) but the tightly wound region surrounding the hole has an inhomogeneous density varying on a scale @xmath152 ( odijk & slok 2003 ) .
such a profile seems to be in qualitative agreement with experiments of olson et al ( 2001 ) on the t4 phage .
the forces of encapsidation have been monitored in single - molecule experiments by smith et al ( 2001 ) .
i shall try to interpret this force semiempirically .
it seems a good ansatz to set the osmotic pressure @xmath180 within a certain interval of the dna concentration in the absence of salt ( see eqs ( 27 ) and ( 31 ) here , and fig 2 of hansen et al ( 2001 ) ) .
we then need the quantity @xmath52 at a certain ionic strength of the buffer .
there is a donnan equilibrium given by eq ( 17 ) and the free energy has been computed by odijk & slok ( 2003 ) .
we therefore obtain for the force @xmath181 .
smith et al ( 2001 ) measured a maximum loading force of 50 pn at a buffer concentration @xmath58 of about 0.1 m. the inner volume of the phi29 capsid can be approximated by an ellipsoid of major axis equal to 25.5 nm and minor axis equal to 19.5 nm ( tao et al 1998 ; p.j .
jardine , private communication , 2000 ) .
hence we have @xmath182 nm@xmath177 and so the unit cell has dimensions @xmath183 nm@xmath157 and @xmath184 nm for a genome length @xmath185 if we disregard the very small inner hole of size @xmath186 nm , computed from eq ( 12 ) upon a first iteration .
we finally calculate @xmath187 and the osmotic pressure @xmath40 is about 29 atm .
this underestimates the theoretical and experimental bulk pressures a bit but the effect of magnesium ions in the force set - up has not been accounted for . however , there are difficulties in the interpretation of the force - length curve below the limit of full packing .
if the inner hole remains fairly small , we may express the derivative as @xmath188 using eqs ( 7 ) and ( 14 ) .
accordingly , @xmath189 .
this is not unreasonable if @xmath190 ( see smith et al 2001 ) but fails utterly when @xmath191 , for @xmath192 increases upon compaction in that case .
it appears that we must have the osmotic pressure increasing as @xmath193 at least . what could be the cause of such a high effective power , granting that the measured force does not reflect dissipative losses ?
first , the crossover concentration demarcating the validity of the oosawa model vis - - vis a model with complete hexagonal detail is about 0.8 monom judging from fig 2 of hansen et al ( 2001 ) ( @xmath194 at this concentration ) .
thus , counterions are gradually decondensing within the anomalous range of contour lengths .
in addition , the donnan effect is also slowly diminishing and the effect of the thin boundary layers is also becoming appreciable .
an increase of the pressure stronger than linear is not surprising though this issue merits further quantitative study .
( parenthetically , a problem possibly related to this could be the curvature in the rate - length curves determined by p.j .
jardine , opposite to the one measured by smith et al ( 2001 ) ( p.j .
jardine , private communications , 2000 , 2001 ) .
he packaged dna in phi29 phages in vitro by adding atp to the solution .
the reaction was effectively blocked with the help of gamma - s - atp and the length of dna injected was quantified as a function of time using a restriction enzyme . )
the inner radii computed above at a high degree of packing are very small .
an important problem that now arises is whether such tiny radii are viable , both physically and biologically .
we are here concerned with the tight curvature of long sections of solenoidal dna independent of the base pair sequence and in the absence of bound proteins .
crick and klug ( 1975 ) already quoted values of 3 to 5 nm for radii of curvature that could pose potential problems with regard to smooth bending .
sussman & trifonov ( 1978 ) estimated a critical radius of kinking of about 2.5 nm by equating the elastic bending energy to the maximum enthalpy of stacking the base pairs assuming one kink per helical repeat .
a more detailed treatment based on van der waals contacts led to a minimum acceptable radius of about 5 nm . more recently , a 51 bp microcircle with a radius of curvature of @xmath195 nm was devised by exploiting current computer modelling techniques ( tung & soumpasis 1995 ) .
42 bp microcircles have actually been constructed in the laboratory ( wolters & wittig 1989 , 1990 ) but the base pair sequence is from a promotor . on the whole ,
a viable minimum radius would thus seem to be 3 nm or only very slightly smaller ( a. travers , private communication , 2003 ) .
the inner radii @xmath152 within phages that we have computed above are dangerously small .
it is possible that the energy of curvature becomes anharmonic at small radii which would increase estimates of @xmath152 somewhat .
any kinking would , of course , render invalid the models outlined here .
_ dna self - friction _
jary and sikorav ( 1999 ) investigated the monomolecular collapse of lambda dna caused by the trivalent polyamine spermidine and the attendant rate of cyclization of the dna .
fortunately , the collapse is very fast so it is possible to monitor the cyclization in the toroidal globular state quite accurately .
the measured rates were between about 1 to 10 @xmath196 .
this compares favourably with the reptation time from eq ( 37 ) ( the viscosity of water @xmath197 cp at room temperature , the contour length @xmath198 @xmath199 m and @xmath200 nm ( widom & baldwin 1983 ) ) .
we conclude that the dna apparently experiences friction from the `` water '' alone , despite the fact that the typical spacing in such a globule is small ( @xmath201 nm ) . in a tightly packed phage , the dynamics of the dna could be markedly different from that in globules , because of the much higher bending stress . in evaluating the frictional forces between two surfaces separated by a very thin layer of intervening fluid ,
one introduces a dimensionless group , @xmath202 in this case , where @xmath203 is the velocity tangential to the surface ( see fig 5 ; persson 2000 ) . in dna encapsidation within a phage , this velocity is so small ( smith et al 2001 ) , that this quantity is minute so we ought te have boundary lubrication rather than hydrodynamic lubrication .
the friction coefficient @xmath148 would be about @xmath204 ( persson 2000 ) . in principle
, there should then be a tendency for the water to be squeezed out between two adjacent dna strands . because there are only a few water molecules between the strands , layering transitions would be possible ( zilberman et al 2001 ) depending on the degree of commensurability of the two dna backbones ( mser et al 2001 ) .
however , eq ( 39 ) would now predict a minimum force more than 100 times larger than the 50 pn force observed in the phi29 phage .
a possible resolution of this paradox is the very recent assessment of sliding friction between surfaces with intervening aqueous nanolayers ( raviv et al 2001 , raviv & klein 2002 ) .
surprisingly , the water still acts almost like a hydrodynamic bulk fluid , possibly because the hydrogen bonds are suppressed under conditions of sliding nanofriction
. the friction coefficient could now be as low as @xmath205 so the 50 pn force ( smith et al 2001 ) may be just enough to overcome the minimum force .
it is also remarked that the free energy of compacting dna into the phi29 phage is about 1 @xmath128 per bp ( smith et al 2001 ) to be compared with the enthalpy of expulsion , 0.80 @xmath128 per bp , measured by calorimetry for bacteriophage t7 ( raman et al 2001 ; serwer 2003 ) , though the packing conditions in these two phages are not identical .
obviously , the problem of dissipative losses , if any , merits further attention .
clearly , our understanding of dna packing in bacteriophages is incomplete . although we may rationalize the high osmotic pressures and loading forces as being caused by the small separation between dna helices , the tight inner radii of the spools are too close to kinking for comfort . at intermediate degrees of packing , we lack a precise theory of the electrostatic interaction in terms of the fraction of counterions condensed on the dna .
the loading forces are very sensitive to detail in the interaction in this regime .
an open problem is the friction at a high degree of compaction and its nature - is it hydrodynamic lubrication ?
lastly , magnesium ions are often present in the buffers used .
their impact on the donnan equlibrium needs to be investigated .
i am grateful to jean - louis sikorav , edouard yeramian and paul jardine for discussions , and the latter for extended correspondence on phages and packaging .
professor andrew travers is thanked for very helpful thoughts on the minimum radius of curvature of dna .
the insights of flodder slok proved most welcome .
: : ao , x. , wen , x. & meyer , r.b .
1991 x - ray scattering from polymer nematic liquid crystals . _ physica a _ * 176 * , 63 - 71 .
: : arsuaga , j. , vzquez , m. , trigueros , s. , de witt sumners & roca , j. 2002 knotting probability of dna molecules confined in restricted volumes : dna knotting in phage capsids .
_ * 99 * , 5373 - 5377 . : : arsuaga , j. , tan r.k.z . ,
vzquez , m. , de witt sumners & harvey , s.c . 2002 investigation of viral dna packaging using molecular mechanics models .
_ biophys .
* 101 - 102 * , 475 - 484 .
: : bhella , d. , rixon , f.j . & dargan , d.j .
2000 cryomicroscopy of human cytomegalovirus virions reveals more densely packed genomic dna than in herpes simplex virus type 1 . _
biol . _ * 295 * , 155 - 161 .
: : booy , f.p . ,
newcomb , w.w .
, trus , b.l . , brown , j.c . , baker , t.s . &
steven , a.c .
1991 liquid crystalline , phage - like packing of encapsidated dna in herpes simplex virus . _ cell _ * 64 * , 1007 - 1015 . : : cerritelli , m.e . ,
cheng , n. , rosenberg , a.h . , mcpherson , c.e .
, booy , f.p . &
steven , a.c .
1997 encapsidated conformation of bacteriophage t7 dna .
_ cell _ * 91 * , 271 - 280 .
: : crick , f.h.c .
& klug , a. 1975 kinky helix .
_ nature _ * 255 * , 530 - 533 .
: : de gennes , p.g . 1977 polymeric liquid crystals : frank elasticity and light scattering . _
cryst . _ * 34 * , 177 - 182 .
: : de gennes , p.g .
1979 _ scaling concepts in polymer physics .
_ cornell university press , ithaca , n.y . : : de vries , r. 1997 thermal undulations in salt - free charged lamellar phases : theory versus experiment .
_ phys . rev
. e _ * 56 * , 1879 - 1886 .
: : durand , d. , doucet , j. & livolant , f. 1992 a study of the structure of highly concentrated phases of dna by x - ray diffraction .
_ j. phys .
ii france 2 _ * 9 * , 1769 - 1783 . : : earnshaw , w.c . &
harrison , s.c .
1977 dna arrangement in isometric phage heads .
_ nature _ * 268 * , 598 - 602 .
: : fixman , m. 1979 the poisson - boltzmann equation and its application to polyelectrolytes .
_ j. chem .
phys . _ * 70 * , 4995 - 5005 .
: : gabashvili , i.s . , grosberg , a.yu . , kuznetsov , d.v . & mrevlishvili , g.m .
1991 theoretical model of dna packing in the phage head .
_ biophysics _ * 36 * , 782 - 789 : : gabashvili , i.s . &
grosberg , a.yu .
1992 dynamics of double stranded dna reptation from bacteriophage . _ j. biomol . struct .
dyn . _ * 9 * , 911 - 920 . : : grimes , s. , anderson , d.l . , baker , t.s . & rossmann , m.g . 2000 structure of the bacteriophage phi29 dna packaging motor . _ nature _ * 408 * , 745 - 750 . : : grimes , s. , jardine , p.j . & anderson , d. 2002 bacteriophage phi29 dna packaging .
_ * 58 * , 255 - 294 .
: : hansen , p.l . ,
podgornik , r. & parsegian , v.a .
2001 osmotic properties of dna : critical evaluation of counterion condensation theory .
_ phys .
e _ * 64 * , 021907 ( 1 - 4 ) .
: : harrison , s.c .
1983 packaging of dna into bacteriophage heads : a model . _ j. mol .
biol . _ * 171 * , 577 - 580 : : hud , n.v .
& downing , k.h .
2001 cryoelectron microscopy of lambda phage dna condensates in vitreous ice : the fine structure of dna toroids .
usa _ * 98 * , 14925 - 14931 .
: : israelachvili , j.n . 1985 _ intermolecular and surface forces _ ,
academic , london .
: : jardine , p.j . & anderson , d. 2003 dna packaging . in _ the bacteriophages _ , ed
. r. calendar , oxford university press , in press .
: : jary , d. & sikorav , j .- l
. 1999 cyclization of globular dna .
implications for dna - dna interactions in vivo .
_ biochemistry _ * 38 * , 3223 - 3227 .
: : kamien , r.d . ,
le doussal , p. & nelson , d.r .
1992 theory of directed polymers .
a _ * 45 * , 8727 - 8750 . : : kassapidou , k. & van der maarel , j.r.c .
1998 melting of columnar hexagonal dna liquid crystals .
j. b. _ * 3 * , 471 - 476 .
: : kindt , j. , tzlil , a. , ben - shaul , a. & gelbart , w.m . 2001 dna packaging and ejection forces in bacteriophage .
* 98 * , 13671 - 13674 .
: : livolant , f. & leforestier , a. 1996 condensed phases of dna : structure and phase transitions .
_ * 21 * , 1115 - 1164 . : : lyuksyutov i. , naumovets , a.g . &
pokrovsky , v. 1992 _ two - dimensional crystals academic _
, san diego , ca .
: : manning , g.s .
1969 limiting laws and counterion condensation in polyelectrolyte solutions .
i. colligative properties . _ j. chem .
phys . _ * 51 * , 924 - 933 : : marenduzzo , d. & micheletti , c. 2003 thermodynamics of dna packaging inside a viral capsid : the role of dna intrinsic thickness , preprint .
: : mser , m.h . , wenning , l. & robbins , m.o .
2001 simple microscopic theory of amontons s laws for static friction .
lett . _ * 86 * , 1295 - 1298 . : : nelson , d.r .
2002 _ defects and geometry in condensed matter physics_. cambridge university press , u.k . : : north , a.c.t . &
rich , a. 1961 x - ray diffraction studies of bacterial viruses .
_ nature _ * 191 * , 1242 - 1247 .
: : odijk , t. 1983 on the statistics and dynamics of confined or entangled stiff polymers .
_ macromolecules _ * 16 * , 1340 - 1344 .
: : odijk , t. 1986 elastic constants of nematic solutions of rodlike and semiflexible polymers .
_ liquid crystals _ * 1 * , 553 - 559 .
: : odijk , t. 1996 dna in a liquid - crystalline environment : tight bends , rings , supercoils .
_ j. chem .
phys . _ * 105 * , 1270 - 1286 .
: : odijk , t. 1998 hexagonally packed dna within bacteriophage t7 stabilized by curvature stress .
_ biophys.j .
_ * 75 * , 1223 - 1227 . : : odijk , t. & slok , f. 2003 nonuniform donnan equilibrium within bacteriophages packed with dna .
_ j. phys .
b. _ in press .
: : olson , n.h .
, gingery , m. , eiserling , f.a . & baker , t.s .
2001 the structure of isometric capsids of bacteriophage t4 .
_ virology _ * 279 * , 385 - 391 . : : oosawa , f. 1971 _ polyelectrolytes _ , marcel dekker , new york . : : park , s.y . , harries , d. & gelbart , w.m . 1998 topological defects and the optimum size of dna condensates .
_ biophys.j . _ * 75 * , 714 - 720 .
: : pereira , g.g . &
williams , d.r.m .
2000 crystalline toroidal globules of dna and other semiflexible polymers : jumps in radius caused by hexagonal packing .
_ europhys .
_ * 50 * , 559 - 564 . : : persson , b.n.j .
2000 _ sliding friction_. springer , berlin , 2nd ed .
: : purohit , p.k . ,
kondev , j. & phillips , r. 2003 mechanics of dna packaging in viruses .
* 100 * , 3173 - 3178 .
: : raman , c.s . , hayes ,
nall , b.t .
& serwer , p. 2001
energy stored in the packaged dna of bacteriophage t7 .
_ biophys .
j. _ * 65 * , a12 .
: : raspaud , e. , da conceiao , m. & livolant , f. 2000 do free dna counterions control the osmotic pressure ?
lett . _ * 84 * , 2533 - 2536 .
: : rau , d.c .
& parsegian , v.a .
1992 direct measurement of the intermolecular forces between counterion - condensed dna double helices .
_ biophys .
j. _ * 61 * , 246 - 259 .
: : raviv , u. , laurat , p. & klein , j. 2001 fluidity of water confined to subnanometre films .
_ nature _ * 413 * , 51 - 54 .
: : raviv , u. & klein , j. 2002 fluidity of bound hydration layers .
_ science _ * 297 * , 1540 - 1543 . : : reimer , s.c .
& bloomfield , v.a .
1978 packaging of dna in bacteriophage heads : some considerations on energetics .
_ biopolymers _ * 17 * , 785 - 794 .
: : ront , gy . , tth , k. , feigin , l.a . ,
svergun , d.i . & dembo , a.t .
1988 symmetry and structure of bacteriophage t7 .
_ comput . math .
_ * 16 * , 617 - 628 . : : selinger , j.v .
& bruinsma , r.f .
1991 hexagonal and nematic phases of chains .
i. correlation functions .
a _ * 43 * , 2910 - 2921 . : : serwer , p. 2003 models of bacteriophage dna packaging motors . _ j. struct .
biol . _ * 141 * , 179 - 188 .
: : simpson , a.a . ,
tao , y. , leiman , p.g . ,
badasso , m.o .
, he y. , jardine , p.j . ,
olson , n.h . , morals , m.c . , grimes , s. , anderson , d.l . , baker , t.s .
& rossmann , m.g .
2000 structure of the bacteriophage phi29 dna packaging motor . _
nature _ * 408 * , 745 - 750 .
: : smith , d.e .
tans , s.j . ,
smith , s.b .
, grimes , s. , anderson , d.l . & bustamante , c. the bacteriophage phi29 portal motor can package dna against a large internal force .
_ nature _ * 413 * , 748 - 752 .
: : strey , h.h . ,
parsegian , v.a .
& podgornik , r. 1999 equation of state for polymer liquid crystals : theory and experiment .
e _ * 59 * , 999 - 1008 . : : strey , h.h . ,
wang , j. , podgornik , r. , rupprecht , a. , yu , l. , parsegian , v.a . & sitora , e.b .
2000 refusing to twist : demonstration of a line hexatic phase in dna liquid crystals .
lett . _ * 84 * , 3105 - 3108 . : : stroud , r.m . , serwer , p. & ross , m.j .
1981 assembly of bacteriophage t7 .
dimensions of the bacteriophage and its capsids .
_ biophys .
j. _ * 36 * , 743 - 757 .
: : sussman , j.l .
& trifonov , e.n .
1978 possibility of nonkinked packing of dna in chromatin .
usa _ * 75 * , 103 - 107 .
: : tao , y. , olson , n.h . , xu , w. , anderson , d.l . ,
rossmann , m.g . & baker , t.s .
1998 assembly of a tailed bacterial virus and its genome release studied in three dimensions .
_ cell _ * 95 * , 431 - 437 . : : tung , c.s .
& soumpasis , d.m .
1995 the construction of dna helical duplexes along prescribed 3-d curves .
_ j. biomol .
* 13 * , 577 - 582 .
: : tzlil , s. , kindt , j.t .
gelbart , w.m . & ben - shaul , a. 2003 forces and pressures in dna packaging and release from viral capsids .
_ biophys .
j. _ * 84 * , 1616 - 1627 .
: : ubbink , j. & odijk , t. 1995 polymer- and salt - induced toroids of hexagonal dna .
_ biophys .
j. _ * 68 * , 54 - 61 .
: : ubbink , j. & odijk , t. 1996 deformation of toroidal dna condensates under surface stress .
_ europhys .
lett . _ * 33 * , 353 - 358 .
: : widom , j. & baldwin , r.l . 1983 monomolecular condensation of lambda - dna induced by cobalt hexammine . _ biopolymers _ * 22 * , 1595 - 1620 .
: : wolters , m. & wittig , b. 1989 .
construction of a 42 base pair double stranded dna microcircle .
acids res . _ * 17 * , 5163 - 5172 .
: : wolters , m. & wittig , b. 1990 . a strategy for the construction of double - stranded dna microcircles with circumferences less than 50 bp .
_ anti - cancer drug design _ * 5 * , 111 - 120 .
: : yamakawa , h. 1971 _ modern theory of polymer solutions . _ harper & row , new york .
: : zilberman , s. , persson , b.n.j . ,
nitzan , a. , mugele , f. & salmeron , m. 2001 boundary lubrication : dynamics of squeeze - ont .
* 63 * , 055103 ( r ) ( 1 - 4 ) .
|
several controversial issues concerning the packing of linear dna in bacteriophages and globules are discussed .
exact relations for the osmotic pressure , capsid pressure and loading force are derived in terms of the hole size inside phages under the assumption that the dna globule has a uniform density .
a new electrostatic model is introduced for computing the osmotic pressure of rodlike polyelectrolytes at very high concentrations . at intermediate packing ,
a reptation model is considered for dna diffusing within a toroidal globule . under tight packing conditions , a model of coulomb sliding friction is proposed .
a general discussion is given of our current understanding of the statics and dynamics of confined dna in the context of to the following experiments : characterization of the liquid crystalline phases , x - ray scattering by phages , osmotic stress measurements , cyclization within globules and single - molecule determination of the loading forces .
|
the differential rotation ( dr ) is believed to be a key ingredient in organizing large - scale magnetic fields in the solar interior .
most of standard solar dynamo scenarios rely heavily on , so - called @xmath3-effect as the amplification process of magnetic fields to reproduce observed solar magnetic activity with cyclic polarity reversals and butterfly - shaped spatiotemporal migrations ( see , e.g. , * ? ? ?
* ; * ? ? ?
* for reviews )
. however , we have not yet arrived at a full understanding of the physical mechanism for maintaining the solar differential rotation .
see , e.g. , @xcite for a review .
the internal rotation profile of the sun has been uncovered by the global helioseismology ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a striking feature of the solar rotation profile is its equatorial acceleration : the plasma near the equator rotates faster than other parts in the convention zone .
while the bulk of the convection zone is characterized by conical iso - rotation contours , the strong radial shear is concentrated in the near - surface layer and the tachocline underlying the convection zone ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the equatorward angular momentum transport due to the turbulent reynolds stress induced by the rotating stratified convection is mainly responsible for the solar - type dr with equatorial acceleration ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the possibility of the opposite profile , or anti - solar type dr , in which the equatorial region is rotating slower , has been suggested numerically since the early 3d hydrodynamic simulation of rotating spherical - shell convection by @xcite over a broad range of parameters and various simulation setups ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , although the solar - type dr has been frequently observed in various types and ages of the stars , the anti - solar type dr has only been reported for a few of k giant stars observed with the doppler imaging technique @xcite .
it is still unclear how common the anti - solar type dr is in the solar - like main - sequence dwarfs and what really separates the two rotation regimes .
bearing in mind these situations , the transition between the solar and anti - solar type rotation profiles has been extensively studied by the simulation of spherical shell convection in recent years .
@xcite found in fully - compressible convection simulation that the transition is characterized by the coriolis number ( inverse of the rossby number which is defined in 2 ) and is only weakly dependent on the density stratification .
@xcite explored the physics of large - scale flows in solar - like stars by using an anelastic large - eddy simulation with stratification resembling the solar interior .
they confirmed that the two regimes of the dr exist even in the solar - like strongly stratified internal structure .
the critical rossby number they obtained was consistent with that in @xcite .
the systematic parameter study of the transition between the solar and anti - solar type rotation profiles was conducted by @xcite with rotating spherical shell simulations of anelastic and boussinesq convections ( see also * ? ? ? * ) .
they found , by combining their massive simulation results with the models of the different research groups , that the transition is controlled by the
convective rossby number " , almost independently of the model setup , such as thickness of the convective envelope and density stratification .
in addition , they found that two kinds of dr profiles are two possible bistable states around the transition , suggesting the hysteresis of the stellar rotation profile .
the bistability of the rotation profile , discovered by @xcite , was confirmed independently by @xcite in hydrodynamic simulations of the compressible convection .
however , the influence of magnetic field on the rotation profile and its transition is still controversial despite the magnetic field is an inevitable outcome of electrically - conducting fluid motions ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . in @xcite ( hereafter myk13 )
, we suggested , for the first time , that the formation of the solar - type dr is associated with the development of the magnetic field : it was found that the anti - solar type dr profile at the early dynamo kinematic stage transits to the solar - type profile after the dynamo - generated magnetic field beginning to affect the convective motion ( see 4.1 in myk13 ) .
such a magnetic influence on the rotation profile was confirmed by independent groups recently .
@xcite performed anelastic convective dynamo simulation of the model with the realistic solar internal structure , solar rotation rate and solar luminosity .
they showed that the dynamo - generated magnetic field plays a crucial role in attaining the solar rotation profile in the actual solar parameter regime , i.e. , the rotation profile becomes the anti - solar type without the magnetic field .
they also reported that the bistability of the rotation profile , discovered in earlier studies , disappears when allowing the evolution of the magnetic field .
the facilitation of the solar - type dr and the disappearance of the bistability of the rotation profile due to the magnetic field were confirmed by @xcite , in which they studied the effect of the magnetic field on the rotation profiles around the transition by fully - compressible convective dynamo simulations . in this paper
, we systematically study the effect of the magnetic field on rotating spherical shell convection and resultant mean dr by comparing the mhd models with their hydrodynamics counterparts ( hd models ) in the broad range of the initial rotation rate .
the bistable nature of the rotation profile is revisited under the influence of the magnetic field .
a self - consistent , fully - compressible yin - yang mhd code which was developed in myk13 and a stellar model consisting of the convection zone and the radiative zone are used for the simulation .
the primary objective of this paper is to explore a key parameter which controls the transition between the anti - solar type and solar - type dr profiles and its dependence on the magnetic field and the initial rotation profile ( 3.13.3 ) .
we cast a spotlight on two diagnostic parameters , i.e. , the rossby number " and the convective rossby number " ( see 2 for the definitions ) .
the magnetic dynamo activity in the mhd models with the different rotation rates is also studied in 3.4 . after discussing the relation between the magnetic dynamo activity and the convection properties in 4
, we summarize our findings in 5 .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] [ table1 ] the simulation setup is almost the same as the model a of myk13 .
we numerically solve a mhd convection system in a spherical shell domain defined by @xmath4 , @xmath5 , and @xmath6 , where @xmath7 , @xmath8 and @xmath9 are radius , colatitude , and longitude , respectively .
our model consists of two - layers , which qualitatively resemble the solar interior : stably stratified layer of thickness @xmath10 in the range @xmath11 and surrounding convective envelope of thickness @xmath12 in @xmath13 .
the basic equations are the fully - compressible mhd equations in the rotating frame with a constant angular velocity @xmath14 which is parallel to the coordinate axis ( @xmath15 ) : @xmath16 \ ; , \ \ \ \ \label{eq2 } \\ \frac{\partial p } { \partial t } & = & - \bm{v}\cdot\nabla p - \gamma p \nabla\cdot \bm{v } \nonumber \\ & & + ( \gamma-1)\left [ \nabla \cdot ( \kappa \nabla t ) + \eta \bm{j}^2 + \phi \right ] \ ; , \label{eq3 } \\ \frac{\partial \bm{a } } { \partial t } & = & \bm{v}\times\bm{b } - \eta \bm{j}\ ; , \label{eq4}\end{aligned}\ ] ] with @xmath17 \ ; , e_{ij } = \frac{1}{2}\left ( \frac{\partial v_i}{\partial x_j } + \frac{\partial v_j}{\partial x_i}\right ) \ ; , \nonumber \\ & & \bm{g } = -g_0/r^2\bm{e}_r\;,\ \bm{b } = \nabla \times \bm{a } \;,\ \bm{j } = \nabla\times \bm{b } \;.\nonumber\end{aligned}\ ] ] here the mass density @xmath18 , pressure @xmath19 , mass flux @xmath20 , magnetic field s vector potential @xmath21 are the basic variables .
@xmath22 is the gravitational acceleration , assumed to be constant .
we assume an ideal gas law @xmath23 with @xmath24 , where @xmath25 is the internal energy .
the viscosity , electrical resistivity , and thermal conductivity are represented by @xmath26 , @xmath27 , and @xmath28 respectively .
the initial condition is a hydrostatic equilibrium which is described by a polytropic temperature distribution with the polytropic index @xmath29 @xmath30 where @xmath31 is the specific heat at constant volume .
we choose @xmath32 and @xmath33 for the upper convective envelope and the lower stable layer , respectively .
the thermal conductivity @xmath28 is determined by requiring a constant luminosity , @xmath34 , defined by @xmath35 , throughout the domain .
non - dimensional quantities are defined by setting @xmath36 where @xmath37 is the initial density at @xmath38 .
the units of length , time , velocity , density , and magnetic field are @xmath39 , @xmath40 , @xmath41 , @xmath37 and @xmath42 , respectively .
the definitions of the prandtl , magnetic prandtl , and rayleigh numbers are @xmath43 where @xmath44 is the thermal diffusivity measured at the middle of the convection zone ( @xmath45 ) , and @xmath46 is the thickness of the convective envelope .
the stratification is controlled by a normalized pressure scale height at the surface @xmath47 , where @xmath48 is the temperature at @xmath49 .
here we use in all the models @xmath50 , yielding a small density contrast between top and bottom boundaries about @xmath33 ( see figure 1 in myk13 ) .
this is a major difference between our model and actual sun . in this paper
, we focus on the two physical non - dimensional quantities , one is the rossby number " ( ratio of inertia to coriolis forces ) and the other is the convective rossby number " ( ratio of buoyancy to coriolis forces ) ( e.g. , * ? ? ?
* ; * ? ? ?
* ) which are given by @xmath51 respectively , where @xmath52 is the wavenumber of the largest convective eddies ( this can be justified because of our weakly stratified system ) , @xmath53^{1/2}$ ] is the mean velocity , and @xmath54 is the taylor number .
angular brackets denote time- and volume - average in the convective envelope " at the saturated state .
the definitions of @xmath55 and @xmath56 are almost identical to those in @xcite and @xcite . note that , in this study , the convective rossby number is only a function of the rotation rate , i.e. , @xmath57 , because the other parameters are fixed in our simulations ( see following paragraphs ) .
the stress - free boundary condition for the velocity is imposed on both the radial boundaries . as for the magnetic field
, we assume the perfect conductor at the inner boundary and the radial field condition at the outer boundary .
a constant energy flux which drives the convective motion is imposed at the inner boundary and the temperature is fixed to be @xmath48 at the outer boundary .
the equations ( [ eq1])([eq4 ] ) are solved by the second - order central difference scheme with spatial discretization with yin - yang grid .
see @xcite and myk13 for details .
non - dimensional parameters @xmath58 , @xmath59 , and @xmath60 ( i.e. , @xmath61 , @xmath62 , and @xmath63 ) are commonly adopted in all the models studied here .
the total grid size for all the simulation runs is @xmath64 ( in @xmath7 ) @xmath65 ( in @xmath66 ) @xmath67 ( in @xmath9 ) @xmath68 ( yin & yang ) . to explore the response of the large - scale flow to the rotation rate
, we vary the magnitude of @xmath69 in the mhd and hd models while keeping the background hydrostatic state unchanged . a random temperature perturbation and small `` seed '' magnetic field ( only for mhd models ) are introduced at the same time in the convection zone when the calculation starts .
our simulation models are summarized in table 1 .
model names are given in the left most column .
the first letter in the model name denotes the presence and absence of the magnetic field .
the labels of mhd models contain the letter m. hd models have a label beginning with h. the second letter denotes the initial rotation profile .
the labels of basic runs with the initial rigid rotation contains the letter r. models started with a solar - type or an anti - solar type dr have a letter s or a. the number indicates the given rotation rate @xmath69 with the decimal point removed .
the mean quantities evaluated at the saturated state ( @xmath55 , @xmath70 , @xmath71 , @xmath72 , and @xmath72 ) , the rossby number ( @xmath0 ) , and the convective rossby number ( @xmath1 ) are shown in table 1 , where @xmath73 is the convective turn - over time , @xmath74 is the the equipartition strength of magnetic field , @xmath75 and @xmath76 are the mean kinetic and magnetic energy densities .
angular brackets with subscript fs " denote time- and volume - average in the full - spherical shell domain at the saturated state . as a measure of the convection strength , the estimated super - criticality level , defined by @xmath77 ,
is shown in the second column , where @xmath78 is the critical rayleigh number ( see caption for the definition of @xmath78 ) . in the second
last and last columns , dr parameter @xmath79 ( see 3.2 for the definition ) and , the resultant dr profile , s ( solar - type ) or as ( anti - solar type ) , are summarized .
we perform three sets of runs . in the first set , we run models from the rigid rotation ( set r : runs mr and hr ) with the initial conditions described in 2 . in the second and third sets , we examine the bistability of the rotation profile by taking either a solar - type ( set s : runs ms and hs ) and an anti - solar type ( set a : runs ma and ha ) solution as initial conditions .
the mhd effects on the convective flows are studied in each set .
after the convective motion sets in , it reaches a saturated state before @xmath80 in the basic run with the initial rigid rotation .
we have run the simulations up to around @xmath81 , which is almost comparable to the magnetic diffusion time , and studied the properties of the convective flow at the equilibrated state .
figure 1 shows , in the mollweide projection , the radial velocity distribution when @xmath82 on spherical surface at @xmath83 .
the normalization unit is the mean velocity , @xmath55 , of each model .
panels ( a)(c ) correspond to the mhd models ( mr030 , mr040 , and mr045 ) and panels ( d)(f ) are for the hd models ( hr040 , hr045 , and hr050 ) .
the lighter and darker tones depict upflow and downflow velocities in the range @xmath84 .
the rotating stratified convection is characterized by up - down asymmetry in all the models : the slower upflow dominant cells surrounded by the faster and narrower downflow lanes ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . as the rotation rate increases , the convective cell shrinks and its structure
is changed from the cellular pattern to the elongated columnar pattern formed in the equatorial region ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the transition from the cellular to columnar convections has been obsered in earlier simulations @xcite and occurs at around @xmath85 ( @xmath86 ) in the mhd model but @xmath87 ( @xmath88 ) in the hd model .
the magnetic field affects the convective motion , suggesting its impacts on the turbulent angular momentum transport and thus the resultant large - scale dr .
shown in figure 2 is the profile of the mean dr , defined by @xmath89 , where the overbar denotes the time- and azimuthal - average .
the normalization unit is @xmath69 of each model .
the time average spans in the range of @xmath90 .
panels ( a)(c ) correspond to the mhd models and panels ( d)(f ) are the hd models .
the red ( blue ) tone denotes the higher ( lower ) angular velocity .
the white lines are iso - rotation contours . the interface between the convective envelope and the radiative layer is denoted by black dashed line . as shown in earlier studies ,
there exists two regimes in the rotation profile : an anti - solar type dr with slower equator [ ( a ) & ( d ) ] and a solar - type dr with faster equator [ ( b),(c),(e ) & ( f ) ] ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
while the anti - solar type dr is established in the model with the smaller rotation rate , the solar - type dr develops in the regime of the higher rotation rate .
the most remarkable difference between the mhd and hd models is the critical rotation rate ( and thus the critical convective rossby number ) for the transition . while the transition between the anti - solar and solar - type dr profiles occurs at around @xmath85 ( @xmath86 ) in the mhd model
, it does at around @xmath87 ( @xmath88 ) in the hd model .
the dr profile is dominated by so - called taylor - proudman balance " both in the mhd and hd models ( e.g. , * ? ? ? * ) and is still far from the solar conical isotachs deduced from the helioseismic measurement ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
nevertheless , it is intriguing that the iso - rotation contour is more conical in the mhd model than in the hd model .
especially , the radial shape of the isotachs is pronounced around the equator in the mhd model near the transition ( see mr035 ) .
when comparing the two models at the same rotation rate , we can see that the rotational shear is generally weaker in the mhd model .
a strong polar jet ( polar vortex ) appeared in the hd model is also suppressed in the mhd model ( see hr045 and hr050 ) . in addition , the mhd effect seems to suppress the ( viscous ) spreading of the rotation profile of the convective envelope into the radiative layer in all the mhd models , as was observed in 4 of myk13 ( see , e.g. , * ? ? ?
* ; * ? ? ?
the reduction of the rotational shear by the influence of the magnetic fields agrees with that has been reported in the earlier studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and is discussed further in 3.2 .
figure 3 shows the time - averaged mean meridional flows .
panels ( a)(c ) correspond to the mhd models and panels ( d)(f ) are the hd models around the transition between the solar and anti - solar type rotation profiles .
the color contour depicts the meridional flow velocity , defined by @xmath91 .
the normalization unit is @xmath55 of each model .
the streamlines are over - plotted with a length proportional to the flow speed .
the white dashed line denotes the interface between the convective and radiative layers .
the circulation flow is primarily counter - clockwise ( clockwise ) in the bulk of the convection zone in the northern ( southern ) hemisphere .
this is consistent with the observations of the solar sub - surface meridional flow ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and a common feature of all the models . however , there is a difference in the circulation pattern .
while a large single - cell is formed in the model with the anti - solar type dr ( see mr030 and hr040 ) , the model with the solar - type dr shows a multiple - cell pattern ( see mr040 and hr050 ) .
a similar transition from single to multiple cell patterns of the meridional flow has been observed before in different settings ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we note that , as well as the rotation pattern , the transition of the meridional flow pattern occurs at lower @xmath69 ( thus higher @xmath1 ) in the mhd model than in the hd model .
the transition between the anti - solar and solar - type drs is a result of the structural difference of the whole three - dimensional convective motion separated at the critical rotation rate .
here we quantitatively examine the dependences of the dr profiles on two key diagnostic parameters , the convective rossby number @xmath1 and the rossby number @xmath0 .
then we address which can better capture the transition between the anti - solar type and solar - type drs .
following @xcite , we quantify the dr by the amplitude of the mean azimuthal flow at the equatorial surface : @xmath92 the positive and negative @xmath93 denote the solar - type and anti - solar type drs .
figure 4(a ) shows the dr parameter ( @xmath79 ) as functions of the convective rossby number ( @xmath1 ) [ bottom axis ] and the corresponding rotation rate ( @xmath69 ) [ top axis ] .
the blue crosses and red diamonds denote the mhd and hd models .
the horizontal and vertical dashed lines indicate @xmath94 and @xmath95 . as stated in 3.1 , the dr profile transits from the solar to anti - solar type regimes as the convective rossby number increases .
the absolute amplitude of @xmath79 is , on average , larger in the anti - solar type dr than in the solar - type dr , and takes the maximum and minimum around the transition .
these are common features of the two models and are compatible with earlier studies @xcite .
the critical value of @xmath1 for the transition is different in the two models .
it occurs when @xmath96 for the mhd model and @xmath97 for the hd model , indicating a more gradual transition under the influence of the magnetic field . when comparing two models at the same @xmath1 , the magnitude of @xmath79 is smaller in the mhd model than in the hd model .
we confirm that the magnetic field has a role in suppressing the dr and helps to produce solar - type dr , which have been suggested in myk13 , @xcite and @xcite .
our intriguing finding is that the critical value of the rossby number for the transition is almost the same between the mhd and hd models . shown in figure 4(b ) is the dependence of the dr parameter on the rossby number ( @xmath0 ) .
the blue crosses and red diamonds are corresponding to the mhd and hd models .
the horizontal and vertical dashed lines denote @xmath94 and @xmath98 .
the dr parameter decreases with the increase of the rossby number in both models around the transition .
the transition between the solar and anti - solar type dr profiles occurs at @xmath99 both in the mhd and hd models .
not only the critical value , the sharpness of the transition is also similar between the two models .
the critical rossby number for the transition obtained here gives a good agreement with that obtained in the hydrodynamic simulations by @xcite .
these indicate that the transition between the two rotation regimes is controlled by the rossby number rather than by the convective rossby number ( i.e. , the rotation rate itself in our study ) regardless of the presence of the magnetic field .
we stress here that @xcite also studied the magnetic influence on the dr profile by comparing hd and mhd models , but did not discuss their @xmath0-dependences . reading the critical rossby number for the transition from table 1 of @xcite and @xcite
, it is found that the transition occurs at a similar value of @xmath0 in their mhd and hd models as well .
our finding is thus consistent with the earlier studies .
the influence of the magnetic field on the dr profiles was also briefly mentioned in @xcite .
however , they focused on the dependence of the dr profile on the local rossby number ( @xmath100 ) " and did not examine its @xmath0-dependence . here
the local rossby number is defined by @xmath101 , where @xmath102 is the rms value of the non - axisymmetric ( fluctuating ) component of the velocity ( including both meridional and azimuthal components ) and @xmath103 is the mean spherical harmonic degree obtained from the kinetic energy spectrum ( e.g. , * ? ? ?
* ; * ? ? ?
since they showed that the critical local " rossby number for the transition is different between mhd and hd models , we can say that the rossby number is even better than the local rossby number at least to describe the transition between the anti - solar and solar - type dr profiles in a unified manner .
in order to reveal a cause of the difference in the critical value of the convective rossby number between the two models , we present , in figure 5 , the dependence of the mean convective velocity ( @xmath55 ) on the convective rossby number for the mhd ( blue crosses ) and hd ( red diamonds ) models . here the regime around the transition is focused .
when comparing two models at the same @xmath1 , the mhd model has the smaller convective velocity than the hd model . the mhd model thus can maintain the solar - type dr even at the lower rotation rate ( thus higher convective rossby number ) in comparison with the hd model . since the lorentz force of the dynamo - generated magnetic field puts the brakes on the convective motion ( e.g. ,
* ; * ? ? ?
* ) , the mhd model becomes more rotationally - constrained than the hd model .
this could be the reason why the magnetic field helps to produce the solar - type dr . @xcite and @xcite also argued that the reduction of the convective velocity by the dynamo - generated magnetic field is responsible for the difference of the critical rotation rate ( i.e. , critical convective rossby number ) for the solar to anti - solar transition between the hd and mhd models .
it is interesting that such a drastic difference between the mhd and hd models observed in @xmath55 ( fig.5 ) is unified in the view of the @xmath0-dependence ( fig.4b ) . as a function of the convective rossby number ( @xmath1 ) .
the blue crosses and red diamonds denote the mhd and hd models with the initial rigid rotation.,width=321 ] the transition between the solar and anti - solar type rotation patterns has been known since @xcite and studied in various settings @xcite . however , the bistability of the rotation profile has not been sufficiently - studied because it has only recently been discovered @xcite .
we revisit here the dependence of the rotation profile on the history both in the mhd and hd models by taking either a solar - type ( set s : runs ms and hs ) and an anti - solar type ( set a : runs ma and ha ) solution as initial conditions . in the set s ,
we perform simulations by starting from the saturated state of mr040 ( hr050 ) with the solar - type dr for the mhd ( hd ) run and decrease the rotation rate .
in contrast , in the set a , we choose the final state of mr025 ( hr030 ) with the anti - solar type dr as the progenitor for the mhd ( hd ) model and increase the rotation rate .
the dr parameter ( @xmath79 ) for these models are shown in figure 6(a ) as a function of the convective rossby number ( @xmath1 ) .
the models ms , ma , hs and ha are denoted by cyan triangles , green circles , pink crosses and orange squares , respectively . the dr parameters of the basic runs ( mr and hr ) are also shown by blue crosses and red diamonds as references . as shown in @xcite and @xcite , the bistability of the rotation profile can be observed near the transition in the hd model . in the set hs
, the transition of the hd solution occurs when @xmath104 while it does when @xmath105 in the set ha .
on the other hand , the set ms started from the solar - type dr reproduces the similar results with the set mr with the initial rigid rotation ( see cyan triangles and blue crosses ) .
this is consistent with the recent studies of @xcite and @xcite who found in their convective dynamo simulations that the bistability is disappeared and the dr profile becomes independent from the hysteresis when allowing the growth of the mhd dynamo .
however , in the set ma which is started from the anti - solar type dr , the bistability seems to be not disappeared ( see green circles ) .
the remaining of the bistability in the set ma of our mhd models may be related to the presence of the stable layer underlying the convective envelope , which is a difference between our simulation model and the models adopted in @xcite and @xcite . carefully studying the rotation profile of the mr030 ( fig.2a ) which has a very similar rotation profile with the progenitor for the set ma
, we can find that the stable layer has a strong equatorial acceleration despite the anti - solar type dr established in the convection zone . since the differential rotation in the stable layer should change the latitudinal temperature and entropy distributions to achieve the thermal wind balance there ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , it may enforce the anti - solar type dr even in the high @xmath69 regime which should provide the solar - type drs .
our speculation on the role of the stable layer in the maintenance of the internal rotation will be quantitatively examined in a subsequent paper .
it is interesting that , by using the rossby number ( @xmath0 ) , we can again unify the transition between the solar and anti - solar type rotation profiles independently of the hysteresis . shown in figure 6(b ) is the dr parameters of all the models as a function of the rossby number .
the symbols have the same meanings as those in figure 6(a ) .
it is found that the transition occurs at @xmath2 in all the simulation sets .
not only the critical value , the sharpness of the transition is also similar between all the cases .
this confirms that the transition between the two rotation regimes is better captured by the rossby number rather than by the convective rossby number .
one important finding in our study is that the initial rotation profile , i.e. , the evolution history of the stellar rotation , has an impact on the convective velocity if we does not allow the growth of the magnetic field .
the dependence of the mean convective velocity ( @xmath55 ) on the convective rossby number ( @xmath1 ) is shown in figure 7 .
the symbols have the same meanings as those in figure 6(a ) . here
the regime around the transition is focused .
the bistable nature of the convective velocity can be seen in the hd models , while the mhd models have similar convective velocity around the transition .
thus , in the hd cases , the set s ( set a ) with the initial solar - type ( anti - solar type ) dr becomes more rotationally - dominated ( inertia dominated ) than the other sets . in contrast , in the mhd cases , the evolution history of the dr does not change the convective velocity and the rotational dominance in the system .
this would be due to the regulation of the convective velocity by the dynamo - generated magnetic field .
these indicate that the bistability appeared in the rotation profile is essentially a consequence of the hysteresis of the convective velocity , which can be weakened by the dynamo - generated magnetic field .
since the convective rossby number denotes the relative importance of the buoyancy force in comparison with the coriolis force , the magnetic effect and the evolution history are not directly reflected in its definition .
in contrast , the rossby number describes the significance of the inertia force arising as the results of the hysteresis and the non - linear interaction between the convective motion and the magnetic field , relative to the coriolis force .
therefore , the effects of the magnetic field and the hysteresis are naturally reflected in it .
this would be the reason why the transition between the solar and anti - solar type rotation profiles can be unified in the view of the @xmath0-dependence . as a function of the convective rossby number ( @xmath1 ) .
the models mr , ms , ma , hr , hs and ha are denoted by blue crosses , cyan triangles , green circles , red diamonds , pink crosses and orange squares , respectively.,width=321 ] the mhd model exhibits a rich variety of the magnetic dynamo activity depending on the rotation rate .
we show the time - latitude diagram of ( a)@xmath106 ( left column ) , ( b)@xmath107 ( middle column ) and ( c)@xmath108 ( right column ) of mr00625 , 025 , 030 , 050 , 060 , and 150 at @xmath109 ( mid - stable zone ) in figure 8 and at @xmath83 ( mid - convection zone ) in figure 9 . here
the angular brackets with subscript @xmath110 denote the azimuthal average at a given depth .
the time and amplitude are normalized by @xmath71 and @xmath70 , respectively .
the red and blue tones denote the positive and negative mean - field strength .
the horizontal dashed - line in each panel denotes the equator .
the upper three models have the anti - solar type dr and the lower three models are characterized by the solar - type dr .
mr025060 are the models near the transition . the model with the smallest rotation rate is dominated by the turbulent magnetic field and has little large - scale component both in the radiative and convective layers ( see mr00625 ) .
this would be due to the weak coriolis force , yielding weak differential rotation . at around the transition of the rotation profile
, the spatiotemporal coherence of the magnetic field becomes remarkable . in the slow rotation regime with the anti - solar type dr , a quasi - steady large - scale magnetic field with a dipole symmetry
is organized ( see mr025 and 030 ) . @xmath106 and @xmath107 are stronger in the convection zone rather than in the radiative zone , while @xmath108 is stronger in the stable zone than in the convection zone . among the three magnetic components ,
@xmath108 is dominant in these models .
the large - scale @xmath111-component is the strongest at the mid - latitude and is antisymmetric about the equator .
this is consistent with the shearing of the @xmath112-component , i.e. , @xmath3-process , by the anti - solar type dr .
the properties of the magnetic dynamo change drastically across the transition of the dr profile .
the large - scale magnetic field shows the polarity reversal in the fast rotation regime with the solar - type dr . in the models near the transition ,
i.e. , mr050 and 060 , the large - scale magnetic component with polarity reversals is built up mainly in the stable zone .
it has commonly @xmath111-dominance .
the large - scale @xmath111-component is the strongest at around the equator in these models unlike the models with the anti - solar type dr in which it is the strongest at the mid - latitude . in mr050 ,
the convection zone is dominated by the turbulent magnetic field .
however , we can observe a weak spatiotemporal coherence of the magnetic field in the convection zone in mr060 .
it is interesting that the cycle of the polarity reversal of the large - scale magnetic field around the equator is shorter in the convection zone than in the radiative zone .
the most intriguing dynamo activity can be observed in the model with a fast rotation beyond the transition ( see mr150 ) .
note that this model has 34 times higher rotation rate than that adopted in the model around the transition . in the stable zone
, the strong large - scale magnetic component with a dipole symmetry is built up .
it concentrates at the high latitudinal region ( @xmath113 ) and seems to have a long cycle period of the polarity reversal ( @xmath114 ) .
the spatial distribution of the @xmath111-field suggests that the @xmath3-effect plays a crucial role in amplifying the large - scale magnetic field in the stable region . in the convective envelope
, two types of the dynamo mode can be clearly observed .
the long - lasting magnetic component with a dipole symmetry , which is similar to that in the stable zone , is found in the high latitude .
in contrast , at around the equatorial region , the oscillatory dynamo mode with a shorter cycle period ( @xmath115 ) and a poleward migration is excited and sustained .
the difference of the dynamo mechanism between the models with the different rotation profiles is discussed in 4 .
( red circles ) and ratio @xmath116 ( blue diamonds ) as a function of the rossby number ( @xmath117 ) for the basic mhd runs . , width=340 ] .
the different line types denote the models with different @xmath69.,width=321 ] the purpose of this paper is to elucidate the effect of the magnetic field on the formation of the rotation profile .
nevertheless , a rich variety of the magnetic dynamo activity observed in the mhd models ( see 3.4 ) makes us interested in their dynamo mechanisms .
here we discuss a possible key ingredient for the oscillatory magnetic dynamo with focusing on its relation to the kinetic helicity of the convection flow . as an indicator of the magnetic dynamo activity , we study the relative amplitude of the magnetic energy to the kinetic energy which is measured by the ratio @xmath118 , where @xmath119 and @xmath72 are mean magnetic and kinetic energy densities at the saturated state ( see 3 for the definitions of @xmath119 and @xmath72 ) .
in addition , the mean kinetic helicity ( net helicity ) , @xmath120 , of each mhd model is also evaluated , where it is defined by @xmath121 with @xmath122 here the angular brackets with subscripts
n " and s " denote the time- and volume - average in the northern and southern hemispheric convection zones at the saturated state .
shown in figure 10 is @xmath123 ( blue line with diamonds [ right axis ] ) and @xmath120 ( red line with circles [ left axis ] ) as a function of the rossby number ( @xmath0 ) for the basic mhd runs ( set mr ) .
the vertical dashed line denotes @xmath98 .
the ratio @xmath116 has a bit complicated profile : in the regime of the small or large rossby number , that is @xmath124 or @xmath125 , @xmath123 decreases with the increase of @xmath0 .
in contrast , at around the transition , the higher @xmath0 rather provides the larger @xmath123 .
the mean kinetic helicity shows a single peak profile with the maximum at around @xmath2 where the transition from the solar to anti - solar type rotation profiles occurs . in the regime of @xmath126 where the solar - type dr is established , @xmath116 anti - correlates with @xmath120 .
by contrast , the ratio decreases with the kinetic helicity in the anti - solar type rotation regime of @xmath127 .
since the temporal property of the large - scale magnetic field also changes from the oscillatory to the stationary across @xmath2 as was described in 3.4 , this suggests that the dynamo mechanism would be associated in some way with the kinetic helicity in the convection zone .
the @xmath0-dependence of the net kinetic helicity , which shows a peak at moderate rotation rate , is deeply related to the formation of the convective column .
since the angular velocity is lower in the regime of the anti - solar type dr or @xmath125 , the up - down asymmetry of the convective motion is lower , leading to less preference for one sign of the kinetic helicity . on the other hand , in the regime of the solar - type dr or @xmath124 ,
the convection has a tendency to align in columnar rolls parallel to the rotation axis ( see 3.1 ) . in the convection columns ,
the velocity is mostly perpendicular to the rotation axis while the vorticity is mostly parallel to the rotation axis , resulting in a small net kinetic helicity ( e.g. , * ? ? ?
* ; * ? ? ?
it was also clearly demonstrated by @xcite with asymptotic numerical analysis of unstratified turbulence that the largest net kinetic helicity is established at moderate rotation rate and decreases as rotation becomes even more rapid ( see also , * ? ? ?
. the anti - correlation between @xmath0 and @xmath116 in the solar - type dr regime would be consistent with the earlier studies ( e.g. , * ? ? ?
* ; * ? ? ?
in addition , the correlation between @xmath0 and @xmath116 around the transition is also compatible with the numerical result of @xcite ( see their table 2 ) .
however , the role of the kinetic helicity in the magnetic dynamo mechanism and its hidden connection to the observed relationship between stellar rotation and magnetic activity ( e.g. , see * ? ? ? * ; * ? ? ?
* for reviews ) remain to be fully elucidated although there are a great deal of work exploring the linkage between them ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . to get a better grasp of the relation of the kinetic helicity and the magnetic dynamo mechanism , in figure 11
, we show the distribution of the time and azimuthally - averaged kinetic helicity , @xmath128 , for the models mr00625 , 025 , 030 , 050 , 060 , and 150 , respectively .
the red and blue tones denote the positive and negative helicity .
the horizontal black dotted - line corresponds to the equator .
the dashed curve denotes the interface between the convection and radiative zones .
the models listed in the upper three panels have the anti - solar type dr ( mr00625 , 025 and 030 ) and the lower three panels correspond to the models with the solar - type dr ( mr050 , 060 and 150 ) . to clearly demonstrate the latitudinal distribution of the kinetic helicity , the radial average of the time- and azimuthally - averaged kinetic helicity , i.e. , @xmath129
, is also shown in figure 12 , where angular brackets with subscript @xmath7 denote the radial average .
the different line - types denote the models with different @xmath69
. it would be trivial that the kinetic helicity has an antisymmetric profile with respect to the equator in the model with a sufficiently fast rotation . in the northern ( southern ) hemisphere , the negative ( positive ) helicity
becomes predominant .
this is because the downflow with the negative ( positive ) helicity is on average stronger than the upflow with the positive ( negative ) helicity in the northern ( southern ) hemisphere in the density stratified system ( e.g. , * ? ? ?
* ; * ? ? ?
the weaker equatorial antisymmetry of the kinetic helicity profile in the model with the smallest rotation ( mr00625 ) would be a result of a weaker coriolis force which provides less impact on the convective motion .
the peak amplitude of the kinetic helicity is little dependent on the rotation rate ( within factor 2 ) .
however , we can find from figure 12 that the latitude for the peak of the kinetic helicity decreases as the rotation rate increases : while the higher latitudinal region has a larger helicity in the models with the anti - solar type dr ( mr00625 , 025 , and 030 ) , the lower latitudinal region has a larger helicity in the models with the solar - type dr ( mr050 , 060 , and 150 ) .
it is the most remarkable that both the hemispheric and inter - hemispheric latitudinal gradients of the kinetic helicity become larger around the equator with the increase of the rotation rate . in the models with the solar - type dr and the oscillatory large - scale magnetic field , the kinetic helicity concentrates on around the equatorial plane and has strong hemispheric and inter - hemispheric latitudinal gradients there .
in contrast , the models with the anti - solar type dr and the stationary large - scale magnetic field do not show the significant latitudinal gradient of the kinetic helicity around the equatorial region .
since , when comparing figures 9 and 11 , the latitude providing the stronger large - scale magnetic field overlaps with that with the larger kinetic helicity , it is expected that the latitudinal distribution of the kinetic helicity is a key to distinguish the dynamo mechanisms .
the larger the latitudinal gradient of the kinetic helicity around the equator becomes , the more the oscillatory large - scale magnetic field would be developed .
recently , @xcite reported an intriguing finding in their dynamo simulation by a forced helical turbulence in a wedge - shaped spherical shell that the oscillatory large - scale magnetic field with equatorial migration can be organized , even without the @xmath3-effect , essentially due to an inter - hemispheric gradient of the kinetic helicity .
they argued that the @xmath130-dynamo mode excited by the helical turbulence with the equatorial antisymmetry is responsible for the oscillatory property .
furthermore , a connection between @xmath130-dynamo mode and solar magnetism was discussed in some recent results of convective mhd dynamos simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
when considering the latitudinal gradient of the kinetic helicity around the equator as an important ingredient of the oscillatory property of the dynamo even in our simulations , we can speculate the reason why there exists a difference of the dynamo property depending on the regime of the rotation profile .
since the model with the smaller rotation rate and thus the anti - solar type dr does not have a sufficient latitudinal gradient of the kinetic helicity around the equator to excite the oscillatory @xmath130-dynamo mode , the @xmath3-effect solely works to build up the large - scale magnetic field and thus yields a stationary solution .
in contrast , the model with the faster rotation and thus the solar - type dr has a sufficient concentration and strong latitudinal gradient of the kinetic helicity around the equator , yielding the oscillatory large - scale magnetic component sustained by the @xmath130-effect in addition to the @xmath3-effect .
the detailed analysis with mode expansion by legendre polynomials on kinetic helicity and magnetic field will be presented in a subsequent paper in which we focus on the magnetic dynamo activity and its dependences on the rotation rate and stratification level .
although a quantitative description of the dynamo mechanism is beyond the scope of this paper , we can speculate from our simulation results that accurate numerical modeling of the kinetic helicity profile in the actual solar interior will offer a way to unveil the mystery of the solar magnetism .
in this paper , the effect of the magnetic field on the mean dr profile established in the convection zone was systematically studied by comparing the mhd and hd models of the rotating spherical shell convection in the broad range of the initial rotation rate .
a fully compressible yin - yang mhd code which was developed in myk13 and a stellar model consisting of the convection zone and the radiative zone were used for the simulation . the critical parameter which controls the transition between the anti - solar type and solar - type dr profiles
was explored with focusing on the rossby number ( @xmath0 ) " and the convective rossby number ( @xmath1 ) " .
in addition , the bistability of the rotation profile , which has only recently been discovered by @xcite , was revisited under the influence of the magnetic field .
our main findings are summarized as follows .
the critical value of the @xmath1 for the transition was higher in the mhd model than the hd model .
the transition was more gradual in the model under the influence of the magnetic field .
comparing two models at the same @xmath1 , the magnitude of the dr parameter ( @xmath79 ) was smaller in the mhd model than in the hd model .
the magnetic field had a crucial role in suppressing the dr and facilitated to produce the solar - type rotation profile . since the lorentz force of the dynamo - generated magnetic field reduced the convective velocity ,
the mhd model became more rotationally - constrained than the hd model .
this would be the reason why the magnetic field facilitated to produce the solar - type dr .
the transition from the solar to anti - solar type rotation profiles occurred at @xmath99 both in the mhd and hd models .
not only the critical value , the sharpness of the transition was also similar between the two models in the view of @xmath0-dependence .
the solar - type dr , accompanied by the columnar convection and multi - cell meridional circulation pattern , was established in the regime @xmath126 .
the anti - solar type dr , characterized by the cellular convections and the single - cell meridional circulation , developed in the regime of @xmath127 regardless of the presence of the magnetic field .
the transition between the two rotation profiles was controlled by the @xmath0 rather than by the @xmath1 .
\3 . from the point of view of the @xmath1-dependence
, the rotation profile showed , as was observed in earlier studies , the bistability near the transition in the hd model while it disapeared when allowing the growth of the dynamo - generated magnetic fields except for the model with taking an anti - solar type solution as the initial condition .
it was found that the transition between the anti - solar and solar - type dr profiles could be unified in the view of @xmath0-dependence independently of the hysteresis .
the transition between the two rotation regimes was again better captured by the @xmath0 rather than by the @xmath1 .
the influence of the magnetic field or the history on the convective velocity would be responsible for the dependence of the rotation profile on it .
such effects were not reflected in the @xmath1 by definition .
in contrast , it could be definitely captured by the @xmath0 .
the transition between the solar and anti - solar type rotation profiles was thus unified in the view of the @xmath0-dependence .
. a rich variety of the magnetic dynamo activity could be observed in the mhd models . in the regime of the anti - solar type dr ( @xmath127 ) , the stationary large - scale magnetic field with a dipole symmetry
was built up both in the stable and convection layers except for the model with the smallest rotation rate .
in contrast , the oscillatory large - scale magnetic component was organized in the model with the solar - type dr profile ( @xmath126 ) .
the latitudinal profile of the kinetic helicity would be a key to distinguish the dynamo mechanisms between the two rotation regimes .
the oscillatory large - scale magnetic field could be organized when the strong latitudinal gradient of the kinetic helicity existed around the equator .
our parametric study was conducted by the simulation model with the same value of the rayleigh number . the supercriticality level @xmath131 ( @xmath132/{\rm ra}_c $ ] ) , which is an essential measure of the turbulence level , was thus not constant and varied with the rotation rate as shown in table 1 ( see the third column ) . in order to compare the models with
the supercriticality level fixed , we should change the ekman number or prandtl number , or density stratification , providing a possible difference in the resultant flow pattern .
gastine et al .
( 2013,2014 ) indicated that , while the transition between the solar and anti - solar type rotation dr profiles would take place at almost the same @xmath1 independently of the density stratification and ekman number , the lower ekman number or the stronger density stratification would produce the steeper transition .
it was also suggested that the higher prandtl number would shifts the critical @xmath1 to the higher value at least in the high ekman number regime .
not only on the flow properties , these parameters would also have an influence on the magnetic dynamo activities ( c.f .
* ; * ? ? ?
* ; * ? ? ?
to verify the universality of the critical rossby number for the transition obtained in this study , further parameter study is indispensable and is planned as our future work .
the recent development of the astroseismology opens up the way to study the large - scale internal flows in the solar - type main sequence stars with different age and thus different rotation rate ( e.g. , * ? ? ?
computer simulation in tandem with the advanced observation will help deepening the understanding of the stellar interior dynamics and stellar dynamo activities in the astroseismology era .
we thank the anonymous referees for constructive comments .
numerical computations were carried out on cray xc30 at center for computational astrophysics , national astronomical observatory of japan .
this work was supported by jsps kakenhi grant nos .
24740125 and 20260052 .
61 natexlab#1#1 al - shamali , f.m . ,
heimpel , m.h . , & aurnou , j.m .
2004 , geophysical and astrophysical fluid dynamics , 98 , 153 aurnou , j. , heimpel , m. , & wicht , j. 2007 , icarus , 190 , 110 baliunas , s. l. , nesme - ribes , e. , sokoloff , d. , & soon , w. h. 1996 , , 460 , 848 basu , s. , & antia , h.m .
2001 , , 324 , 498 beaudoin , p. , charbonneau , p. , racine , e. , & smolarkiewicz , p.k .
2013 , , 282 , 335 brandenburg , a. , sokoloff , d. , & subramanian , k. 2012 , , 169 , 123 brandenburg , a. , saar , s. h. , & turpin , c. r. 1998 , , 498 , l51 brown , b. p. , miesch , m. s. , browning , m. k. , brun , a. s. , & toomre , j. 2011 , , 731 , 69 browning , m.k .
2008 , , 676 , 1262 brummell , n.h .
, hurlburt , n.e . , & toomre , j. 1996 , , 473 , 494 brummell , n.h . ,
hurlburt , n.e . , & toomre , j. 1998 , , 493 , 955 brun , a.s . , & toomre , j. 2002 , , 570 , 865 brun , a.s . ,
miesch , m.s . , & toomre , j. 2004 , , 614 , 1073 busse , f.h .
1970 , journal of fluid mechanics , 44 , 441 cattaneo , f. , emonet , t. , & weiss , n. 2003 , , 588 , 1183 charbonneau , p. , christensen - dalsgaard , j. , henning , r. , et al . 1999 , , 527 , 445 charbonneau , p. 2005 , living reviews in solar physics , 2 , 2 charbonneau , p. 2010 , living reviews in solar physics , 7 , 3 chan , k.l .
2010 , iau symposium , 264 , 219 chaplin , w.j . ,
appourchaux , t. , elsworth , y. , et al . 2010 , , 713 , l169 christensen , u.r .
, & aubert , j. 2006 , geophysical journal international , 166 , 97 fan , y. , & fang , f. 2014 , , 789 , 35 gastine , t. , wicht , j. , & aurnou , j.m .
2013 , icarus , 225 , 156 gastine , t. , yadav , r.k . ,
morin , j. , reiners , a. , & wicht , j. 2014 , , 438 , l76 giles , p.m. , duvall , t.l . ,
scherrer , p.h . , & bogart , r.s .
1997 , , 390 , 52 gilman , p.a . 1977 , geophysical and astrophysical fluid dynamics , 8 , 93 glatzmaier , g.a . , & gilman , p.a . 1981 , , 47 , 103 glatzmaier , g.a . ,
& gilman , p. a. 1982 , , 256 , 316 gough , d.o . , & mcintyre , m.e .
1998 , , 394 , 755 gdel , m. 2007 , living reviews in solar physics , 4 , 3 guerrero , g. , smolarkiewicz , p.k . ,
kosovichev , a.g . , & mansour , n.n .
2013 , , 779 , 176 howe , r. 2009 , living reviews in solar physics , 6 , 1 jennings , r. l. , & weiss , n. o. 1991 , , 252 , 249 kageyama , a. , & sato , t. 2004 , geochemistry , geophysics , geosystems , 5 , 9005 kpyl , p. j. , korpi , m. j. , & brandenburg , a. 2009 , , 697 , 1153 kpyl , p.j .
, mantere , m.j . , & brandenburg , a. 2011 , astronomische nachrichten , 332 , 883 kpyl , p. j. , mantere , m. j. , cole , e. , warnecke , j. , & brandenburg , a. 2013 , , 778 , 41 kpyl , p.j . ,
kpyl , m. j. , & brandenburg , a. 2014 , , 570 , aa43 karak , b.b . ,
kapyla , p.j .
, kapyla , m.j . , & brandenburg , a. 2014 , arxiv:1407.0984 kitchatinov , l.l . , & ruediger , g. 1995 , , 299 , 446 knobloch , e. , rosner , r. , & weiss , n. o. 1981 , , 197 , 45p komm , r.w . ,
howard , r.f . , & harvey , j.w .
1993 , , 147 , 207 kovri , z. , bartus , j. , strassmeier , k.g . , et al .
2007 , , 474 , 165 kovari , z. , kriskovics , l. , knstler , a. , et al .
2014 , arxiv:1411.1774 krause , f. , rdiger , g. 1974 , astronomische nachrichten , 295 , 93 krause , f. , & raedler , k .- h . 1980 , mean - field magnetohydrodynamics and dynamo theory ( oxford : pergamon ) kitchatinov , l.l . ,
& ruediger , g. 1995 , , 299 , 446 kosovichev , a.g . ,
schou , j. , scherrer , p.h . , et al .
1997 , , 170 , 43 masada , y. 2011 , , 411 , l26 masada , y. , yamada , k. , & kageyama , a. 2013 , , 778 , 11 masada , y. , & sano , t. 2014 , , 66 , s2 masada , y. , & sano , t. 2014 , , 794 , ll6 miesch , m.s . , elliott , j.r . , toomre , j. , et al .
2000 , , 532 , 593 miesch , m. s. 2005 , living reviews in solar physics , 2 , 1 mitra , d. , tavakol , r. , kpyl , p.j . , & brandenburg , a. 2010 , , 719 , l1 moffatt , h.k .
1978 , magnetic field generation in electrically conducting fluids , cambridge university press pedlosky , j. 1987 , geophysical fluid dynamics , by joseph pedlosky .
springer , new york .
1987 reiners , a. 2012 , living reviews in solar physics , 9 , 1 rempel , m. 2005 , apj , 622 , 1320 ruediger , g. 1989 , differential rotation and stellar convection .
sun and the solar stars , berlin : akademie verlag , 1989 schou , j. , antia , h.m . ,
basu , s. , et al .
1998 , , 505 , 390 schrinner , m. , petitdemange , l. , & dormy , e. 2012 , , 752 , 121 simard , c. , charbonneau , p. , & bouchat , a. 2013 , , 768 , 16 spiegel , e. a. , & zahn , j. p. 1992
, , 265 , 106 sprague , m. , julien , k. , knobloch , e. , & werne , j. 2006 , journal of fluid mechanics , 551 , 141 spruit , h. c. , nordlund , a. , & title , a. m. 1990 , , 28 , 263 steffen , m. , & freytag , b. 2007 , astronomische nachrichten , 328 , 1054 strassmeier , k.g .
, kratzwald , l. , & weber , m. 2003 , , 408 , 1103 thompson , m.j . , christensen - dalsgaard , j. , miesch , m.s .
, & toomre , j. 2003 , ara&a , 41 , 599 ulrich , r.k .
2010 , , 725 , 658 weber , m. , strassmeier , k.g . , & washuettl , a. 2005 , astronomische nachrichten , 326 , 287 zhao , j. , bogart , r.s . ,
kosovichev , a.g . ,
duvall , t.l . , jr . , & hartlep , t. 2013 , , 774 , ll29
|
effects of magnetic field on stellar differential rotation are studied by comparing magnetohydrodynamic ( mhd ) models and their hydrodynamic ( hd ) counterparts in the broad range of rotation rate and in varying initial rotation profile .
fully - compressible mhd simulations of rotating penetrative convection are performed in a full - spherical shell geometry .
critical conditions for the transition of the differential rotation between faster equator ( solar - type ) and slower equator ( anti - solar type ) are explored with focusing on the rossby number ( @xmath0 ) " and the convective rossby number ( @xmath1 ) " .
it is confirmed that the transition is more gradual and the critical value for it is higher in the mhd model than the hd model in the view of the @xmath1-dependence .
the rotation profile shows , as observed in earlier studies , the bistability near the transition in the hd model , while it disappears when allowing the growth of magnetic fields except for the model with taking anti - solar type solution as the initial condition .
we find that the transition occurs at @xmath2 both in the mhd and hd models independently of the hysteresis .
not only the critical value , the sharpness of the transition is also similar between the two models in the view of the @xmath0-dependence .
the influences of the dynamo - generated magnetic field and/or the hysteresis on convective motion are reflected in the @xmath0 .
this would be the reason why the transition is unified in the view of the @xmath0-dependence .
we finally discuss the @xmath0-dependence of magnetic dynamo activities with emphasis on its possible relation to the kinetic helicity profile .
|
fascinating insights into the formation of atomic chains consisting only of a few atoms suspended between two electrodes have been obtained by transmission electron microscopy and mechanically controllable break junction techniques.@xcite these experiments triggered the imagination of scientists to use such atomic - scale junctions as future electronic devices by exploiting their unique properties , e.g. ballistic electronic transport.@xcite due to their reduced dimensions , atomic chains have also been predicted to develop magnetic moments even for elements nonmagnetic in bulk such as ir , pt , or pd.@xcite recently , an indirect proof has been given that transition - metal chains in break junctions are in general magnetic.@xcite mastering the intriguing magnetic properties of suspended chains would enable spintronic applications based on the unique possibility to probe , control and switch the magnetic states by spin - polarized currents.@xcite the orientation of the magnetic moments is stabilized against thermal fluctuations by the magnetocrystalline anisotropy energy ( mae ) arising from spin - orbit coupling .
theoretical studies of this key quantity performed for idealized systems such as free - standing infinite monowires ( mws ) and small clusters suggest giant values for @xmath1- and @xmath0-transition - metals and switching of the easy axis upon stretching the wires . @xcite
even the size of the magnetic moment itself can crucially depend on the magnetization direction , an effect coined as colossal magnetic anisotropy.@xcite the ballistic conductance in such wires varies with the orientation of the magnetization direction with respect to the chain axis giving rise to ballistic anisotropic magnetoresistance ( bamr).@xcite while the predicted effects probably occur in real atomic - scale junctions , the theoretical studies have either focused on idealized systems or neglected varying interatomic distances in the chains.@xcite however , for small suspended chains variations in the interatomic distance due to stretching the electrodes are unavoidable and can provide a unique way to control the properties of the system , not available in higher dimensions.@xcite here , we go beyond such theoretical studies by explicitly taking the contacts into account and by investigating the development of magnetism in atomic - scale junctions upon increasing the distance between the electrodes .
we perform first - principles calculations for w , ir and pt junctions including spin - orbit coupling and structural relaxations and focus on atomic chains of three atoms suspended between two bulk - like bcc-(001 ) electrodes as shown in fig .
these transition - metals are common tip materials in scanning tunneling microscopy , and ir and pt have been shown to form long atomic chains in break junction experiments.@xcite in particular we concentrate on the influence of varying tip - to - tip separation on the magnetic properties of the junctions .
we observe that relaxations of the chain atoms upon stretching are in accordance to observed expermentally trends of chain formation for these elements.@xcite we also find that large spin moments develop in the suspended trimers upon stretching and on the central atom they reach similar values as in infinite monowires.@xcite interestingly , the w trimer behaves as a single spin impurity as only the central atom develops a sizable magnetic moment . by including spin - orbit coupling ( soc )
, we obtain the mae of the suspended chains and observe large values of 10@xmath230 mev per atom .
surprisingly , we find that the contribution to the mae from central chain atoms and atoms in contact with the electrodes can have opposite signs and are of similar giant magnitude of a few tens mev per atom
. the magnitude of this effect , unknown in bulk or on surfaces may lead to a strong tendency towards non - collinear magnetic order in atomic - scale junctions even when the exchange interactions are small or profoundly collinear.@xcite we prove the generality of this phenomenon by showing that it occurs for chains of different length with different geometry of the contacts .
further , the mae may exhibit changes of sign upon stretching , i.e. the easy magnetization direction switches from along the chain to perpendicular to it .
we demonstrate that switching may be observable experimentally by measuring the ballistic conductance of the junctions .
( color online ) upper panel : three - atoms chain suspended between two bcc(001)-contacts . @xmath3 denotes the length of the unit cell consisting of 17 atoms , while @xmath4 and @xmath5 give the distance between the first plane of the contact to the apex atom and from the apex atom to the central atom of the trimer , respectively .
lower panel : relaxation of the @xmath5 and @xmath4 bonds in suspended trimers of w , ir , and pt upon increasing @xmath3 . for comparison ,
the bond length for ideal stretching with fixed apex atom positions is given ( dashed line).,title="fig : " ] + ( color online ) upper panel : three - atoms chain suspended between two bcc(001)-contacts .
@xmath3 denotes the length of the unit cell consisting of 17 atoms , while @xmath4 and @xmath5 give the distance between the first plane of the contact to the apex atom and from the apex atom to the central atom of the trimer , respectively .
lower panel : relaxation of the @xmath5 and @xmath4 bonds in suspended trimers of w , ir , and pt upon increasing @xmath3 . for comparison ,
the bond length for ideal stretching with fixed apex atom positions is given ( dashed line).,title="fig : " ] we have performed first - principles calculations within the generalized gradient approximation ( gga ) to the density functional theory ( dft ) .
we employed the full - potential linearized augmented plane - wave method for one - dimensional ( 1d ) systems ( 1d - flapw),@xcite as implemented in the ` fleur ` code.@xcite basis functions were expanded up to @xmath6 a.u.@xmath7 and we have used 8 @xmath8-points in one half of the 1d brillouin zone .
considering more @xmath8-points does not change the values of the spin moments and magnetic anisotropy energies significantly .
the contact structure has first been relaxed separately in all three dimensions and then kept fixed for all contact separations @xmath3 ( fig . 1 ) .
the apex atoms of the trimer have been relaxed along the chain direction ( @xmath9 ) upon varying @xmath3 , while the central atom is fixed by symmetry ( fig .
[ fig1 ] ) .
calculations have been performed for junctions stretched from the equilibrium up to @xmath10 = 3 a.u . for pt and ir trimers we considered the ferromagnetic ground state , while w trimer always converges to the antiferromagnetic solution .
spin - orbit coupling soc was included into the calculations in a non - perturbative way.@xcite we define the magnetic anisotropy energy ( mae ) as the total energy difference between configurations with a magnetization along the chain axis ( @xmath9-direction ) , and perpendicular to it ( @xmath11-direction ) .
the mae was calculated employing the force theorem , and for several electrode stretchings , @xmath10 , we checked the force theorem results for the trimer s mae against self - consistent calculations , finding good qualitative agreement .
taking into account the magnitude of the obtained mae we can safely neglect the magnetic dipole - dipole contribution . in order to check the applicability of our 1d geometry for the electrodes , which allows us to efficiently perform the demanding calculations with structural relaxations and including soc
, we have also considered a geometry with electrodes represented by surfaces , c.f .
due to the immense computational cost , we have restricted our test to the most critical case of a pt trimer and chose an elongation of @xmath12 a.u .
for these calculations we used the film version of the ` fleur ` code .
we performed a single super - cell calculation in which the trimer was suspended between two electrodes consisting of a four pt atom square base deposited onto three layers of fcc pt(001 ) .
this configuration will be referred to as `` 2d '' in the following , fig . 2 .
the ideal bulk values of pt with an in - plane lattice constant of 5.23 a.u .
were taken for the four contact atoms and the bulk - like leads .
the positions of the trimer s atoms were set to their relaxed values at @xmath12 a.u .
obtained within the wire geometry of fig . 1 .
the in - plane separation between the neighboring trimers was set to 15.7 a.u .
we used a basis functions cut - off of @xmath13 a.u.@xmath7 and 15 @xmath8-points in the irreducible part of the 2d brillouin zone , while 144 @xmath8-points in the whole brillouin zone were used in order to determine the mae with the force theorem . along the trimer s axis only the @xmath14-point was used .
( color online ) 2d configuration , and b1 and b2 cases of the pt trimer considered for the separation of pt electrodes of @xmath12 and 1.8 a.u .
, respectively .
in 2d configuration the trimer is suspended between two electrodes consisting of a four pt atom square base deposited onto three layers of fcc pt(001 ) .
green color in b1 and b2 marks the pt atoms of the contacts in which the spin moment was artificially quenched . in accordance to fig . 1 , red color marks the central atom and the blue color marks the apex atoms of suspended trimer .
for more details see text . ]
under the condition of reduced coordination of transition metal atoms in a monoatomic chain suspended between thicker leads , a situation realized e.g. in a break junction experiment , the distance @xmath15 between the chain atoms becomes of utter importance . upon pulling apart the leads during the chain elongation process@xcite @xmath15 is constantly changing which has dramatic consequences for the magnetic properties of suspended chains , in particular for heavy @xmath1 and @xmath0 transition - metals.@xcite by increasing the separation @xmath10 between the leads and accounting at the same time for the changes in the chain s interatomic distances @xmath2 which was neglected in most of the previous _ ab initio _ studies of magnetism in suspended chains @xmath2 we aim at mimicking a real break junction experiment .
the relaxations of the trimers along the chain s symmetry @xmath9-axis are shown in fig .
[ fig1 ] , where the bond lengths @xmath5 and @xmath4 are presented as a function of contact separation @xmath3 . for w with an almost half - filled @xmath0-shell and large bulk cohesive energy almost the entire increase in contact separation enters the @xmath5 bond length , while the @xmath4 bond between the apex atom and the contact is nearly unaffected .
in contrast , for ir and pt , some of the stretching transfers into an increased bond length at the contact . the apex atom of the ir trimer has a stronger bond to the contact than to the central atom , which is reflected in the linear regime of @xmath16 for @xmath17 a.u . for pt ,
the @xmath5 bond is significantly stronger resulting in a remarkable stretching of the @xmath4 bond even at large electrode separations , which in a realistic break junction will eventually result in chain elongation.@xcite overall , our predicted trend is in accordance with experimental observations of increased probability for chain formation when going from w to pt.@xcite ( color online ) local spin moments of the central ( @xmath18 ) and apex ( @xmath19 ) atoms ( denoted as `` c - atom '' and `` a - atom '' , respectively ) of the suspended ( a ) w , ( b ) ir and ( c ) pt trimers as a function of the stretching of the contacts ( without soc ) . for comparison the values for an infinite free monowire ( mw )
are given by dashed lines .
the interatomic distance in the infinite mw corresponds to the distance @xmath5 for a given @xmath10 . for pt triangles
down and triangles up stand for the values of the apex ( filled ) and central ( open ) spin moments in b2 and 2d cases , respectively .
for details see text . ]
the magnetic properties of suspended @xmath0 tm atoms are very sensitive to the local environment . in most previous studies it was common to model the contacts by semi - infinite surface electrodes.@xcite
this restriction on the geometry of the contacts is quite strong in particular for @xmath0 tm break junctions , as the surface geometry leads to suppressed spin moments of the atoms in suspended wire and contact atoms closest to it.@xcite however , it is well - established that the shape of the contacts in a break junction is rather a thinning wire - like geometry , with the reduced coordination number of the contact atoms adjacent to the chain which greatly enhances their tendency towards magnetism.@xcite in this respect our contact geometry is probably closer to the real situation , although it might overestimate the tendency of the contacts to magnetism . in order to investigate this point quantitatively ,
we have performed calculations in the setup shown in fig . 2 , which are discussed further below .
the calculated spin moments @xmath20 inside the atomic spheres of the suspended trimers upon increasing the distance between the contacts @xmath10 are shown in fig .
[ fig2 ] . for w
, the central atom reveals a sizable magnetic moment of 1.7@xmath21 already at @xmath22 which further increases upon stretching , while the moments of all other atoms are negligible . in this respect
the w trimer presents a unique system of a single spin impurity between nonmagnetic leads . at small
stretching , the spin moments of the ir trimer are rather small with significantly larger moments of the central atom . upon further stretching the apex and central moments rise rapidly , reaching large values of 1.4@xmath21 and 2@xmath21 , respectively . both , the @xmath18 spin moments of the central w and ir chain atom follow the behavior in an infinite monowire very closely as seen in fig .
[ fig2](a ) and ( b ) .
the pt trimer shows a trend similar to ir upon stretching , only with smaller moments @xmath2 this is in contrast to an infinite pt mw , where the spin moment is zero for a large interval of interatomic distances without soc .
it is resurrected upon including soc , but only when the magnetization points along the @xmath9-axis , an effect coined as colossal magnetic anisotropy.@xcite in contrast , for the suspended pt trimer , the moments of the trimer atoms are non - zero already without soc for a wide range of @xmath5 . upon including soc
the spin moments do not change dramatically and the moments are non - zero for both magnetization directions in the whole range of @xmath10 , which manifests the subtle magnetism of pt chains and its sensitivity to local environment . in the critical case of pt
, we analyze in more detail the influence of the contact geometry on the spin moments of the trimer . at the stretching of @xmath23 a.u .
we performed a calculation in which we artificially quenched the spin moments in the contacts by applying a small magnetic field inside the muffin - tin spheres of the pt atoms , situation referred to as `` b2 '' in the following ( see fig . 2 ) .
we observe that @xmath19 drops significantly from 0.55@xmath21 to 0.18@xmath21 , while @xmath18 drops from 0.89@xmath21 to 0.54@xmath21 ( fig .
3(c ) ) @xmath2 a value , very close to that reported in ref .
[ ] , where contacts were modelled by infinite bulk electrodes . however , quenching the spin moment in the deeper parts of the contacts only ( `` b1''-case , fig .
2 ) @xmath2 the situation which is probably closer to that in real experiments @xmath2 almost does not affect the values of @xmath18 and @xmath19 . finally , we compare the results obtained within our 1d geometry of the electrodes with the situation of surface - like electrodes , as shown in fig .
for a separation of @xmath12 a.u .
@xmath18 is reduced in the 2d geometry by only 25% from 1.03@xmath21 in the wire geometry to 0.79@xmath21 , while the moment of the apex atom is affected stronger , dropping from 0.64@xmath21 to 0.26@xmath21 ( fig .
3(c ) ) . in a more realistic situation
the coordination of the contact atoms next to the apex atoms is more reduced and the moments of the trimer s atoms will increase approaching the values obtained in our wire geometry
. however , we have to conclude that magnetism in suspended short pt chains with their small moments is extremely sensitive to slightest changes in the contact geometry .
therefore , observing this magnetism experimentally will be a hard task , even considering the cluster s large magnetic anisotropy energies .
( color online ) magnetic anisotropy energies of suspended trimers of ( a ) w , ( b ) ir and ( c ) pt as a function of the stretching of the contacts @xmath10 .
the total trimer s anisotropy @xmath24 ( black half - filled circles ) is decomposed into the contribution from the central @xmath25 ( red open circles ) and apex atoms @xmath26 ( blue filled circles ) ( for definitions of the quantities see text ) .
black dashed line stands for the mae of the infinite mws ( per atom ) . for pt squares stand for the @xmath26 ( filled ) and @xmath25 ( open ) tetramer values , while triangles up and triangles down stand for the corresponding values in b2 and 2d cases .
for details see text . ]
recent theoretical studies,@xcite which are in accordance with experiments,@xcite clearly state that the interatomic distance @xmath5 in suspended chains of @xmath0 elements consisting of several atoms can reach large values of [email protected] a.u . upon stretching , which corresponds to @xmath27 a.u .
, c.f .
therefore , in a real pt , ir , or w break junction experiment the central atom will reveal strong fingerprints of spin - polarization at sufficiently small temperatures . in order to study the thermal stability of magnetism in the suspended trimers ,
next we concentrate on their magnetic anisotropy energies . in a transport break junction experiment ,
the data is obtained by averaging over thousands of measurements , which differ from each other by the thinning history of the wire and the contact geometry .
therefore , in order to analyze this data with respect to the magnetism of the suspended chain , it is necessary to disentangle the contributions of intrinsic trimer s magnetic anisotropy energy @xmath24 from that originating from the contacts . in our calculations , we do this by switching soc off in the contact s atomic spheres .
moreover , we individually switch soc on and off in the apex and central atoms , to determine their separate contributions , @xmath28 and @xmath29 , respectively , to the total mae of the trimer , @xmath24 . for w with essentially one magnetic atom in the trimer
we define @xmath24 as @xmath30 , while for ir and pt we define the total mae of the trimer per atom as @xmath31 . @xcite the absolute value of trimer s mae , shown in fig . 4 with a black solid line for w , ir and pt , reach large absolute values of 10 to 30 mev per atom at most electrode separations .
these values , which are one to two orders of magnitude larger than those in most of transition - metal nanostructures , give us confidence that the magnetism of suspended trimers can be tackled experimentally .
for all three elements the calculated values of anisotropy energies @xmath24 are of the order of magnitude of those predicted for infinite monowires ( dashed line in fig . 4 and refs .
interestingly , the trimer s mae displays a non - trivial dependence on @xmath10 quite different from that in ideal chains for ir and w , while for pt both anisotropy energies are close in their trend . for w , fig .
4(a ) , the only contribution to @xmath24 stems from the strongly spin - polarized central atom .
starting at about 35 mev and a magnetization along the chain , there is a switch to a perpendicular direction upon stretching by @xmath32 a.u . upon further increasing the contact separation ,
the perpendicular magnetization is stabilized by a sizable mae of 10@xmath220 mev .
a switch of the magnetization direction can be also observed for ir at significant stretching of around @xmath33 a.u . , fig .
4(b ) . in general , for ir
, the behavior of @xmath24 upon stretching is quite smooth , and the @xmath9-direction of the magnetization is favored by considerable 15@xmath230 mev over a wide interval of contact separations .
detecting traces of magnetism in pt break junctions at small stretching will present a significant challenge as can be seen from fig .
4(c ) . in the regime of @xmath34 a.u .
the value of @xmath24 is on the order of a few mevs , posing the question of whether the sensitive magnetization will survive temperature fluctuations .
moreover , in this regime , as pointed out above , the spin moments of the apex and central atoms depend strongly on contact geometry and details of the thinning process
. this will influence the contribution of these atoms to @xmath24 , causing frequent measurement - to - measurement oscillations in its sign and magnitude . beyond a distance of @xmath35 a.u .
the well - developed magnetization of the trimer is pointing rigidly along the chain axis with a mae of about 15 mev @xmath2 a value somewhat smaller than that of an infinite mw of corresponding interatomic distance , c.f .
4(c ) . as a general trend
, we observe in fig .
4 competing contributions to the mae from central and apex atoms for ir and pt atomic junctions . at most separations
a positive value from the apex atom , @xmath36 , favors a perpendicular magnetization direction , while negative contribution from the central atom , @xmath37 , forces the magnetization along the trimer s axis . due to this competition
the trimer s mae behaves qualitatively differently from that in the infinite mw as a function of the interatomic distance for ir .
it significantly quenches the central atom s mae , so that the resulting total anisotropy is on average smaller than anticipated from the infinite mw both for ir and pt .
the origin of this competition is the different local symmetry of apex and central atoms and it exists also in longer chains or chains made of other elements . to demonstrate this point we additionally calculated the mae of a suspended four - atom chain of pt atoms with the contacts as in fig . 1 .
the @xmath28 and @xmath29 anisotropy energies for the two apex and two central atoms , respectively , are displayed by squares in fig .
4(c ) ( per atom ) and reveal the same effect : @xmath28 and @xmath29 are close in their values but opposite in sign .
due to its large magnitude , this effect may even compete with exchange interactions in suspended small chains and lead to a non - collinear magnetic order .
with increasing chain length the anisotropy energies of the atoms in the center of the chain will eventually approach that of an atom in the infinite monowire ( this can be already seen in fig .
4(c ) for pt when going from a trimer to a tetramer ) , and the whole cluster will behave as a superparamagnet . for smaller suspended clusters , however , which are much more probable to occur in an experiment , the effect of competition between central and apex atoms will be dramatic .
( color online ) spin - decomposed local density of states ( ldos ) for a pt trimer including soc for a separation between the electrodes of @xmath23 a.u .
( a ) and ( b ) : @xmath38 ( @xmath39,@xmath40 ) and @xmath41 ( @xmath42,@xmath43 ) @xmath44-ldos of the central ( `` c - atom '' ) and apex ( `` a - atom '' ) atoms for the magnetization along the @xmath9-axis .
( c ) ldos of the central atom for two different magnetization directions ( @xmath9 and @xmath11 ) . ]
we demonstrate that the effect of the competition between the apex and central atoms for the mae value is also stable with respect to the geometry of the contacts . for this purpose
we calculate the mae of the pt trimer in 2d - configuration , as well as in b1 and b2 cases , and present the calculated values in fig .
as far as the 2d - case is concerned , we observe that despite significant changes in the spin moments , c.f .
3(c ) , the values of @xmath28 and @xmath29 are very close to those calculated within the wire geometry , moreover , they are also opposite in sign .
their competition results in rather close values for the total @xmath24 of 7 and 18 mev for the 2d and wire - like geometry of the contacts , respectively , both favoring the @xmath9-direction of the trimer s magnetization .
notably , both values are very close to that of 12 mev reported by smogunov _ et al . _ in ref . [ ] for the suspended between semi - infinite electrodes pt trimer , although one has to keep in mind the difference between the geometries of all three cases . in the b1-situation , in which we do not quench the spin moments of the contact atoms of the electrode ,
the mae is almost not affected compared to the wire setup of fig . 1 .
on the other hand , in the b2-case , in which we quench the spin moments inside the entire electrode , the spin moments of the trimer atoms are strongly reduced , c.f . fig .
3(c ) , and so is the mae : @xmath29 and @xmath28 drop to almost half their value .
however , the latter two contributions to the total mae @xmath24 are still of opposite sign , proving the generality of the phenomenon of competing anisotropies for this type of atomic junctions .
the latter effect is responsible for the value of the total mae @xmath24 of 9 mev , smaller than the value of 20 mev obtained within the wire - geometry .
to understand the effect , we analyze the local density of states ( ldos ) and choose the pt trimer as an example .
in an infinite mw , the position of the localized @xmath15-states with respect to the fermi energy ( @xmath45 ) is responsible for the formation of the orbital moment and the direction of the magnetization.@xcite according to the symmetry these @xmath15-states can be subdivided into @xmath38 ( @xmath39,@xmath40 ) and @xmath41 ( @xmath42,@xmath43 ) groups , not taking into account @xmath46 @xmath47-hybrid.@xcite in a pt ( ir ) infinite mw , the position of the x - edge ( @xmath14-edge ) of the @xmath41 degenerate @xmath43 and @xmath42 bands at @xmath45 is responsible for the easy magnetization direction along the chain.@xcite in a trimer at larger stretching , the @xmath0-ldos of a central atom qualitatively follows that of an atom in an infinite monowire ( not shown ) , which explains the same sign of the mae ( fig .
[ fig3 ] ) . a small splitting between the @xmath42 and @xmath43 states due to the presence of the leads can be already seen for the central atom ( fig .
5(a ) ) , however , for the apex atom the effect is dramatic and has crucial consequences for the apex mae .
interaction with the contacts locally breaks the @xmath48 symmetry of an infinite chain which results in large splitting between the @xmath43- and @xmath42-orbitals in the @xmath41-band and their shift towards the lower energies ( fig .
on the other hand , the degeneracy between the @xmath39 and @xmath40-states ( @xmath38-band ) is locked by symmetry .
this leads to the rearrangement of the states around @xmath45 and as a result the in - plane direction of the apex spin moment becomes favorable ( fig .
4 ) . in general
, it should be possible to deduce the predicted switches in the magnetization direction of the suspended chains from transport measurements in these systems .
as an example , in fig .
5(c ) we present the ldos of the central atom in a pt trimer at @xmath23 a.u . upon changing the magnetization direction significant changes in the ldos around the fermi energy @xmath45
can be observed ( shaded area in fig .
5(c ) ) , which will inevitably result in large changes of the experimentally measured conductance .
as far as the @xmath0 tms are concerned , giant values of the ballistic anisotropic magnetoresistance should be achievable due to strong spin - orbit coupling in these metals , which is able to modify the electronic structure significantly in response to the changes of the magnetization direction in real space .
by performing _
ab initio _ calculations of suspended trimers of w , ir and pt including structural relaxations as a function of the separation between the leads , we demonstrate that the chain atoms develop significant spin moments upon stretching .
we investigate the stability of these spin moments via evaluating the magnetic anisotropy energy of the trimers .
our calculations reveal large mae of the whole trimers of the order of 10 to 30 mev per atom .
interestingly , we observe that the total mae presents a competition between large contributions from the apex and central atoms .
we argue that this effect is general and can occur in suspended chains of different elements and different length , leading to non - trivial real space textures of magnetic anisotropy energy which might even lead to complex magnetic ordering in atomic - scale junctions . under the condition of large predicted values of mae huge magnetic fields would be required to change the magnetization of the chains , therefore , new ways to achieve this goal have to be established . one of the possibilities lies in using the ability to control the magnetization direction in the junction by changing the distance between the leads .
we show that such switches of the magnetization can happen and will result in strong features in the measured conductance .
distinguishing these features from the changes in conductance due to chain elongation or atomic rearrangements presents a considerable challenge .
39 a. sokolov , c. zhang , e.y .
tsymbal and j. redepenning , nat . nanotech .
* 2 * , 171 ( 2007 ) r.h.m .
smit , c. untied , a.i .
yanson and j.m .
van ruitenbeek , phys .
lett . * 87 * , 266102 ( 2001 ) t. kizuka , phys .
b * 77 * , 155401 ( 2008 ) .
a. smogunov , a. dal corso and e. tosatti , nat . nanotech . * 3 * , 22 ( 2008 ) j. velev , r.f .
sabirianov , s.s .
jaswal and e.y .
tsymbal , phys .
* 94 * , 127203 ( 2005 ) .
a. delin and e. tosatti , phys .
b * 68 * , 144434 ( 2003 ) a. delin , e. tosatti and r. weht , phys . rev
. lett . * 92 * , 057201 ( 2004 ) a. thiess , y. mokrousov , s. heinze and s. blgel , phys . rev . lett . * 103 * , 217201 ( 2009 ) .
p. lucignano , r. mazzarello , a. smogunov , m. fabrizio and e. tosatti , nat .
materials * 8 * , 563 ( 2009 ) .
calvo , j. fernndez - rossier , j.j .
palacios , d. jacob , d. natelson and c. untied , nature * 458 * , 1150 ( 2009 ) .
y. mokrousov , g. bihlmayer , s. heinze and s. blgel , phys .
lett . * 96 * , 147201 ( 2006 ) t.o .
strandberg , c.m .
canali and a.h .
macdonald , phys .
b * 77 * , 174416 ( 2008 )
. l. fernndes - seivane and j. ferrer , phys .
lett . * 99 * , 183401 ( 2007 ) .
a. smogunov , a. dal corso and e. tosatti , phys .
b * 78 * , 014423 ( 2008 ) .
smelova , d.i .
bazhanov , v.s .
stepanyuk , w. hegert , a.m. saletsky and p. bruno , phys .
b * 77 * , 033408 ( 2008 ) . m. czerner , b.yu .
yavorskiy and i. mertig , phys .
b * 77 * , 104411 ( 2008 ) .
a. thiess , y. mokrousov , s. blgel and s. heinze , nano letters * 8 * , 2144 ( 2008 ) .
s. mankovsky , s. bornemann , j. minr , s. polesya , h. ebert , j.b .
staunton and a.i .
lichtenstein , arxiv:0902.3336v1 ( 2009 ) .
y. mokrousov , g. bihlmayer and s. blgel , phys .
b * 72 * , 045402 ( 2005 ) . c. li , a.j .
freeman , h.j.f .
jansen and c.l .
fu , phys .
b * 42 * , 5433 ( 1990 ) c. untiedt , m.j .
caturla , m.r .
calvo , j.j .
palacios , r.c .
segers and j.m .
van ruitenbeek , phys .
lett . * 98 * , 206801 ( 2007 ) . l.g .
rego , a.r .
rocha , v. rodrigues and d. ugarte , phys .
b * 67 * , 045412 ( 2003 ) . d. cheng , w.y .
kim , s.k .
min , t. nautiyal and k.s .
kim , phys .
lett . * 96 * , 096104 ( 2006 ) .
e. z. da silva , a. j. r. da silva , a. fazzio , phys .
lett . * 87 * , 256102 ( 2001 ) s. b. legolas , d. s. galvo , v. rodrigues , d. ugarte , phys .
. lett . * 88 * , 076105 ( 2002 ) in reality the value of the apex mae is obtained by substracting the mae of the central atom from the mae of the whole trimer .
g. van der laan , j. phys .
matter * 10 * , 3239 ( 1998 ) a. thiess , y. mokrousov and s. heinze ( in preparation ) .
|
using first - principles calculations , we study the magnetism of @xmath0 transition - metal atomic junctions including structural relaxations and spin - orbit coupling . upon stretching monatomic chains of w , ir , and pt suspended between two leads , we find the development of strong magnetism and large values of the magnetocrystalline anisotropy energy ( mae ) of up to 30 mev per chain atom .
we predict that switches of the easy magnetization axis of the nanocontacts upon elongation should be observable by ballistic anisotropic magnetoresistance measurements . due to the different local symmetry
, the contributions to the mae of the central chain atoms and chain atoms in the vicinity of the leads can have opposite signs which reduces the total mae .
we demonstrate that this effect occurs independent of the chain length or geometry of the electrodes .
|
the primary differences between the winds of b stars and those of o stars are the much lower densities of b star winds and the relative importance of euv / x - ray emission in b stars . in this paper we will examine the ionization and temperature structure of b star winds , with a focus on the euv and x - ray emission properties of these objects .
we will see how the differences in the wind and euv / x - ray properties cause strong departures from the thermal and ionization conditions found in the more well - studied and well - understood o stars .
the ionization balance in b star winds is both more complex , and harder to determine , both theoretically and observationally , than in o star winds .
in fact , because of the more uncertain ionization corrections to the uv mass - loss rates the actual mass - loss rates of b stars are not well known .
in addition to this crucial difference between the high wind densities of o stars and the low wind densities of b stars , the relative proximity of the nearest b stars provides us with some important information that does not exist for o stars . due to the very low column densities toward @xmath2 cma and @xmath3 cma
, we have information about the extreme ultraviolet ( euv ) emission properties both of the photosphere and of the wind of these two objects , while we have no comparable information about o stars .
as we shall see , this euv emission plays an important role in the setting the ionization conditions ( and possibly the thermal conditions ) in the low - density b star winds .
the dense winds of o stars give rise to ultraviolet absorption lines that show both p cygni emission peaks and broad absorption troughs .
however , in b stars the winds are not strong enough to generate emission components , and the observed blue - shifted absorption components are seen to be quite weak in the few ions that are observable in the spacecraft ultraviolet .
grigsby & morrison ( 1995 ) have shown that uv wind features ( at least in those ions accessible to iue ) do not exist beyond about b2 . in principle
, radiation - driven winds can exist if there is at least one optically thick line in the wind .
gayley ( 1995 ) has developed a new formalism for describing the radiation force in standard line - driven winds ( castor , abbott , & klein 1975 ; hereafter cak ) .
he has shown that the line opacity in a stellar wind has the remarkably simple character that its opacity is @xmath4 times stronger than free electron scattering .
this includes the effects of elemental abundances and assumes a matching between the wavelengths at which the line opacity exists and those at which the photospheric spectrum is strongest .
therefore , stars with eddington factors @xmath5 have the potential for line - driven winds .
however , puls ( 2000 ) has recently shown that for b stars , the opacity and photospheric flux may not be well matched , making it harder for b stars to drive a strong outflow . in any case , this @xmath5 limit corresponds to stars that have somewhat later spectral types than the b2 limit seen in the uv wind lines .
the disappearance of these spectral wind features may , however , be due to ionization effects , rather than the disappearance of winds altogether . as we have already pointed out ,
determining the mass - loss rates of b stars is very difficult because the ionization corrections to uv absorption line observations is so uncertain , and no other reliable mass - loss diagnostics exist for most of these stars . despite the findings of grigsby and morrison ( 1995 ) ,
there is , in fact , some evidence for winds in mid- to late - b , and even early - a , stars .
much of the evidence is indirect and centers on the x - ray emission ( which we will discuss in 3 ) , but one intriguing piece of more direct evidence involves the detection of a blue - shifted absorption component in the lyman-@xmath6 line of the a0 v star sirius ( bertin et al . 1995 ) .
this feature is consistent with a wind having a terminal velocity of @xmath7 km s@xmath8 and a mass - loss rate of @xmath9 .
so , it seems that at least some stars with @xmath10 have winds , although they may _ not _ be driven ( primarily ) by radiation pressure acting on lines . in the remainder of this paper
we discuss some results from large surveys but concentrate on detailed observations of the b2 ii mk standard @xmath2 cma . with a wind - broadened ( @xmath11 km s@xmath8 ) , but weak , 1548 , 1551 profile and a mass - loss rate of @xmath12 to @xmath13 ( cassinelli et al .
1995 ; cohen et al .
1999 ) this star has a much lower density wind than o stars , but , unlike some later b stars , the existence of its wind is not in doubt .
we note that with @xmath14 the multiplier factor of @xmath4 implies that the force from spectral lines exceeds gravity at the star s surface .
this star has @xmath15 k and @xmath16 .
it is very well studied and , most importantly , is the only hot star observable with all three of the extreme ultraviolet explorer ( ) spectrometers .
this unique data set alone argues for concentrating on when trying to understand radiation - driven winds in b stars .
direct measurements of stellar photospheric continua below the lyman edge at 912 were impossible until the early 1990s when the satellite was launched . even with this instrument , only a small number of stars could be observed , due to a combination of low intrinsic fluxes and high interstellar opacity . of these stars ( aside from white dwarfs ) , all of the observations are of coronal emission lines from late - type stars , except for two stars : and .
these are the only two stars for which there are direct measurements of the photospheric flux below the lyman edge .
this is due primarily to their fortuitous position in a tunnel of very low interstellar column density . in general ,
any study that has required the photospheric euv fluxes of hot stars including studies of the warm ionized component of the interstellar medium as well as studies of stellar wind conditions has relied on model atmospheres .
there are also a very small number of _ indirect _ observations of euv fluxes of b stars .
one of these involves fabry - perot spectroscopy of h@xmath6 fluxes of small regions around b stars .
assuming that all stellar lyman continuum radiation is reprocessed into h@xmath6 in these regions , kutyrev , reynolds , & tufte ( 1993 ) and reynolds ( 1995 ) placed lower limits to the stellar lyman continuum fluxes that in some cases exceed the predictions of model atmospheres .
the direct measurements of the lyman continuum fluxes of and ( b1 iii ) with revealed fluxes that are 10 to 100 times higher than the predictions of model atmospheres ( vallerga , vedder , & welsh 1993 ; cassinelli et al . 1995 ; cassinelli et al .
the discrepancy seems to be even higher in the continuum ( @xmath17 ) , which is observed in but not , although the correction for interstellar attenuation is more uncertain at these wavelengths .
non - lte , line - blanketed tlusty models ( hubeny & lanz 1995 ) actually provide worse fits to the data , as the additional physical effects tend to both cool the outer wind , and lead to an overpopulation of the ground state with respect to lte . both these effects lower the lyman continuum flux .
cassinelli et al .
( 1995 ; 1996 ) suggest that whatever the physical mechanism , the euv excesses and the observed ir excesses can both be accounted for by a modest increase in the temperature with respect to models in the extreme outer layers of the atmosphere ( @xmath18 ) , perhaps of 1000 to 2000 k. the ir continuum is dominated by free - free opacity , and as such , is in lte .
furthermore , the ir continuum near 15 is formed in the _ same physical layers _ as the euv continuum near 600 .
the 1000 k to 2000 k temperature increase that can explain the observed lyman continuum excess in and can also explain the roughly 10% ir excess observed with .
although non - lte effects could in principle explain the observed euv excess , only the existence of a temperature above the predictions of models high in the atmosphere can _ simultaneously _ explain the euv and ir excesses .
it is not clear how representative these two stars are , or how large the region of the hr diagram is for which b star euv fluxes have been underestimated .
however , the indirect evidence of region with high h@xmath6 fluxes indicates that these stars are not unique . to investigate the influence of the photospheric euv flux levels of b stars on the wind ionization
, we have calculated non - lte ionization equilibrium models for the wind of using both the observed spectral energy distribution , with its high euv flux , and a spectrum consistent with non - lte models , with its much lower euv flux .
these calculations are identical to the ones described in macfarlane , cohen , & wang ( 1994 ) , where the authors used a version of the nltert code ( macfarlane 1997 ) that accounts for spherical expansion . in this code ,
rate equations for multi - level atoms are solved self - consistently with the radiation field .
we show a comparison in figure [ fig : photospheric_ionization ] of the helium and carbon ionization balance as a function of radius for the two different cases . the effect of the factor of 100 higher euv flux is not major , but is nonetheless significant . in both cases , he@xmath0 and c@xmath19 dominate , but
the higher photospheric flux changes the second most abundant ionization state to the one _ above _ the dominant stage from the one _ below _ it .
this does not change the bulk properties of the wind , but it can have a strong impact on the strengths of the observed uv resonance lines . specifically , note that the mass - loss rate determined from the 1548 , 1551 doublet will be roughly a factor of 20 higher if the ionization correction fails to account for x - ray photoionization .
the effect of the higher euv fluxes in b stars on the wind temperature has yet to be explored .
however , it is clear that the hydrogen ionization rate will be strongly increased by the higher fluxes , and photoionization is the primary heating mechanism in ob star winds ( see drew 1989 ) .
the short - wavelength euv ( @xmath20 ev @xmath21 ev ) and soft x - ray line emission from early - b stars is usually understood in the context of the wind - shock emission that is assumed to operate in o stars .
the x - ray emission from mid- and late - b stars is harder to understand in terms of the standard picture of wind shocks , however .
much of the information we have about the x - ray properties of b stars comes from large low - resolution surveys with and , but a significant amount of complementary information has been obtained from higher resolution observations of a small number of b stars , including .
the surveys have indicated that the @xmath22 relation that holds for o stars extends to about b1 but then breaks down , with @xmath23 decreasing faster than @xmath24 beyond this spectral subtype ( cohen , cassinelli , & macfarlane 1997 ) .
the breakdown in the law occurs just at the point where hot star winds are becoming optically thin to x - rays , as discussed by owocki & cohen ( 1999 ) , who show that the x - ray luminosity is expected to scale as @xmath25 .
however , the slope in the relationship is not observed to be this steep , implying that the x - ray filling factors in mid - b stars approach , and in some cases exceed , unity ( cohen et al .
there are several possible implications to these surprising data : ( 1 ) the x - ray emitting gas is clumped to enhance the emission measure for a given total mass ; ( 2 ) the mass - loss rates of these b stars are actually higher than abbott s ( 1982 ) cak parameters imply ; or ( 3 ) the x - rays do not arise ( entirely ) in the wind .
note that this filling - factor problem is not as significant an issue for very early b stars and early - b stars lying above the main sequence , like , which exhibits an emission measure filling factor of roughly 0.1 ( cohen et al . 1996 ) .
the spectrometers have access not only to photospheric radiation in the so - called `` long - wavelength '' spectrometer ( @xmath26 @xmath27 ) , but also to line emission in the short- and medium - wavelength spectrometers ( @xmath28
@xmath29 ) . for
, five highly ionized iron lines are detected in the data with greater than @xmath30 significance .
these lines , along with emission lines formed in the photoionized cool component of the stellar wind and the photospheric continuum , are shown in figure [ fig : euve_data ] .
these iron lines have a combined _ emergent _ luminosity on par with the total x - ray luminosity observed with , and significantly more _ intrinsic _ luminosity ( the difference being the result of continuum opacity in the wind ) .
they can be fit simultaneously with the data to derive a temperature distribution of the hot plasma on .
the results of this modeling indicate that there is significantly more plasma with temperatures at or below @xmath31 k than plasma with temperatures near @xmath32 k ( cohen et al . 1996 ) .
this result is consistent with the data for other b stars ( cohen et al . 1997 ) .
the analysis of the combined and data also indicates that the wind is somewhat optically thick at very soft x - ray energies , and quite optically thick at euv energies .
the amount of attenuation derived from these data indicates that the hot plasma on is spatially distributed throughout the wind ( cohen et al .
1996 ) . the information about the temperature distribution of the x - ray emitting plasma on , as well as the wind absorption , allow us to generate and constrain models of the wind ionization state of the bulk , unshocked wind . in much
the same way we compared ionization state models for different values of the photospheric euv flux , let us now examine the effect of x - rays on the wind ionization .
using an input photospheric spectrum for which we have adjusted the euv continuum flux to match the data , we calculated a model with an x - ray source distributed throughout the wind and compared it to the model with no x - rays . in figure
[ fig : xrays_ionization ] we show the results of this comparison for helium and carbon .
it can immediately be seen that the x - rays influence is much greater than the photospheric euv radiation s . with the x - ray photoionization in the calculation , the primary ionization state changes from he@xmath0 to he@xmath19 and from c@xmath19 to c@xmath33 .
there is also a general spreading - out of the ionization stages , with three different stages of carbon each accounting for at least 10% of the total .
an explanation for the strong effect of x - rays in determining the ionization states of b star winds ( as opposed to o stars winds in which the photosphere sets the ionization rate , and the x - rays cause only a perturbation ) has been offered by macfarlane et al .
( 1994 ) . in the hotter o stars ,
the photospheric radiation field is strong above the edge at 228 ( 54.4 ev ) and the x - ray field dominates the photosphere only above about 100 ev .
by contrast , for the cooler b stars , the x - ray field dominates the much weaker continuum and even some of the continuum , starting at about 40 ev .
this causes the x - rays to have a stronger effect on the wind ionization in b stars .
in fact , the x - rays boost the mean ionization state about one full stage on average for b stars .
indeed , in surprisingly high levels of wind ionization are seen in oxygen . both and detected wind broadening in the 630 line .
see figure [ fig : orfeus ] for the higher - resolution spectral measurements .
_ both _ x - rays and a high photospheric euv flux are necessary to produce significant amounts of o@xmath33 in the wind of this star .
it should be noted that the attenuation of x - rays in the bulk stellar wind involves primarily very soft x - rays and short - wavelength euv emission , and that the majority of photoionization involves helium ( from singly to fully ionized ) .
the ionization of helium by the x - rays leads to recombination radiation , primarily in the ly@xmath6 line at 304 .
this line is coincident with an line , which is pumped , leading to a radiative cascade that ( usually ) culminates in the emission of a photon at 374 . until it was seen in
, this line had never been detected in an astrophysical object outside the solar system , although the near - uv lines emitted by transitions among the high - lying levels involved in the bowen mechanism are often seen in nebulae . in figure
[ fig : grotrian ] we show an energy level diagram for the bowen mechanism in .
cohen et al .
( 1999 ) have shown that the 304 and 374 fluxes are quite sensitive to the mass - loss rate .
they find a value of @xmath12 to @xmath13 , which is about a factor of three less than the prediction from the cak parameters compiled by abbott ( 1982 ) .
the 304 line is wind broadened in the spectrum of .
the broadening is consistent with a line that is optically thick out to velocities of @xmath34 km s@xmath8 .
this is further indication that the x - ray emission is distributed throughout the wind , and furthermore , that it interacts with the bulk wind and causes ionization throughout the wind . finally , it has recently been realized that the bowen fluorescence can resolve via one of two transitions on the singlet side of at 394 and 435 , in addition to the 374 triplet ( cohen et al . 1999 ) .
the relative strength of the singlet and triplet lines is sensitive ultimately to the wind temperature , and so these lines may prove to be the most quantitative and direct temperature diagnostic of any hot star wind .
an entirely new era of astrophysical x - ray spectroscopy is imminent with the recent launches of and .
the spectral diagnostics that will soon become available will allow for the determination of the density , velocity , and temperature of the hot regions in the winds of b stars , and other hot stars .
we will be able to answer questions such as : are the shocked regions denser than the ambient wind ? and what is the ionization distribution in the denser structures in b star winds ? in order to demonstrate the types of analysis that will be possible with these new data and also to explore directly the clumping properties of b star winds we briefly discuss some new interpretations of recent medium - resolution spectra of the x - ray binary , vela x-1 , which contains a b0.5 i star as the mass - losing companion .
the accretion of the b star s wind onto the compact companion produces hard x - rays , which are reprocessed in the wind .
when viewed during eclipse , virtually all of the radiation is coming from the b star s wind , primarily via recombination in the highly ( photo-)ionized portion of the wind and via x - ray fluorescence in the less ionized , denser portion of the wind .
recently , sako et al . ( 1999 ) have identified these two different spectral components in the data ( see figure [ fig : velax1 ] ) .
much more quantitative analysis will be possible with the higher resolution spectra that become available over the next few years . however , sako et al .
( 1999 ) extract a significant amount of information from this marginal spectrum .
they fit a mass - loss rate and velocity @xmath3 parameter for the b star s wind , based on observations of the highly ionized recombination lines which are formed in the hot , low density component .
they also derive the overall mass in the cool dense component and put some constraints on its density from the observations of the x - ray fluorescence lines .
the results of the analysis show the highly ionized low density component accounts for less than 10% of the mass , but occupies 95% of the volume .
conversely , the cold , dense clumps occupy less than 5% of the volume but represent about 90% of the mass of the wind .
it is notable that this type of wind structure is similar to that seen in numerical simulations of the line - force instability ( owocki , castor , & rybicki 1988 ; feldmeier 1995 ; feldmeier , puls , & pauldrach 1997 ; owocki , runacres , & cohen , these proc . ) .
furthermore , to maintain clumping structure far out in the wind flow , over - dense and under - dense regions must be close to pressure equilibrium .
the `` ambient '' wind ( the mass - weighted mean wind , which is dominated by the dense clumps ) has a temperature somewhat below the stellar effective temperature ( @xmath35 k ) and the x - ray emitting component has a typical temperature of @xmath36 k. pressure equilibrium therefore requires the hot component to be of order 100 times less dense than the cool component .
using the numbers from the observations of vela x-1 , the clumps are 20 times denser than the mean wind , while the hot , low - density component is 10 times less dense .
so the density contrast between the two components is 200 .
owocki & cohen ( 1999 ) have introduced a simple empirical model for x - ray transfer through a two - component wind in an effort to explain the @xmath37 law . the total x - ray flux generated scales as the observed emission measure
, @xmath38 where @xmath39 is a function of atomic parameters and of the ionization distribution in the emitting plasma , and the emission measure is given by @xmath40 here the volume integral is over regions with plasma at x - ray emitting temperatures ( @xmath41 ) .
owocki & cohen ( 1999 ) scale this to the `` whole wind '' emission measure , as @xmath42 where the mass filling factor , @xmath43 , is the fraction of the mass of the wind ( above some appropriate minimum radius ) that is heated to x - ray emitting temperatures , and @xmath44 is the typical density in the post - shock region compared to the local value of the density in the idealized smooth , mean wind .
note that each of these factors can be thought of as scaling one of the density terms inside the volume integral .
the product @xmath45 is the normalization factor essentially the emission measure filling factor and the terms can not be separated using low - resolution data ; but using spectroscopic techniques , enough information can be determined about both components of the wind to separate them . for the b star in vela x-1 , @xmath46 , and @xmath47 .
this normalization factor @xmath48 is consistent with the observations of o stars , and the b0.5 i component of vela x-1 is a supergiant and so has a denser wind that is more like an o star s .
thus , a consistent picture emerges of hot star winds , in which a pervasive hot , low - density component is in approximate pressure equilibrium with a cold , denser component presumably consisting of numerous small clumps .
the numbers are roughly consistent with numerical simulations of the line - force instability , perhaps having stochastic perturbations at the base ( feldmeier et al .
1997 ) for o and early - b stars .
they are also consistent with the overall levels of x - ray flux seen in o and early - b stars , as well as with the detailed x - ray reprocessing spectrum of vela x-1 , as observed with ( sako et al .
1999 ) . for the mid- and late - b stars , however , recall that much higher filling factors ( @xmath45 ) are required to match the observations . in these objects ,
the hot x - ray emitting gas must also be dense ( _ i.e. _ @xmath49 ) to prevent @xmath50 .
it is possible , for example , that magnetic loops confine a radiation - driven wind in some of these stars , forcing different flows to collide at high velocities , and containing the resulting hot gas , which preserves its relatively high density ( babel & montmerle 1997 ) .
high - resolution x - ray spectroscopy will be able to provide density and temperature diagnostics of the emitting plasma in these stars , which will reveal whether the hot material in these winds exists in the dense phase .
the x - ray and euv properties of b stars are significant , and have a controlling effect on the ionization conditions in the wind . with the improved understanding of these properties and their effects , additional modeling of the temperature structure in ob star winds
is warranted .
the stronger than expected photospheric euv continua provide a small but important increase to the overall ionization conditions in the winds of b stars , but the x - ray and especially the short - wavelength euv emission lines play the dominant role .
the ionization balance is changed completely by x - ray photoionization .
furthermore , the x - rays ( @xmath51 ev ) that are seen by instruments like represent just a small fraction of the photons emitted by the hot plasma in the winds of b stars , which are predominantly euv photons .
finally , the interaction of this euv / x - ray emission with the cooler , denser wind component involves a complex reprocessing of the absorbed x - ray photons into uv and euv emission lines via the bowen fluorescence mechanism .
more and more , the detailed spectral information that is becoming available will drive our understanding of the structure of hot star winds , including those of b stars .
the picture that has started to emerge is one of a clumped , two - component medium , in which most of the bulk properties of the wind are defined by the dense clumps in a sense , these clumps _ are _ the wind but most of the volume is filled by a much lower density component .
the information we have covered leads us to identify the following questions which still need to be addressed : what role does clumping play in b star winds , and how do clumps originate and evolve ? how does x - ray induced ionization feed back into the line force , wind dynamics , and mass - loss rate ?
how are thermal conditions affected by the high photospheric euv fluxes ? and what is the nature of winds for stars later than about b2 ?
abbott , d. c. 1982 , , 259 , 282 babel , j. , & montmerle , t. 1997 , , 323 , 121 bertin , p. , lamers , h. j. g. l. m. , vidal - madjar , a. , ferlet , r. , & lallement , r. 1995 , , 302 , 899 cassinelli , j. p. , et al .
1995 , , 438 , 932 cassinelli , j. p. , et al .
1996 , , 460 , 949 castor , j. i. , abbott , d. c. , & klein , r. i. 1975 , , 195 , 157 cohen , d. h. , cassinelli , j. p. , & macfarlane , j. j. 1997 , , 487 , 867 cohen , d. h. , cooper , r. g. , macfarlane , j. j. , owocki , s. p. , cassinelli , j. p. , & wang , p. 1996
, , 460 , 506 cohen , d. h. , macfarlane , j. j. , cassinelli , j. p. , & owocki , s. p. 1999
, presented at the _
. division of plasma physics meeting _ drew , j. e. 1989 , , 71 , 267 feldmeier , a. 1995 , , 299 , 523 feldmeier , a. , puls , j. , & pauldrach , a. w. a. 1997 , , 322 , 878 gayley , k. g. 1995 , , 455 , 410 grigsby , j. a. , & morrison , n. d. 1995 , , 442 , 794 hubeny , i. & lanz , t. , 1995 , , 439 , 875 kutyrev , a. s. , reynolds , r. j. , & tufte , s. l. 1993 , , 182 , 6003 macfarlane , j. j. , cohen , d. h. & wang , p. 1994
, , 437 , 351 macfarlane , j. j. 1997 , _ university of wisconsin fusion design memo _ , 1050 owocki , s. p. , castor , j. i. , & rybicki , g. b. 1988 , , 335 , 914 owocki , s. p. , & cohen , d. h. 1999 , , 520 , 833 owocki , s. p. , runacres , m. , & cohen , d. h. 2000 , in `` thermal and ionization aspects of flows from hot stars : observations and theory '' , asp conf . ser .
204 , eds .
h. j. g. l. m. lamers & a. sapar , 183 puls , j. 2000 , , in press reynolds , r. j. 1995 , private communication sako , m. , leidahl , d. a. , kahn , s. m. , & paerels , f. 1999 , , 525 , 921 vallerga , j. v. , vedder , p. w. , & welsh , b. y. 1993 , , 414 , l65
|
we discuss the euv and x - ray properties of b stars , focusing on ( b2 ii ) which is the _ only _ star with both emission lines and a photospheric continuum detected with .
we explore the modest effects of the photospheric euv continua on the wind , as well as the much stronger effects of the short - wavelength euv and soft x - ray emission lines .
attenuation of the euv and soft x - ray emission by the wind plays an important role , and leads to the reprocessing of x - rays via he@xmath0 ionization and the bowen mechanism in the wind .
finally , we explore some of the new diagnostics that will shortly become available with the next generation of high spectral resolution x - ray telescopes .
all of this analysis is presented in the context of a two component stellar wind a dense component ( clumps ) that contains most of the mass but fills a negligible fraction of the volume , and a rarefied component that fills most of the volume but accounts for only a small fraction of the mass .
# 1#2#1@xmath1#2
|
introduction \1 .
rappels et notations \2 .
les anneaux @xmath10 et @xmath11 \3 .
application aux reprsentations @xmath0-adiques \4 . proprits de @xmath12 \5 .
structures diffrentielles sur les @xmath3-modules et monodromie @xmath0-adique \6 .
extensions de reprsentations semi - stables diagramme des anneaux de priodes index des notations bibliographie
dans tout cet article @xmath13 est un corps parfait de caractristique @xmath0 , et @xmath14 est le corps des fractions de lanneau des vecteurs de witt coefficients dans @xmath13 .
soit @xmath15 une extension finie totalement ramifie de @xmath14 , ce qui fait que le corps rsiduel de @xmath16 sidentifie @xmath13 et que le corps @xmath15 est complet pour la valuation @xmath17 qui tend celle de @xmath14 .
soit @xmath18 le complt dune clture algbrique @xmath19 de @xmath15 . on sintressera
aux reprsentations @xmath0-adiques @xmath1 du groupe de galois absolu @xmath20 cest - - dire aux @xmath21-espaces vectoriels de dimension finie @xmath22 munis dune action linaire et continue de @xmath23 .
fontaine a construit dans @xcite une quivalence de catgories @xmath24 entre la catgorie de toutes les reprsentations de @xmath23 et la catgorie des @xmath25-modules tales ( le foncteur inverse est not @xmath26 ) .
un @xmath25-module est un espace vectoriel de dimension finie sur un corps local @xmath27 de dimension @xmath28 muni dune action semi - linaire dun frobenius @xmath29 et dune action semi - linaire de @xmath30 commutant celle de @xmath29 . un tel module est tale si @xmath29 est de pente @xmath31 ( `` unit root '' ) .
le corps @xmath27 est isomorphe ( non canoniquement ) lanneau des sries @xmath32 en lindtermine @xmath33 o la suite @xmath34 est une suite borne dlments de @xmath14 et @xmath35 .
laction de @xmath29 et de @xmath30 est assez complique en gnral mais si @xmath15 est non - ramifi ( @xmath36 ) , alors on peut choisir @xmath33 de telle sorte que @xmath37 et @xmath38 ( @xmath39 est le caractre cyclotomique ; dans le corps du texte , @xmath33 est gal un lment @xmath40 construit via la thorie du corps des normes et laction de @xmath29 et de @xmath30 provient dune action naturelle ; si @xmath36 , on a @xmath41 - 1 $ ] ) .
lanneau @xmath27 est assez dsagrable mais il contient lanneau @xmath42 des sries surconvergentes , cest - - dire lanneau des sries @xmath43 o la suite @xmath34 est une suite borne dlments de @xmath14 et il existe @xmath44 tel que la srie @xmath45 converge sur une couronne non - vide du type @xmath46 .
le thorme principal de @xcite est que tout @xmath25-module tale provient par extension des scalaires dun `` module surconvergent '' ; plus prcisment on a le rsultat suivant si @xmath47 est un @xmath25-module tale , alors lensemble des sous-@xmath42-modules libres de type fini stables par @xmath29 et @xmath30 admet un plus grand lment @xmath48 et on a @xmath49 .
le fait que lensemble des sous-@xmath42-modules libres de type fini stables par @xmath29 et @xmath30 admet un plus grand lment est un rsultat de cherbonnier @xcite , et moyennant ce rsultat , le thorme est nonc dans @xcite de manire duale ( via le foncteur @xmath26 ) sous la forme `` toute reprsentation de @xmath23 est surconvergente '' .
grce ce rsultat , on va pouvoir tensoriser au - dessus de @xmath42 avec lanneau de robba @xmath50 ( apparaissant dans la thorie des quations diffrentielles @xmath0-adiques ) , constitu des sries @xmath43 telles que @xmath51 et quil existe @xmath44 tel que la srie @xmath45 converge sur une couronne non - vide du type @xmath46 ( sans condition de croissance au voisinage de @xmath52 ) . ceci va nous permettre , si @xmath1 est une reprsentation @xmath0-adique de @xmath23 , dune part de retrouver les invariants @xmath53 et @xmath54 associs @xmath1 par la thorie de hodge @xmath0-adique et , dautre part , en utilisant laction infinitsimale de @xmath30 , dassocier @xmath1 un module connexion @xmath55 sur lanneau de robba @xmath50 .
cet article rassemble un certain nombre de rsultats que lon peut obtenir via ltude de ce module connexion : caractrisation des reprsentations absolument cristallines , semi - stables , de de rham et @xmath18-admissibles et dmonstration de la conjecture de `` monodromie @xmath0-adique '' .
soient @xmath56 $ ] ( anneau muni des actions videntes de @xmath29 et de @xmath30 ) , @xmath57 si @xmath1 est une reprsentation @xmath0-adique de @xmath23 , soient @xmath53 , @xmath54 et @xmath58 les modules associs @xmath1 par la thorie de hodge @xmath0-adique .
le premier rsultat est que lon peut retrouver ces modules [ introcomp ] si @xmath1 est une reprsentation @xmath0-adique de @xmath23 , alors @xmath59)^{\gamma_k } \
\text { et } \ { \mathbf{d}_{\mathrm{cris}}}(v)=({\mathbf{d}^{\dagger } _
{ \mathrm{rig}}}(v)[1/t])^{\gamma_k}.\ ] ] de plus : 1 .
si @xmath1 est une reprsentation semi - stable , alors @xmath60 = { \mathbf{d}_{\mathrm{st}}}(v ) \otimes_f { \mathbf{b}^{\dagger } _ { \mathrm{log } , k}}[1/t]\ ] ] 2 .
si @xmath1 est une reprsentation cristalline , alors @xmath61 = { \mathbf{d}_{\mathrm{cris}}}(v ) \otimes_f { \mathbf{b}^{\dagger } _ { \mathrm{rig } , k}}[1/t]\ ] ] dans
lnonc ci - dessus le @xmath62 qui apparat est un lment de @xmath50 sur lequel @xmath30 agit via @xmath63 et tel que @xmath64 ; si @xmath36 , on a @xmath65 .
les modules @xmath53 et @xmath54 sont naturellement munis dun frobenius @xmath29 , dun oprateur de monodromie @xmath66 et dune filtration .
le thorme prcdent permet de retrouver les actions de @xmath29 et @xmath66 , mais la recette permettant de retrouver la filtration est suffisament peu ragotante pour ne pas tre explicite dans cet article .
un corollaire du thorme [ introcomp ] est le rsultat suivant : [ introhf ] si @xmath1 est une reprsentation cristalline de @xmath67 , alors @xmath1 est de hauteur finie .
ce rsultat avait t conjectur par fontaine ( voir @xcite ) et dmontr par colmez de manire trs dtourne ( la dmonstration utilisait deux versions @xcite de la dmonstration de la loi de rciprocit conjecture par perrin - riou @xcite ) dans le cas dun corps rsiduel fini .
la dmonstration donne dans cet article utilise la place le thorme ci - dessus et un rsultat danalyse @xmath0-adique dmontr par kedlaya @xcite .
signalons que benois @xcite a par ailleurs dmontr la loi de rciprocit de perrin - riou pour les reprsentations cristallines de hauteur finie ce qui , combin avec le rsultat ci - dessus , fournit une dmonstration de cette loi de rciprocit dans le cas gnral , nutilisant que la thorie des @xmath25-modules , ralisant ainsi un programme de fontaine .
soit @xmath68 un lment de @xmath30 assez proche de @xmath69 .
la srie qui dfinit @xmath70 converge vers un oprateur diffrentiel @xmath71 ; si @xmath36 on a @xmath72 .
si @xmath73 est lanneau des sries convergeant sur la couronne @xmath74 ( o @xmath75 $ ] et @xmath76 ) , alors @xmath71 stabilise @xmath73 pour @xmath77 assez grand et on dispose de plus de morphismes injectifs @xmath78 $ ] qui vrifient @xmath79 .
soit aussi @xmath80 ; si @xmath36 on a donc @xmath81 , cest une autre drivation de @xmath50 , qui est dailleurs une base du
@xmath50-module de rang @xmath69 des drivations de @xmath50 .
ce qui prcde correspond au cas de la reprsentation triviale et , plus gnralement , si @xmath1 est une reprsentation @xmath0-adique de @xmath23 et si @xmath68 est un lment de @xmath30 assez proche de @xmath69 , la srie qui dfinit @xmath82 converge vers un oprateur diffrentiel @xmath83 de @xmath55 au - dessus de @xmath50 muni de @xmath71 . plus
prcisment , si @xmath77 est assez grand , @xmath83 stabilise le sous - module @xmath84 ce qui permet dassocier toute reprsentation @xmath0-adique de @xmath23 un module diffrentiel sur une couronne .
une consquence presque immdiate de [ introcomp ] est le thorme suivant : [ introconn ] si @xmath1 est une reprsentation @xmath0-adique , alors il existe @xmath77 tel que la restriction de @xmath1 @xmath85 est semi - stable ( respectivement cristalline )
si et seulement si @xmath83 est une connexion unipotente ( respectivement triviale ) sur @xmath86 $ ] .
le module diffrentiel @xmath55 est points singuliers rguliers , mais a une infinit de singularits dans la couronne ; en effet la srie @xmath87 sannule en tous les @xmath88 , o @xmath89 est une racine dordre @xmath90 de @xmath69 .
on nest donc pas dans le cadre classique des quations diffrentielles @xmath0-adiques , mais si la reprsentation @xmath1 est de de rham , on peut supprimer les singularits et retomber sur un objet classique ( voir le thorme [ ndrintro ] ci - dessous ) . en localisant en @xmath88 ( o @xmath89 est un racine primitive dordre @xmath90 de lunit ) , ce qui revient considrer lapplication @xmath91 , on retombe sur le module diffrentiel considr par fontaine dans @xcite , ce qui permet en particulier de retrouver le module @xmath58 partir du noyau de cette connexion , et de montrer que la reprsentation @xmath1 est de de rham si et seulement si @xmath83 na que des singularits apparentes en les @xmath88 . il est donc naturel desprer que lon peut supprimer toutes les singularits simultanment ou , ce qui revient au mme , construire un sous - objet de rang maximal , stable par @xmath92 .
de fait , on a le rsultat suivant : [ ndrintro ] soit @xmath1 une reprsentation de de rham de @xmath23 , do nt les poids de hodge - tate sont ngatifs .
alors il existe un unique sous @xmath50-module @xmath93 de @xmath55 , libre de rang @xmath22 et stable par @xmath94 . de plus , @xmath93 est stable par @xmath29 et @xmath30 .
le @xmath50-module @xmath93 est donc muni dune connexion et dun oprateur de frobenius , et andr a rcemment montr quun tel module est quasi - unipotent .
une consquence de ceci , alli au thorme [ introconn ] , est une dmonstration de la conjecture de monodromie pour les reprsentations @xmath0-adiques @xcite : [ intromono ] si @xmath1 est une reprsentation de de rham , alors @xmath1 est potentiellement semi - stable .
enfin , en utilisant lapplication @xmath95 , on retombe sur le module de sen ( le @xmath96-espace vectoriel associ @xmath1 via la thorie de sen , voir @xcite ) , la connexion devenant un oprateur @xmath97-linaire do nt les valeurs propres sont les `` poids de hodge - tate gnraliss '' .
en particulier , @xmath1 est @xmath18-admissible ( ce qui quivaut @xmath1 de hodge - tate et tous ses poids de hodge - tate sont nuls ) si et seulement si cet oprateur est nul . on en dduit le fait que @xmath83 est divisible par @xmath62 , et donc que @xmath98 est un @xmath99-isocristal surconvergent , avec un frobenius tale .
utilisant ce fait et des techniques dquations diffrentielles @xmath0-adiques ( plus prcisment un thorme de
tsuzuki @xcite ) on obtient une dmonstration du rsultat suivant , d sen @xcite , qui caractrise les reprsentations
@xmath18-admissibles : [ introsen ] si @xmath1 est une reprsentation @xmath0-adique de @xmath23 , alors les deux conditions suivantes sont quivalentes : 1 . @xmath1 est @xmath18-admissible ; 2 . le sous - groupe dinertie de @xmath23 agit sur @xmath1 travers un quotient fini .
ce thorme est dailleurs quivalent au thorme de tsuzuki , voir la remarque [ remasentsu ] . en utilisant le thorme de monodromie @xmath0-adique , on peut montrer que : [ introordi ] 1 .
toute reprsentation ordinaire de @xmath23 est semi - stable ; 2 .
une reprsentation de de rham , qui est une extension de deux reprsentations semi - stables , est elle - mme semi - stable ; 3 .
si @xmath1 est une reprsentation semi - stable do nt les poids de hodge - tate sont tous @xmath8 , alors @xmath100 est un isomorphisme .
ces trois rsultats taient connus dans le cas dun corps rsiduel fini , o on peut les dduire de calculs de dimensions de groupes de cohomologie galoisienne ( ce qui nest plus possible si le corps rsiduel nest pas fini ) .
le ( 1 ) avait t dmontr dans ce cas l par perrin - riou @xcite comme corollaire des calculs de bloch et kato @xcite , le ( 2 ) par hyodo @xcite , et le ( 3 ) par bloch et kato .
cet article comporte six chapitres subdiviss en paragraphes .
le premier chapitre est consacr des rappels sur les reprsentations @xmath0-adiques et les anneaux de fontaine .
dans le deuxime on donne la construction des anneaux @xmath10 et @xmath11 , qui sont fondamentaux pour ce qui suit . on donne
dans le troisime chapitre une application de ces constructions : comment retrouver @xmath53 et @xmath54 partir du @xmath25-module @xmath101 . le quatrime chapitre est consacr ltude de la connexion @xmath71 sur lanneau @xmath50 , ce qui permet de dfinir dans le cinquime chapitre lquation diffrentielle associe une reprsentation @xmath0-adique ; on donne des applications la thorie de fontaine , la thorie de sen , et aux reprsentations de de rham ( preuve de la conjecture de monodromie @xmath0-adique ) .
enfin dans le sixime chapitre on donne quelques applications du thorme de monodromie @xmath0-adique .
le thorme [ introcomp ] est la runion du thorme [ isomcomp ] et de la proposition [ isomcomp2 ] .
le corollaire [ introhf ] est lobjet du thorme [ hf ] , le thorme [ introsen ] correspond la proposition [ sen ] et le thorme [ introconn ] la proposition [ conilp ] .
le thorme [ ndrintro ] et sa rciproque sont dmontrs dans le paragraphe [ ndrsec ] .
le thorme [ intromono ] correspond au thorme [ monofont ] .
le premier point de [ introordi ] est le corollaire [ ordisst ] , le ( 2 ) est le thorme [ extsst ] , et le ( 3 ) est le thorme [ expbk ] .
article comporte deux appendices pour le rendre plus lisible : un diagramme des anneaux de priodes , et un index des notations .
cet article est une version amliore de ma thse avec pierre colmez .
je tiens le remercier davoir partag avec moi ses ides et ses connaissances , ainsi que pour le temps et lnergie quil ma consacrs . lors de la rdaction de cet article , jai eu des conversations enrichissantes avec pierre berthelot , bruno chiarellotto , gilles christol , richard crew , jean - marc fontaine et kiran kedlaya .
je remercie le referee pour ses nombreuses suggestions , qui mont permis de grandement amliorer cet article .
je tiens aussi remercier les organisateurs du `` dwork trimester '' padova pour leur hospitalit .
mon sjour padova a t une occasion unique de discuter du contenu de cet article .
ce chapitre est entirement constitu de rappels sur la thorie des reprsentations @xmath0-adiques . on se reportera
@xcite ou aussi @xcite pour la dmonstration des faits qui sont rappels ici .
pour ce qui est de la thorie des quations diffrentielles @xmath0-adiques et des isocristaux surconvergents , le lecteur pourra se reporter @xcite .
la principale stratgie ( due fontaine , voir par exemple @xcite ) pour tudier les reprsentations @xmath0-adiques dun groupe @xmath102 est de construire des @xmath21-algbres topologiques @xmath103 munies dune action du groupe @xmath102 et de structures supplmentaires de telle manire que si @xmath1 est une reprsentation @xmath0-adique , alors @xmath104 est un @xmath105-module qui hrite de ces structures , et que le foncteur qui @xmath1 associe @xmath106 fournisse des invariants intressants de @xmath1 . on dit quune reprsentation @xmath0-adique @xmath1 de @xmath102 est _
@xmath103-admissible _ si on a @xmath107 en tant que @xmath102-modules .
soit @xmath13 un corps parfait de caractristique @xmath0 , @xmath14 le corps des fractions de lanneau des vecteurs de witt sur @xmath13 et @xmath15 une extension finie totalement ramifie de @xmath14 .
soit @xmath109 une clture algbrique de @xmath14 et @xmath110 sa compltion @xmath0-adique . on pose @xmath20 , cest aussi le groupe des automorphismes continus @xmath15-linaires de @xmath18 .
le corps @xmath18 est un corps complet algbriquement clos de corps rsiduel @xmath111 . on pose aussi @xmath112 et @xmath96 est dfini comme tant la runion des @xmath97 .
soit @xmath113 le noyau du caractre cyclotomique @xmath114 et @xmath115 le groupe de galois de @xmath116 qui sidentifie via le caractre cyclotomique un sous groupe ouvert de @xmath117 .
soient @xmath118 et @xmath119 lensemble des @xmath120 tels que @xmath121 .
si @xmath122 et @xmath123 sont deux lments de @xmath108 , alors on dfinit leur somme
@xmath124 et leur produit @xmath125 par : @xmath126 ce qui fait de @xmath108 un corps de caractristique @xmath0 do nt on peut montrer quil est algbriquement clos .
si @xmath127 soit @xmath128 .
cest une valuation sur @xmath108 pour laquelle celui - ci est complet ; lanneau des entiers de @xmath108 est @xmath119 et lidal maximal est @xmath129 .
lanneau @xmath131 des vecteurs de witt coefficients dans @xmath119 et @xmath132 = \ { \sum_{k\gg -\infty } p^k [ x_k],\ x_k \in { \widetilde{\mathbf{e}}^+}\}\ ] ] o @xmath133 \in { \widetilde{\mathbf{a}}^+}$ ] est le relvement de teichmller de @xmath134 .
cet anneau est muni dun morphisme danneaux @xmath135 dfini de la manire suivante : @xmath136\right ) = \sum_{k{\geqslant}0 } p^k x_k^{(0)}\ ] ] soient @xmath137 avec @xmath138 et @xmath139 , @xmath140 - 1 $ ] , @xmath141 - 1 $ ] , @xmath142 et @xmath143 .
alors @xmath144 est lidal principal engendr par @xmath145
. de mme soit @xmath146 avec @xmath147 , alors @xmath144 est aussi engendr par @xmath148-p$ ] .
remarquons que @xmath149 est un lment de @xmath119 tel que @xmath150 . on pose @xmath151 et on dfinit @xmath152 comme tant la clture sparable de @xmath153 dans @xmath108 ainsi que @xmath154 et @xmath155 lanneau des entiers et lidal maximal de @xmath152 .
remarquons que , par dfinition , @xmath152 est sparablement clos , et que lon retrouve @xmath108 partir de @xmath152 en prenant le complt de sa clture radicielle .
@xcite pour une application de la thorie du corps de normes @xcite la construction dune application @xmath156 de la limite projective des @xmath157 relativement aux applications normes dans @xmath119 do nt la principale proprit est la suivante : [ youpla ] lapplication @xmath158 induit une bijection de @xmath159 sur lanneau des entiers @xmath160 de @xmath161 .
on remarquera que la proposition ci - dessus est nonce dans ( * ? ? ? * i.1.1 ) avec la restriction supplmentaire que @xmath162 est finie mais ceci nest pas ncessaire ( en revanche il est important que @xmath163 soit finie ) .
de plus on peut montrer de la mme manire que @xmath164 est une extension finie sparable de @xmath153 de degr @xmath75 $ ] et de groupe de galois @xmath165 ( si @xmath163 est galoisienne ) , et que le groupe de galois @xmath166 sidentifie @xmath113 .
enfin @xmath167 est un anneau de valuation discrte de la forme @xmath168 $ ] o @xmath169 est limage dune suite @xmath170 duniformisantes compatibles pour les normes des @xmath157 .
lanneau @xmath172 est dfini comme tant le complt pour la topologie @xmath144-adique de @xmath173 ( on remarquera que @xmath130 est complet pour cette topologie ) : @xmath174 cest un anneau de valuation discrte , didal maximal @xmath144 ; la srie qui dfinit @xmath175)$ ] converge dans @xmath172 vers un lment @xmath62 , qui est un gnrateur de lidal maximal , ce qui fait que @xmath176 $ ] est un corps , muni dune action de @xmath67 et dune filtration dfinie par @xmath177 pour @xmath178 . on dit quune reprsentation @xmath1
de @xmath23 est de de rham si elle est @xmath171-admissible ce qui quivaut ce que le @xmath15-espace vectoriel @xmath179 est de dimension @xmath180 .
lanneau @xmath181 est dfini comme tant @xmath182 et @xmath183 $ ] . on peut
dailleurs remplacer @xmath145 par nimporte quel gnrateur de @xmath144 , par exemple @xmath148-p$ ] .
cet anneau se plonge canoniquement dans @xmath171 ( les sries dfinissant ses lments convergent dans @xmath171 ) et en particulier il est muni de laction de galois et de la filtration induites par celles de @xmath171 , ainsi que dun frobenius @xmath29 , qui tend lapplication @xmath184 dduite de @xmath185 dans @xmath119 .
on remarquera que @xmath29 ne se prolonge pas par continuit @xmath171 .
on pose @xmath186 ( on remarquera que lanneau @xmath187 est lanneau @xmath188 de @xcite ) .
la notation sexplique par le fait que lon a @xcite un isomorphisme : @xmath189 . on dit quune
reprsentation @xmath1 de @xmath23 est cristalline si elle est @xmath190-admissible ou , ce qui revient au mme , @xmath191$]-admissible ( les priodes des reprsentations cristallines vivent dans des sous @xmath14-espaces vectoriels de dimension finie , stables par @xmath29 , de @xmath190 , et donc en fait dans lanneau
@xmath192 $ ] ) ; ceci quivaut ce que le @xmath14-espace vectoriel @xmath193 \otimes_{{\mathbf{q}_p } } v)^{g_k}\ ] ] est de dimension @xmath180 .
alors @xmath53 est muni dun frobenius et dune filtration induits par ceux de @xmath190 , et @xmath194 ce qui fait quune reprsentation cristalline est aussi de de rham .
la srie qui dfinit @xmath195/\overline{\pi}^{(0)})$ ] ( aprs avoir choisi @xmath196 et o @xmath197 ) converge dans @xmath172 vers un lment @xmath198}$ ] qui est transcendant sur @xmath199 et on pose @xmath200}]$ ] et @xmath201}]$ ] . on dit quune
reprsentation @xmath1 est semi - stable si elle est @xmath202-admissible ou , ce qui revient au mme , @xmath203$]-admissible ( mme raison que ci - dessus ) ; ceci quivaut ce que le @xmath14-espace vectoriel @xmath204 \otimes_{{\mathbf{q}_p } } v)^{g_k}\ ] ] est de dimension @xmath180 .
la dfinition de @xmath202 donne ci - dessus est lgrement diffrente de celle de fontaine , mais le foncteur @xmath205 est le - mme .
alors @xmath54 est muni dun frobenius , dune filtration et dun oprateur de monodromie @xmath206}$ ] qui vrifie @xmath207 ( voir @xcite pour une justification du signe `` @xmath208 '' ) , et @xmath209
. de plus @xmath1 est cristalline si et seulement si elle est semi - stable et @xmath210 sur @xmath54 . on utilisera aussi
le @xmath14-espace vectoriel @xmath211 .
soit @xmath213 lanneau des vecteurs de witt coefficients dans @xmath108 et @xmath214 $ ] .
soit @xmath215 le complt de @xmath216 $ ] dans @xmath130 pour la topologie de celui - ci , cest aussi le complt @xmath0-adique de @xmath217[\pi^{-1}]$ ] .
cest un anneau de valuation discrte complet do nt le corps rsiduel est @xmath218 .
soit @xmath219 le complt pour la topologie @xmath0-adique de lextension maximale non ramifie de @xmath220 $ ] dans @xmath212 . on dfinit alors
@xmath221 et @xmath222 .
ces anneaux sont munis dune action de galois et dun frobenius dduits de ceux de @xmath108 . on pose @xmath223
et @xmath224 $ ] .
quand @xmath36 les deux dfinitions concident .
si @xmath1 est une reprsentation @xmath0-adique de @xmath23 soit @xmath225 .
on sait @xcite que @xmath101 est un @xmath27-espace vectoriel de dimension @xmath226 muni dun frobenius et dune action rsiduelle de @xmath30 qui commutent ( cest un @xmath25 module ) et que lon peut rcuprer @xmath1 grce la formule @xmath227 .
le corps @xmath219 est une extension totalement ramifie ( lextension rsiduelle est purement insparable ( elle est `` radicielle '' ) ) de degr @xmath0 de @xmath228 .
le frobenius @xmath229 est injectif , mais nest donc pas surjectif , et il est utile den dfinir un inverse gauche par la formule : @xmath230 tout lment @xmath231 peut scrire de manire unique sous la forme @xmath232 $ ] o les @xmath233 sont des lments de @xmath108 .
si @xmath234 , on pose : @xmath235 si @xmath236 on dfinit @xmath237 comme tant le plus petit entier @xmath77 tel que @xmath238 . la srie @xmath239 $ ] converge dans @xmath171 si et seulement
si la srie @xmath240 converge dans @xmath18 ( voir ( * ? ? ?
* ii.25 ) , et la construction qui prcde [ kertheta00 ] ) . on en dduit notamment
pour @xmath77 entier tel que @xmath241 une application injective ( on montrera cela et plus en [ kertheta00 ] , [ iotainj ] ) @xmath242 qui envoie @xmath239 $ ] sur la somme de
la srie @xmath243 $ ] dans @xmath172 .
soient @xmath244 , @xmath245 et @xmath246 .
est lensemble des @xmath248 tels quen plus , @xmath249 pour tout @xmath250 , @xmath251 , et @xmath252 o @xmath253 . on pose @xmath254 , @xmath255 , @xmath256 et @xmath257 ( en gnral , si @xmath258 est un anneau muni dune action de @xmath113 , on pose @xmath259 ) .
pour situer ces anneaux , le lecteur est invit se reporter au diagramme des anneaux de priodes qui se trouve en appendice .
la notation adopte ici diffre un peu de celle de @xcite .
ce que nous appelons @xmath260 ( repectivement @xmath261 ) y est not @xmath262 ( respectivement @xmath263 ) . on dit quune reprsentation @xmath0-adique
@xmath1 est surconvergente si @xmath101 possde une base sur @xmath27 constitue dlments de @xmath264 .
rappelons le rsultat principal @xcite ce sujet : [ toutsur ] toute reprsentation @xmath1 de @xmath23 est surconvergente , cest - - dire quil existe @xmath265 tel que @xmath266 . on a dautre
part une description assez prcise des anneaux @xmath267 , comme le montre la proposition suivante : [ rappeldag ] il existe @xmath268 et @xmath269 do nt limage modulo @xmath0 est une uniformisante @xmath270 de @xmath164 .
si @xmath271 , alors tout lment @xmath272 peut scrire @xmath273 o @xmath51 et o la srie @xmath32 est holomorphe et borne sur la couronne @xmath274 .
soit @xmath275 et @xmath276 lensemble des sries @xmath277 avec @xmath51 une suite borne telle que tout @xmath278 , on ait @xmath279 , et @xmath280 . les rappels du chapitre prcdent montrent que lapplication @xmath281 de @xmath282 dans @xmath42 est un isomorphisme .
soit @xmath283 lensemble des sries @xmath277 avec @xmath51 et telles que tout @xmath278 , on ait @xmath279 .
alors @xmath284 est lanneau de robba coefficients dans @xmath14 do nt il a t question dans lintroduction : @xmath285.\ ] ] afin de faire le lien entre reprsentations @xmath0-adiques et quations diffrentielles , on `` dfinit '' donc un anneau @xmath286 tel que @xmath287 contienne @xmath288 : @xmath289 } , \frac{[{\overline{\pi}}^s]}{p } \right\ } \left [ \frac{1}{p } \right].\ ] ] cet anneau contient @xmath290 , et aussi @xmath187 : il va donc servir comme intermdiaire entre les @xmath25-modules et la thorie de hodge @xmath0-adique .
il y a un certain nombre de choses techniques dmontrer pour vrifier que sa dfinition a bien un sens ( les trois premiers paragraphe de ce chapitre ) , et le lecteur est invit ne les consulter quen cas de besoin .
dans tout ce chapitre , @xmath292 et @xmath293 sont deux lments de @xmath294 \cup \{+\infty\}$ ] tels que @xmath295 .
rappelons que pour @xmath296 on a pos @xmath76 .
dans toute la suite , on notera @xmath133^r$ ] pour @xmath297 $ ] , mme si @xmath292 nest pas entier .
on convient que @xmath298^{+\infty}=1/[{\overline{\pi}}]$ ] et que @xmath299^{+\infty}/p=0 $ ] .
soient @xmath300 } & = { \widetilde{\mathbf{a}}^+}\ { \frac{p}{[{\overline{\pi}}]^r } , \frac{[{\overline{\pi}}]^s}{p } \ } \\ : & = { \widetilde{\mathbf{a}}^+}\{x , y\}/([{\overline{\pi}}]^r x - p , py-[{\overline{\pi}}]^s , xy-[{\overline{\pi}}]^{s - r } ) \\
\text{et}\ \ { \widetilde{\mathbf{b}}}_{[r;s ] } & = { \widetilde{\mathbf{a}}}_{[r;s]}[1/p]\end{aligned}\ ] ] o , si @xmath301 est un anneau complet pour la topologie @xmath0-adique , @xmath302 dnote la compltion @xmath0-adique de @xmath303 $ ] cest - dire que @xmath304 o @xmath305 est une suite qui tend vers @xmath31 selon le filtre des complmentaires des parties finies .
soit @xmath306 et @xmath307^r x - p , py-[{\overline{\pi}}]^s , xy-[{\overline{\pi}}]^{s - r})$ ] .
alors 1 .
@xmath308 ; 2 .
@xmath309 est ferm dans @xmath301 pour la topologie @xmath0-adique .
le ( 2 ) suit du ( 1 ) puisque @xmath309 est complet pour la topologie @xmath0-adique et que le ( 1 ) montre quune suite dlments de @xmath309 qui tend vers @xmath31 dans @xmath301 tend vers @xmath31 dans @xmath309 . montrons donc le ( 1 ) .
soit @xmath310 , @xmath311 ce qui revient dire que @xmath312 .
quitte modifier @xmath313 par des lments de @xmath314^{s - r } ) \subset p^n i$ ] on peut supposer que @xmath315 o @xmath90 divise @xmath316 et @xmath317 .
si @xmath318 cest que lon peut crire @xmath319^r x - p ) \\
+ b(x , y)(py-[{\overline{\pi}}]^s)+c(x , y)(xy-[{\overline{\pi}}]^{s - r})\end{gathered}\ ] ] et les relations @xmath320^r x -p ) & = [ { \overline{\pi}}]^r(xy-[{\overline{\pi}}]^{s - r})-(py-[{\overline{\pi}}]^s ) \ \text{et } \\
x(py-[{\overline{\pi}}]^s ) & = p(xy-[{\overline{\pi}}]^{s - r})-([{\overline{\pi}}]^r x - p)[{\overline{\pi}}]^{s - r}\end{aligned}\ ] ] montrent que quitte modifier @xmath321 on peut supposer que @xmath322 et @xmath323 . on voit alors
que @xmath324 et donc finalement que @xmath325^r x - p)+b(y)(py-[{\overline{\pi}}]^s)$ ] .
montrons que @xmath326 ( la preuve pour @xmath327 est la mme ) .
posons @xmath328 . on a alors
@xmath329 et @xmath330^r c_{i-1}-pc_i$ ] ce qui fait si @xmath331 et @xmath332 , alors @xmath90 divise @xmath333^r - pc_{i+j+1}$ ] et donc aussi @xmath334^{r(n-1-j ) } p^j(c_{i+j}[{\overline{\pi}}]^r - pc_{i+j+1})=[{\overline{\pi}}]^{rn } c_i - p^n c_{i+n}$ ] et donc @xmath90 divise @xmath335 dans @xmath130 .
lanneau @xmath336}$ ] est spar pour la topologie @xmath0-adique ( il est clairement complet ) . de plus on a une application naturelle surjective @xmath336}$ ] dans le complt @xmath0-adique de @xmath337^r,[{\overline{\pi}}]^s / p]$ ]
do nt le noyau est limage de ladhrence de @xmath309 dans @xmath336}$ ] et donc nul , ce qui fait que @xmath336}$ ] sidentifie aussi au complt @xmath0-adique de @xmath337^r,[{\overline{\pi}}]^s / p]$ ] .
tout lment de @xmath336}$ ] peut scrire de la manire suivante : @xmath338^r } \right)^k a_k + \sum_{k > 0 } \left ( \frac{[{\overline{\pi}}]^s}{p } \right)^k b_k\ ] ] avec @xmath339 , @xmath340 deux suites de @xmath130 qui convergent vers @xmath31 ( cette criture est bien sr non - unique ) . cest une consquence immdiate de la dfinition . on se contentera de remarquer que @xmath341^r}\right)^k \cdot \left(\frac{[{\overline{\pi}}]^s}{p}\right)^{\ell } = \begin{cases } { [ { \overline{\pi}}]}^{k(s - r ) } \left(\frac{[{\overline{\pi}}]^s}{p}\right)^{\ell - k } \text { si $ k { \leqslant}\ell$ } \\ { [ { \overline{\pi}}]}^{\ell(s - r ) } \left(\frac{p}{[{\overline{\pi}}]^r}\right)^{k - \ell } \text { si $ k { \geqslant}\ell$ } \end{cases}\ ] ] si @xmath342 et @xmath343 sont deux lments de @xmath119 qui vrifient @xmath344 et @xmath345 , alors @xmath336}={\widetilde{\mathbf{a}}^+}\{p/[\rho],[\sigma]/p\}$ ] .
cest vident .
remarquons que si on a @xmath346 , alors il y a une inclusion ( les deux anneaux en prsence sont intgres et ont mme corps des fractions ) : @xmath347^{r_1}},\frac{[{\overline{\pi}}]^{s_1}}{p } ] \hookrightarrow { \widetilde{\mathbf{a}}^+}[\frac{p}{[{\overline{\pi}}]^{r_2}},\frac{[{\overline{\pi}}]^{s_2}}{p}]\ ] ] cette inclusion se prolonge en un morphisme @xmath348 } { \rightarrow}{\widetilde{\mathbf{a}}}_{[r_2;s_2]}$ ] qui est toujours injectif .
le morphisme du haut se factorise en @xmath348 } { \rightarrow}{\widetilde{\mathbf{a}}}_{[r_1;s_2 ] } { \rightarrow}{\widetilde{\mathbf{a}}}_{[r_2;s_2]}$ ] et on peut donc supposer @xmath349 ou @xmath350 .
supposons par exemple @xmath351 ( lautre cas se traite de la mme manire ) .
alors il suffit de montrer que le morphisme compos @xmath352 } { \rightarrow}{\widetilde{\mathbf{a}}}_{[r;s_2 ] } { \rightarrow}{\widetilde{\mathbf{a}}}_{[r;r]}$ ] est injectif ( tout ceci pour simplifier la notation ) . on va donc montrer que si @xmath353 , alors @xmath336 } { \rightarrow}{\widetilde{\mathbf{a}}}_{[r;r]}$ ] est injectif .
soient @xmath354 $ ] et @xmath355 $ ] avec @xmath356 tels que @xmath357 et @xmath358 de telle sorte que @xmath300 } & = { \widetilde{\mathbf{a}}}\{x , y\}/(\alpha x - p , py-\alpha\beta , xy-\beta ) \
\text{et } \\ { \widetilde{\mathbf{a}}}_{[r;r ] } & = { \widetilde{\mathbf{a}}}\{x , y\}/(\alpha x - p , py-\alpha , xy-1).\end{aligned}\ ] ] le fait que lapplication naturelle @xmath359 du premier anneau dans le second est injective est quivalent au fait que si @xmath360 , alors @xmath361 appartient lidal de @xmath362 engendr par @xmath363 , ce que nous allons maintenant dmontrer .
commenccons par remarquer que quitte modifier @xmath364 par des lments de lidal @xmath365 on peut supposer ( en remplaant @xmath366 par @xmath69 ) que @xmath367 supposons
que lon ait @xmath368 les relations @xmath369 montrent que lon peut supposer , quitte modifier @xmath321 , que @xmath370 et @xmath371 .
comme @xmath361 ne contient pas de terme en @xmath366 cest alors que @xmath324 . on a donc montr que @xmath372 .
posons @xmath373 .
pour terminer la dmonstration il faut montrer que @xmath317 est un multiple de @xmath374 .
un calcul facile montre que si @xmath375 , alors @xmath376 ce qui fait que @xmath377 divise @xmath378 dans @xmath130 et donc que @xmath374 divise @xmath379 .
reste choisir @xmath77 assez grand pour que @xmath374 divise @xmath380 ce qui montre alors que @xmath374 divise @xmath317 .
on utilise ces injections pour dfinir , si @xmath309 est un intervalle de @xmath381 : @xmath382 \subset i \cap { \mathbf{r } } } { \widetilde{\mathbf{b}}}_{[r;s]}$ ] .
soient @xmath383 deux intervalles ferms , ce qui fait que @xmath384 , on dfinit une valuation @xmath0-adique @xmath385 sur @xmath386 en dcidant que @xmath387 si et seulement si @xmath388 et que limage de @xmath385 est @xmath389 .
par dfinition @xmath390 muni de @xmath385 est un espace de banach @xmath0-adique .
de plus le complt de @xmath386 pour @xmath385 sidentifie a @xmath390 .
[ valanno ] comme @xmath291 est un anneau on a notamment @xmath391 .
le but de cette partie est de dgager quelques proprits de ces anneaux .
commenons par remarquer que le groupe de galois @xmath67 agit sur @xmath130 et que cette action stend lanneau @xmath337^r,[{\overline{\pi}}]^s / p]$ ] et le stabilise , ce qui fait que laction de @xmath67 stend par continuit une action par isomtries sur son complt @xmath0-adique et par suite tous les @xmath291 et @xmath390 . de mme le frobenius @xmath29 stend en un morphisme @xmath392^r},\frac{[{\overline{\pi}}]^s}{p } ] { \rightarrow}{\widetilde{\mathbf{a}}^+}[\frac{p}{[{\overline{\pi}}]^{pr}},\frac{[{\overline{\pi}}]^{ps}}{p}]\ ] ] et se prolonge donc en une application de @xmath291 dans @xmath393 pour tout @xmath309 . [ vw ] si @xmath394 $ ] , alors @xmath395 et si @xmath396 scrit @xmath397 $ ] , alors la valuation @xmath398 vrifie @xmath399 o @xmath400 est le plus grand entier @xmath401 .
le premier point suit de la dfinition et de plus si @xmath396 vrifie @xmath402 , alors la somme qui le dfinit converge dans @xmath291 ce qui fait que @xmath403 .
reste montrer que si @xmath404 , alors @xmath402 .
comme @xmath291 est le complt @xmath0-adique de @xmath337^r,[{\overline{\pi}}]^s / p]$ ] il suffit de montrer que @xmath405 si @xmath406 , si @xmath407^r$ ] et si @xmath408^s / p$ ] ce qui est clair .
comme @xmath409 on en dduit que @xmath410 est une valuation @xmath0-adique do nt limage est @xmath389 et telle que @xmath402 si et seulement si @xmath404 ce qui fait que @xmath399 .
beaucoup de ces anneaux sont dj connus : 1 .
@xmath411}={\mathbf{a}_{\mathrm{max}}}^+$ ] et @xmath412}={\mathbf{b}_{\mathrm{max}}}^+$ ] ; 2 .
@xmath413 ; 3 .
@xmath414}$ ] et @xmath415}$ ] ; 4 .
@xmath416}$ ] et @xmath417}$ ] ; 5 .
@xmath418}$ ] et @xmath419}$ ] . le @xmath420 est une
consquence du @xmath421 et du fait que par dfinition , @xmath422 .
les @xmath423 et @xmath424 sont vidents , et le @xmath425 est contenu dans ( * ? ?
* remarque ii.1.3 ) .
reste montrer le @xmath421 qui suit du fait que par dfinition @xmath426/p-1\}$ ] et @xmath411}={\widetilde{\mathbf{a}}^+}\ { [ { \widetilde}{p}]/p\}$ ] ( et @xmath427 puisque @xmath428=a[x-1]$ ] ) . [ modp ] on a @xmath336}/(p ) = { \widetilde{\mathbf{e}}^+}[x,{\overline{\pi}}^{s - r}x^{-1}]/({\overline{\pi}}^s , { \overline{\pi}}^r x)$ ] .
notamment si @xmath429 , alors @xmath430}/(p ) = { \widetilde{\mathbf{e}}^+}/({\overline{\pi}}^r)[x , x^{-1}]$ ] . soit @xmath431 et @xmath432^{s - r } , p - x[{\overline{\pi}}]^r , [ { \overline{\pi}}]^s - p y)$ ] de telle sorte que @xmath336}$ ] sidentifie @xmath433 et donc que @xmath336}/(p)=(a / i)/(p)$ ] . on a une suite exacte @xmath434 et la multiplication par @xmath0 induit un diagramme : @xmath435 et comme @xmath433 est sans @xmath0-torsion , le lemme du serpent montre que @xmath436 sidentifie au quotient du @xmath437 par limage de @xmath309
dans ce dernier .
dans notre situation on a @xmath438 $ ] et limage de @xmath309 sidentifie @xmath439 do le lemme . attention au fait que dans le lemme prcdent , @xmath440 est lidal de lanneau @xmath441 $ ] engendr par @xmath442 et @xmath443 . on va maintenant dfinir des morphismes de @xmath336}$ ]
dans @xmath172 . on va montrer que si @xmath444 , alors lapplication @xmath445 ralise une injection de @xmath390 dans @xmath172 .
pour cela , soit @xmath446 $ ] avec @xmath296 .
si @xmath447 , alors on peut crire @xmath448 } \right)^k + \sum_{j { \geqslant}0 } b_j \left ( \frac{[{\widetilde}{p}]}{p } \right)^j \\ = \sum_{k { \geqslant}0 } a_k \left[\left ( \frac{p}{[{\widetilde}{p}]}-1 \right ) + 1 \right]^k + \sum_{j { \geqslant}0 } b_j \left [ \left ( \frac{[{\widetilde}{p}]}{p}-1 \right ) + 1 \right ] ^j \\ = \sum_{\ell { \geqslant}0 } \left ( \frac{p}{[{\widetilde}{p } ] } - 1 \right)^{\ell } \sum_{k { \geqslant}\ell } \binom{k}{\ell } a_k + \sum_{m { \geqslant}0 } \left ( \frac{[{\widetilde}{p}]}{p } - 1 \right)^m \sum_{j { \geqslant}m } \binom{j}{m } b_j \end{gathered}\ ] ] et comme les @xmath449 et @xmath450 tendent vers @xmath31 , les sries @xmath451 convergent dans @xmath130 et la srie du haut est donc convergente pour la topologie @xmath144-adique et converge dans @xmath172 vers un lment @xmath452 .
[ kertheta00 ] lapplication @xmath453 est un morphisme injectif de @xmath454 dans @xmath172 et si @xmath455 , alors le noyau du morphisme compos @xmath456 est @xmath457/p-1){\widetilde{\mathbf{a}}}_i$ ] .
montrons tout dabord que @xmath458/p-1){\widetilde{\mathbf{a}}}_i$ ] .
soit donc @xmath459 $ ] avec @xmath460 et @xmath404 tel que @xmath461 .
on peut crire @xmath448^r } \right)^k + \sum_{j { \geqslant}0 } b_j \left ( \frac{[{\widetilde}{p}]^s}{p } \right)^j \\ = \sum_{\ell { \geqslant}0 } \left ( \frac{p}{[{\widetilde}{p}]^r } - [ { \widetilde}{p}^{1-r } ] \right)^{\ell } \sum_{k { \geqslant}\ell } \binom{k}{\ell } [ { \widetilde}{p}^{1-r}]^{k-\ell } a_k \\ + \sum_{m { \geqslant}0 } \left ( \frac{[{\widetilde}{p}]^s}{p } - [ { \widetilde}{p}^{s-1 } ] \right)^m \sum_{j { \geqslant}m } \binom{j}{m } [ { \widetilde}{p}^{s-1}]^{j - m } b_j \\ = \sum_{\ell { \geqslant}1 } \left ( \frac{p}{[{\widetilde}{p}]^r } - [ { \widetilde}{p}^{1-r } ] \right)^{\ell } \sum_{k { \geqslant}\ell } \binom{k}{\ell } [ { \widetilde}{p}^{1-r}]^{k-\ell } a_k
\\ + \sum_{m { \geqslant}1 } \left ( \frac{[{\widetilde}{p}]^s}{p } - [ { \widetilde}{p}^{s-1 } ] \right)^m \sum_{j { \geqslant}m } \binom{j}{m } [ { \widetilde}{p}^{s-1}]^{j - m } b_j \\ + \sum_{k { \geqslant}0 } ( a_k [ { \widetilde}{p}^{1-r}]^k + b_k [ { \widetilde}{p}^{s-1}]^k ) . \end{gathered}\ ] ] les deux premiers termes tant des sries convergeant dans @xmath172 ( ne pas oublier que @xmath462 et @xmath463 ) et do nt la somme est dans le noyau de @xmath464 , et le troisime terme tant un lment de @xmath130 ( mme argument pour la convergence ) qui est annul par @xmath465 et qui scrit donc @xmath466-p ) y$ ] avec @xmath467 .
montrons que @xmath468^k + b_k [ { \widetilde}{p}^{s-1}]^k)=([{\widetilde}{p}]/p-1)z\ ] ] avec @xmath469 . on a @xmath470^{\ell }
+ b_{\ell } [ { \widetilde}{p}^{s-1}]^{\ell})= \\ \sum_{\ell { \geqslant}1 } a_{\ell } \left(\left ( \frac{p}{[{\widetilde}{p}]^r } \right)^{\ell}- [ { \widetilde}{p}^{1-r}]^{\ell}\right ) + \sum_{\ell { \geqslant}1 } b_{\ell } \left(\left ( \frac{[{\widetilde}{p}]^s}{p } \right)^{\ell } -[{\widetilde}{p}^{s-1}]^{\ell } \right ) = \\ \left(\frac{[{\widetilde}{p}]-p}{p}\right ) \biggl ( -\sum_{k { \geqslant}1}\left ( \frac{p}{[{\widetilde}{p}]^r } \right)^{k } \sum_{\ell { \geqslant}k } a_{\ell } [ { \widetilde}{p}^{1-r}]^{\ell - k } \\ + [ { \widetilde}{p}^{s-1 } ] \sum_{j { \geqslant}0}\left ( \frac{[{\widetilde}{p}^s]}{p } \right)^j \sum_{\ell { \geqslant}j+1 } b_{\ell}[{\widetilde}{p}^{s-1}]^{\ell - j-1}\biggr)\end{gathered}\ ] ] comme le montre un petit calcul , ce qui montre lassertion quant au noyau de @xmath471 sur @xmath291 .
enfin montrons que @xmath472 est injectif .
si @xmath473 avec @xmath474 , alors @xmath475 et donc @xmath476 est divisible par @xmath148/p-1 $ ] .
comme @xmath472 est un morphisme danneau et que @xmath172 est intgre , cest que @xmath477/p-1 ) x_1 $ ] avec @xmath478 et @xmath479 .
en itrant ce procd on en dduit que @xmath480/p-1)^n { \widetilde{\mathbf{a}}}_{j_0}$ ] .
reste donc montrer que @xmath481/p-1)^n { \widetilde{\mathbf{a}}}_{j_0}=0 $ ] . limage de cette intersection dans @xmath482 sidentifie @xmath483 $ ] qui est nulle . on en dduit
quun lment de lintersection @xmath484/p-1)^n { \widetilde{\mathbf{a}}}_{j_0}$ ] est infiniment divisible par @xmath0 et donc nul .
[ kertheta0 ] si @xmath485 on pose @xmath486 .
lapplication @xmath487 est un morphisme injectif de @xmath488 dans @xmath172 et si @xmath444 , alors le noyau du morphisme compos @xmath489 est @xmath490/p-1){\widetilde{\mathbf{a}}}_i$ ] . cest une consquence immdiate de la proposition prcdente , tant donn que lapplication @xmath491 est une bijection , et que @xmath492/p-1)=[{\widetilde}{p}]/p-1 $ ] .
si @xmath444 , alors @xmath493 ralise une injection de @xmath291 et @xmath390 dans @xmath172 .
[ phiqn ] le noyau de @xmath494 est @xmath495 .
en effet , @xmath496-p)={\varphi}^n([{\widetilde}{p}]-p)$ ] , @xmath497 et on sait que @xmath148-p$ ] et @xmath145 engendrent le mme idal de @xmath130 .
[ unicite ] les inclusions naturelles de @xmath498 et @xmath499 dans @xmath454 induisent une suite exacte @xmath500 .
la flche est @xmath501 est surjective car il suffit de dcomposer une criture dun lment de @xmath454 en deux .
ensuite @xmath130 est contenu dans lintersection de @xmath498 et de @xmath499 et il reste donc montrer que linclusion @xmath502 est aussi une surjection . on va dabord montrer que cest vrai modulo @xmath503 ( on remarquera que modulo @xmath0 la flche nest plus injective ) .
rappelons que les anneaux @xmath498 et @xmath499 sidentifient @xmath504)$ ] et @xmath505y - p)$ ] et que @xmath506 $ ] .
limage de @xmath499 dans cet anneau sidentifie @xmath507 $ ] et celle de @xmath498 @xmath508 $ ] ce qui fait que limage de leur intersection ( qui est un sous - ensemble de lintersection de leurs images ) est un sous - anneau de @xmath509 et donc que la flche @xmath510 est surjective modulo @xmath511 .
si lon prend @xmath476 dans @xmath512 il existe donc @xmath467 tel que @xmath513 .
cela veut dire que @xmath514 dune part et @xmath515{\widetilde{\mathbf{a}}^+}$ ] dautre part ( il suffit dappliquer le lemme [ modp ] tous ces anneaux ) .
comme @xmath0 divise @xmath148 $ ] dans @xmath498 il existe @xmath516{\widetilde{\mathbf{a}}^+}$ ] tel que @xmath517 . on conclut en itrant ce procd ( comme @xmath130
est complet pour la topologie @xmath0-adique ) .
dans ce paragraphe on introduit lanneau @xmath10 .
les anneaux @xmath518 et @xmath10 sont dfinis par @xmath519 et @xmath520 .
on munit @xmath518 de la topologie de frchet dfinie par lensemble des @xmath385 o @xmath309 parcourt lensemble des intervalles ferms de @xmath521 . on dfinit
aussi @xmath522 comme tant lanneau des entiers de @xmath518 pour la valuation @xmath523}$ ] .
[ kertheta1 ] on a @xmath524 .
tant donne la remarque [ phiqn ] il suffit de montrer que @xmath525/p-1){\widetilde{\mathbf{b}}}_i$ ] pour tout @xmath526 ce qui suit de [ kertheta0 ] . on a une suite exacte @xmath527 on va dabord montrer que si @xmath528 , alors on a @xmath529 } \oplus { \widetilde{\mathbf{b}}}_{[0;r_n ] } { \rightarrow}{\widetilde{\mathbf{b}}}_{[r;r_n ] } { \rightarrow}0\ ] ] il est clair que tout lment de @xmath530}$ ] scrit comme somme dlments des deux autres et il faut vrifier que deux critures diffrentes diffrent par un lment de @xmath173 ce qui revient montrer que @xmath531 } \cap { \widetilde{\mathbf{b}}}_{[0;r_n ] } = { \widetilde{\mathbf{b}}^+}$ ] , ou encore en appliquant @xmath445 que @xmath532 } \cap { \widetilde{\mathbf{b}}}_{[0;r_0 ] } = { \widetilde{\mathbf{b}}^+}{}$ ] ce qui suit du lemme [ unicite ]
. montrons maintenant le lemme .
si @xmath533 , alors pour tout @xmath77 on peut crire ( puisque @xmath534}$ ] ) : @xmath535 avec @xmath536}$ ] et @xmath537}$ ]
. remarquons que @xmath538 est une autre criture de ce type ( puisque @xmath539}$ ] et @xmath540}$ ] ) et donc que @xmath541 ce qui fait que quitte modifier @xmath542 et @xmath543 par des lments de @xmath173 on peut supposer que @xmath544 et @xmath545 ce qui fait que @xmath546 avec @xmath547 et @xmath548 } = { \widetilde{\mathbf{b}}^{+}_{\mathrm{rig } } } $ ] .
[ complet ] lanneau @xmath518 est complet pour sa topologie de frchet et contient @xmath260 comme sous - anneau dense .
le fait que @xmath518 est complet suit du fait que chacun des @xmath549}$ ] est complet pour @xmath550}$ ] .
ensuite montrons que @xmath260 est dense .
soient @xmath533 et @xmath551 trois rels .
alors comme @xmath552}$ ] , on peut lcrire comme @xmath553^t}{p}\right)^k\ ] ] avec @xmath554 et @xmath555 , si @xmath556 . on a alors @xmath557^t}{p}\right)^n { \widetilde{\mathbf{a}}}_{[r;t ] } \subset \left(\frac{[{\overline{\pi}}]^t}{p}\right)^n { \widetilde{\mathbf{a}}}_{[r;s]}\ ] ] et un petit calcul montre qualors @xmath550}(x - x_n ) { \geqslant}n(t / s-1)$ ] .
un argument dextraction diagonale permet de trouver une suite qui converge vers @xmath476 pour la topologie de frchet
. [ princmax ] pour @xmath533 et @xmath558 \ni r$ ] , on a @xmath559}(x ) ; v_{[t;t]}(x ) \}$ ] .
un petit argument montre que cest vrai avec @xmath560 la place de @xmath385 pour @xmath561 ; on conclut par densit et continuit .
si @xmath562 est un lement de @xmath119 qui vrifie @xmath563 , alors la topologie dfinie par @xmath385 sur @xmath522 est plus fine que la topologie @xmath564$]-adique cest - - dire que si @xmath565 , alors @xmath566 pour la topologie @xmath564$]-adique .
de plus les topologies @xmath564$]-adiques et @xmath523}$]-adiques sont quivalentes .
soit @xmath567 une suite telle que @xmath568 pour @xmath569 et soit @xmath570 .
alors un petit calcul montre que @xmath571^{-m } ) { \geqslant}0 $ ] pour @xmath569 et donc @xmath523}(y_i [ a]^{-m } ) { \geqslant}0 $ ] pour @xmath569 ce qui revient dire que @xmath572^m { \widetilde{\mathbf{a}}^{\dagger , r}_{\mathrm{rig } } } $ ] pour @xmath569 et donc que @xmath573 tend vers @xmath31 pour la topologie @xmath564$]-adique .
de mme si @xmath574^m { \widetilde{\mathbf{a}}^{\dagger , r}_{\mathrm{rig } } } $ ] , alors @xmath523}(y ) { \geqslant}\frac{(p-1)m \lambda}{pr}$ ] et donc les deux topologies sont quivalentes . on prendra garde au fait que cela nest plus vrai si @xmath309 nest plus rduit @xmath575 $ ] .
lanneau @xmath247 est complet pour la valuation @xmath523}$ ]
. dans ( * ? ? ?
* ii.1.2 ) on montre que @xmath247 est spar complet pour la topologie @xmath564$]-adique si @xmath576 .
ce paragraphe est consacr la construction dun anneau @xmath11 qui est @xmath10 ce que @xmath202 est @xmath190 . on commence par construire une application logarithme .
il existe une et une seule application @xmath578 $ ] de @xmath108 dans @xmath579 $ ] qui vrifie @xmath580=\log[x]+\log[y]$ ] , @xmath581=0 $ ] si @xmath582 , @xmath198}=x$ ] et @xmath583=\sum_{n>0 } ( -1)^{n-1}\frac{([x]-1)^n}{n } \text { si } v_p(x^{(0)}-1 ) { \geqslant}1\ ] ] soit @xmath584 lensemble des @xmath585 tels que @xmath586 .
pour @xmath587 la srie @xmath581=\sum_{n>0 } ( -1)^{n-1}([x]-1)^n / n$ ] converge dans @xmath181 et @xmath580=\log[x]+\log[y]$ ] par un argument de sries formelles . on en dduit notamment que @xmath581={\varphi}(\log[x]/p)$ ]
ce qui fait que limage de @xmath584 par @xmath588 est en fait incluse dans @xmath187 .
si @xmath589 est tel que @xmath590 , alors il existe @xmath77 tel que @xmath591 et le @xmath588 stend donc lensemble des @xmath476 tels que @xmath590 .
ensuite soit @xmath592 .
on peut crire @xmath593 avec @xmath594 et @xmath595 de manire unique ce qui montre que @xmath588 stend @xmath596 .
enfin @xmath119 est un anneau de valuation et le choix de @xmath198}$ ] achve de dterminer lapplication @xmath597 $ ] .
[ log ] il existe une et une seule application @xmath598 $ ] telle que @xmath599)=\log [ x]$ ] , @xmath600 et @xmath601 .
si @xmath602 est exactement divisible par @xmath603 , alors il existe @xmath292 tel que lon peut crire @xmath604(1-pz)$ ] avec @xmath605 , o @xmath606 est limage
de @xmath607 dans @xmath119 , et la srie @xmath608 converge dans @xmath247 qui est complet pour la topologie @xmath0-adique ce qui permet dtendre @xmath588 par multiplicativit @xmath130 ( et aussi @xmath173 ) . on pose @xmath609 $ ] muni de laction de @xmath67 et du frobenius donns par @xmath610 et @xmath611)$ ]
ce qui fait que lon peut prolonger lapplication @xmath612 pour @xmath77 assez grand par @xmath613}$ ] .
la proposition suivante montre que @xmath493 est injectif et commute aux actions de @xmath67 et de frobenius ( l o ce - dernier est dfini ) , et ds lors on crira @xmath614}]$ ] .
soit aussi @xmath615 .
[ iotainj ] lapplication @xmath616 { \rightarrow}{\mathbf{b}_{\mathrm{dr}}}^+$ ] qui tend @xmath493 par @xmath617}$ ] est injective , commute laction de galois et sa restriction @xmath618 est @xmath445 .
les deux derniers points sont triviaux .
pour montrer le premier , il suffit de montrer que llment @xmath198}$ ] est transcendant sur le corps des fractions de @xmath619 et cela revient au mme de montrer que @xmath620)$ ] est transcendant sur @xmath621 .
montrons tout dabord que @xmath622 .
soit @xmath623/p$ ] et @xmath624 lanneau des lments de @xmath172 qui appartiennent @xmath625 $ ] .
rappelons que fontaine a montr que @xmath626 ( cf @xcite ) .
la dmonstration de la proposition [ kertheta00 ] montre que si @xmath627 , alors @xmath628 ( en regardant la srie qui donne limage de @xmath476 dans @xmath172 ) et donc si @xmath629 , alors on a @xmath630 tels que @xmath631 ( puisque @xmath632 ) et le rsultat de fontaine montre que cela nest pas possible et donc que @xmath622 .
montrons maintenant que @xmath633 est transcendant sur @xmath621 pour tout @xmath77 .
un petit calcul montre quil existe @xmath634 tel que @xmath635 .
soit @xmath636 le polynme minimal de @xmath633 .
alors en appliquant @xmath637 et en comparant les coefficients il vient @xmath638 ce qui fait que @xmath639 sidentifie un lment @xmath640 de @xmath171 stable par @xmath67 et donc que @xmath641 , avec @xmath642 , et on vient de voir que cela est impossible .
par la proposition [ log ] il existe un lment @xmath643 ( remarquons que lon a @xmath644 $ ] puisque la srie qui dfinit @xmath645/\pi)$ ] converge dans @xmath499 ) .
on munit @xmath11 de loprateur de monodromie @xmath66 dfini par @xmath646 cest - - dire que @xmath647 .
un calcul facile montre que @xmath66 commute laction de @xmath67 . [ logpietro ] llment @xmath648 est plus agrable que @xmath198}$ ] , par exemple @xmath649 $ ] si @xmath650 .
dans ce paragraphe on dcrit les invariants de @xmath10 sous laction de @xmath113 .
soit @xmath651 .
[ decomp ] soit @xmath309 un intervalle qui contient @xmath652 $ ] et @xmath653 $ ] . on a une suite exacte @xmath654 . on a vu que lon
a une suite exacte @xmath655 et en prenant les invariants par @xmath113 on trouve @xmath656 et pour montrer le lemme il suffit de montrer que @xmath657 .
si @xmath658 , alors comme @xmath299 $ ] est inversible dans @xmath659 ( sauf si @xmath660 , auquel cas le lemme est trivial ) , on a @xmath661^{-1 } \in { \widetilde{\mathbf{b}}}_j^{h_k}$ ] et il existe @xmath77 tel que @xmath662^{-1 } ) \in h^1(h_k , p^{-n } { \widetilde{\mathbf{a}}^+})$ ] et donc @xmath663^{-1 } ) [ { \overline{\pi } } ] ) \in h^1(h_k , w({\mathfrak{m}}_{{\widetilde{\mathbf{e}}^+}}))$ ] . or ce dernier espace de cohomologie est nul ( ( * ? ? ?
* iv.2.4 ) appliqu la reprsentation triviale ) et on en dduit que @xmath664 .
[ youpi ] dans le cas o @xmath665 , on obtient la suite exacte : @xmath666 [ bmaxhk ] si @xmath667 , alors il existe une suite @xmath316 dlments de @xmath668 , qui tend vers @xmath31 , telle que @xmath669 . on se ramne immdiatement montrer que si @xmath670 , alors il existe une suite @xmath316 dlments de @xmath671 , telle que @xmath669 . on va dabord montrer que @xmath672 est surjective .
si @xmath36 , cest bien connu ( on se ramne montrer que cest vrai modulo @xmath0 , et cela rsulte alors du fait que les @xmath673 \cap [ 0;1 [ \}$ ] forment une base de @xmath674 ( voir ( * ? ? ?
* iii.2.1 ) ) ) ; ensuite soient @xmath675 llment construit
dans la proposition [ youpla ] , @xmath676 , et @xmath677 .
rappelons @xcite que si @xmath556 et @xmath678 , alors @xmath679 .
comme @xmath680 on a , pour @xmath556 , @xmath681 et donc , si @xmath556 , alors @xmath682 est gal modulo @xmath683 une uniformisante de @xmath157 .
ceci implique que lapplication @xmath684 { \rightarrow}{\mathcal{o}}_{k_n}/\mathfrak{a}$ ] est surjective , et donc que @xmath685_{n { \geqslant}0 } { \rightarrow}{\mathcal{o}}_{{\widehat}{k}_{\infty}}$ ] est surjective ( puisquelle lest modulo @xmath683 et que les deux @xmath686-modules en prsence sont complets pour la topologie @xmath683-adique ) .
comme @xmath687)$ ] , ceci montre que @xmath672 est surjective .
soit maintenant @xmath670 .
on dfinit deux suites @xmath688 et @xmath316 de @xmath689 et @xmath671 de la manire suivante : @xmath690 , et si @xmath691 , alors comme @xmath692 , il existe @xmath693 tel que @xmath694 .
ceci montre que @xmath695 est dans le noyau de @xmath465 et est donc divisible par @xmath696 dans @xmath498 . on pose @xmath697 .
il est clair qualors , @xmath669 .
montrons que @xmath698 : si @xmath699 alors il existe une suite @xmath700 , telle que @xmath701 , et un petit calcul montre que cela force @xmath698 .
[ dagrig ] lanneau @xmath702 est dense dans @xmath703 pour la topologie de frchet de ce dernier . cela revient montrer que si lon se fixe un intervalle @xmath704 $ ] avec @xmath556 , alors pour tout @xmath705 il existe une suite @xmath706 telle que @xmath707 .
tant donne la dcomposition du corollaire [ youpi ] il suffit de le faire pour @xmath708}^{h_k}$ ] avec @xmath709 . on se
fixe @xmath710 tel que @xmath711 ( cest possible car pour tout @xmath309 compact , @xmath712 pour @xmath385 ) .
le lemme [ bmaxhk ] auquel on applique @xmath713 montre que tout lment @xmath708}^{h_k}$ ] scrit @xmath714 o @xmath715 est une suite de @xmath668 qui tend vers @xmath31 , et donc que @xmath716 o @xmath573 est une suite borne de @xmath668 .
comme @xmath711 , la suite @xmath717 tend vers @xmath31 pour @xmath385 et donc , pour montrer la proposition , il suffit de prendre @xmath718 .
soit @xmath12 le complt de @xmath267 pour la topologie de frchet . on va
donner une description nettement plus agrable des @xmath12 . on a dj vu quil existe @xmath268
et @xmath269 do nt limage modulo @xmath0 est une uniformisante de @xmath720 et que si @xmath271 , alors tout lment @xmath272 peut scrire @xmath273 o @xmath51 et o la srie @xmath32 est holomorphe et borne sur la couronne @xmath721 . lanneau @xmath267 est alors muni de la topologie induite par celle de @xmath518 qui devient la topologie de la convergence sur les couronnes dfinies par un intervalle compact ce qui fait que @xmath12 est le complt de @xmath267 pour cette topologie et donc que soit @xmath722 lensemble des sries @xmath277 avec @xmath51 et telles que tout @xmath278 , @xmath723 et soit @xmath724 .
alors lapplication @xmath725 qui @xmath313 associe @xmath726 est un isomorphisme .
la fin de ce paragraphe est consacr la dmonstration de la proposition suivante : [ rm ] si @xmath292 est un entier assez grand , alors il existe des applications @xmath727 , telles que : 1 .
@xmath728 est une section continue de linclusion @xmath729 ; 2 .
@xmath728 est @xmath730-linaire ; 3 .
si @xmath705 , alors @xmath731 ; 4 .
si @xmath732 , et @xmath733 , alors @xmath734 . soit @xmath735 ,
si @xmath736 , soit @xmath737 , pour @xmath77 assez grand tel que @xmath738 .
rappelons ( voir ( * ? ? ?
* iii.2 ) , par exemple ) que tout lment @xmath739 scrit de manire unique sous la forme : @xmath740 , o @xmath741 est une suite de @xmath742 qui tend vers @xmath31 . on en dduit
que tout lment @xmath743 scrit de manire unique sous la forme : @xmath744 a_i(x)$ ] , o @xmath745 est une suite de @xmath746 qui tend vers @xmath31 . on pose @xmath747 a_i(x)$ ] . en particulier , @xmath748 .
rappelons que si @xmath292 est un entier @xmath8 , alors @xmath749 , et que pour dfinir les @xmath728 sur @xmath750 dans ( * ? ? ?
* iii.2 ) , on tend @xmath728 @xmath750 par la formule @xmath751 .
pour construire les @xmath728 sur @xmath703 , et dmontrer la proposition [ rm ] , on va suivre un chemin semblable .
[ riginter ] soit @xmath292 un entier @xmath8 . on a @xmath752 . on va montrer que @xmath753,[{\overline{\pi}}^s]/p\}$ ] , le lemme en suit car si @xmath754 , alors @xmath755 $ ] est une unit de @xmath247 .
soit @xmath756 , et @xmath556 .
on peut crire @xmath546 , avec @xmath757 , et @xmath758})^{h_k}$ ] par le lemme [ decomp ] .
le lemme [ bmaxhk ] montre que @xmath759 peut scrire @xmath760 , o @xmath761 est une suite de @xmath762 qui tend vers @xmath31 .
on peut dailleurs remplacer @xmath145 par nimporte quel lment @xmath633 de @xmath763 , tel que les idaux @xmath764 et @xmath765 de @xmath763 soient gaux .
on peut notamment crire @xmath766^{r_n}/p)^i$ ] , o @xmath335 est une suite de @xmath762 qui tend vers @xmath31 .
ceci tant vrai pour tout @xmath767 , le lemme en rsulte .
si on tend @xmath728 @xmath768 , par @xmath769 alors , si @xmath770 , on a @xmath731 .
cela suit du fait que @xmath771 , dune part , et que @xmath772 , dautre part .
comme on a dfini les applications @xmath728 de @xmath773 dans lui - mme , et que @xmath774 ( par le lemme [ riginter ] ) , on en dduit des applications @xmath775 .
les deux premiers points de la proposition [ rm ] sont clairs sur la dfinition pour @xmath36 .
le troisime suit ( toujours si @xmath36 ) du lemme prcdent .
cela dfinit @xmath728 dans le cas o @xmath36 .
dans le cas gnral , @xmath776 est une extension de degr @xmath777 si @xmath292 est assez grand .
soient @xmath778 , @xmath779 une base de @xmath267 sur @xmath780 , et @xmath781 la base duale pour la forme linaire @xmath782 .
alors , si @xmath705 , on a @xmath783 avec @xmath784 , et on pose @xmath785 .
comme @xmath728 est @xmath780-linaire , cette dfinition ne dpend pas du choix de la base @xmath786 .
les deux premiers points de la proposition sont clairs sur la dfinition . on a dj montr le troisime , dans le cas o @xmath36 , et dans le cas gnral , si on crit @xmath787 , alors @xmath788 car @xmath789 pour chaque @xmath790 .
reste voir que @xmath728 commute @xmath733 .
cest vrai sur @xmath763 , car lcriture que lon utilise pour dfinir @xmath728 est unique .
la manire do nt on a construit @xmath728 montre quil en est de mme sur @xmath791 .
comme @xmath792 , avec @xmath793 , on a @xmath794 : @xmath728 commute donc @xmath733 . on dfinit aussi @xmath795 $ ] , cet anneau est stable par les actions de @xmath29 et de @xmath30 tant donn que @xmath796 et @xmath797 et que les sries qui dfinissent @xmath798 et @xmath799 convergent dans @xmath50 . remarquons que si @xmath650 , alors @xmath800 $ ] .
[ rmlog ] on prolonge les @xmath728 en une section @xmath730-linaire de linclusion de @xmath801)$ ] dans @xmath802 $ ] : ils commutent toujours @xmath30 , et @xmath731 si @xmath803 $ ] .
ce chapitre est consacr lapplication des constructions de la section prcdente la caractrisation des reprsentations @xmath0-adiques semi - stables ( et cristallines ) en termes du @xmath25-module qui leur est associ .
on montre en particulier que si @xmath1 est une reprsentation @xmath0-adique , alors @xmath804)^{\gamma_k } \ \text { et } \\ { \mathbf{d}_{\mathrm{cris}}}(v ) & = ( { \mathbf{b}^{\dagger } _ { \mathrm{rig } , k } } \otimes_{{\mathbf{b}^{\dagger } } _ k } { \mathbf{d}^{\dagger } } ( v)[1/t])^{\gamma_k}\end{aligned}\ ] ] on commence par tablir un rsultat de rgularisation par le frobenius , qui est la base des applications suivantes .
le frobenius @xmath29 est une bijection de @xmath390 sur @xmath805 et induit donc une bijection de @xmath518 sur @xmath806 ainsi que de @xmath577 sur @xmath807 puisque @xmath808})=p\cdot{\log[\overline{\pi}]}$ ] .
[ regul ] soit @xmath809 un entier positif .
alors @xmath810 montrons le premier point : comme @xmath811 il scrit de manire unique sous la forme @xmath812 $ ] et de mme @xmath813 $ ] .
cest que @xmath815 ce qui implique que @xmath816 et donc ( en laissant tendre @xmath293 vers @xmath817 ) que @xmath818 ce qui fait que @xmath406 .
passons au deuxime point .
pour tout @xmath293 on peut crire @xmath819 avec @xmath820 et @xmath821 .
par le lemme [ unicite ] on a @xmath822 et dautre part @xmath823 ce qui fait que @xmath824 et que quitte modifier @xmath825 on peut supposer que @xmath826 . on a alors
@xmath827 et donc @xmath828 .
[ phireg ] soit @xmath292 et @xmath633 deux entiers positifs et @xmath829 . on suppose quil existe @xmath830 telle que @xmath831 .
alors @xmath832 .
soit @xmath833 et @xmath834}^n$ ] .
soit @xmath835 tel que @xmath836 et @xmath837 .
lhypothse reliant @xmath301 et @xmath838 peut scrire : @xmath839 et comme @xmath840}^n)=p^{-n } { \log[\overline{\pi}]}^n$ ] , on en dduit que si @xmath841 , alors comme @xmath842 et @xmath843 , on a @xmath844 . on itre ce procd
et il en sort que @xmath845 . on est
en mesure dappliquer le lemme [ regul ] @xmath846 et la proposition suit .
soient @xmath847 $ ] et @xmath848 le thorme [ toutsur ] montre que @xmath55 et @xmath849 sont des @xmath50- et @xmath719- modules libres de rang @xmath180 .
si @xmath850 est un @xmath67-module soit @xmath851 le tordu de @xmath850 par @xmath852 ( twist de tate ) . on a @xmath853 soit @xmath854 .
cest un @xmath14-espace vectoriel de dimension finie ( puisque @xmath493 ralise une injection de @xmath855 dans @xmath856 ) stable par frobenius et la proposition [ phireg ] implique que @xmath857 ce qui fait @xcite que @xmath858 .
comme @xmath859 cela dmontre le rsultat .
soit @xmath860 , rappelons que @xmath211 .
si @xmath1 a ses poids de hodge - tate ngatifs , alors @xmath861 et en gneral @xmath862 pour @xmath22 assez grand .
[ comp1 ] si @xmath1 est une reprsentation @xmath0-adique , alors @xmath863 est un @xmath14-espace vectoriel de dimension finie , et le morphisme induit par linclusion de @xmath618 dans @xmath11 @xmath864 est un isomorphisme de @xmath865-modules .
si @xmath866 , alors @xmath867 est un @xmath14-espace vectoriel de dimension finie @xmath868d$ ] , car @xmath493 ralise une injection de @xmath869 dans @xmath58 , qui est un @xmath15-espace vectoriel de dimension finie @xmath870 .
si lon prend @xmath871d+1 $ ] lments de @xmath863 , ils vivent dans @xmath869 pour @xmath556 , et vrifient donc une relation de dpendance @xmath14-linaire . cest donc que @xmath863 est un @xmath14-espace vectoriel de dimension @xmath868d$ ] .
passons maintenant au deuxime point .
soient @xmath872 et @xmath873 des @xmath21- et @xmath14- bases de @xmath1 et @xmath863 .
il existe une matrice @xmath874 telle que @xmath875 ( les @xmath876 et @xmath877 sont des vecteurs colonnes ) .
soit @xmath878 la matrice de @xmath29 dans la base @xmath876 ( qui est inversible car @xmath879 est une bijection ) . on a alors
@xmath880 et donc @xmath881 ; la proposition [ phireg ] montre que @xmath882 et donc que @xmath883 , ce qui permet de conclure .
une reprsentation @xmath1 poids ngatifs est donc semi - stable si et seulement si elle est @xmath11-admissible et elle est cristalline si et seulement si elle
est @xmath10-admissible , cest - - dire si elle est @xmath11-admissible et que ses priodes sont tues par @xmath66 .
de plus , si les poids de hodge - tate de @xmath1 ne sont pas ngatifs , alors en tordant @xmath1 on en dduit que @xmath1 est semi - stable si et seulement si elle est @xmath884$]-admissible et elle est cristalline si et seulement si elle est @xmath885$]-admissible .
[ comparaison ] si @xmath1 est semi - stable on a un isomorphisme de comparaison : @xmath886 \otimes_f { \mathbf{d}_{\mathrm{st}}}(v ) = { \widetilde{\mathbf{b}}^{\dagger } _ { \mathrm{log } } } [ 1/t ] \otimes_{{\mathbf{q}_p } } v\ ] ] ceci rsulte du fait que dans ce cas on a dj : @xmath887 \otimes_f { \mathbf{d}_{\mathrm{st}}}(v ) = { \widetilde{\mathbf{b}}^{+}_{\mathrm{log } } } [ 1/t ] \otimes_{{\mathbf{q}_p } } v\ ] ] il suffit alors de tensoriser les deux membres par @xmath884 $ ] au - dessus de @xmath203 $ ] .
[ isomcomp ] si @xmath1 est une reprsentation @xmath0-adique , alors @xmath59)^{\gamma_k } \ \text { et } \ { \mathbf{d}_{\mathrm{cris}}}(v)=({\mathbf{d}^{\dagger } _ { \mathrm{rig}}}(v)[1/t])^{\gamma_k}\ ] ]
notamment @xmath1 est semi - stable ( respectivement cristalline ) si et seulement si @xmath888)^{\gamma_k}$ ] ( respectivement @xmath889)^{\gamma_k}$ ] ) est un @xmath14-espace vectoriel de dimension @xmath180 .
le deuxime point est une consquence immdiate du premier .
ensuite comme @xmath890 \subset ( { \widetilde{\mathbf{b}}^{\dagger } _ { \mathrm{log } } } [ 1/t]\otimes_{{\mathbf{q}_p } } v)$ ] ( et que @xmath86 \subset ( { \widetilde{\mathbf{b}}^{\dagger } _ { \mathrm{rig } } } [ 1/t]\otimes_{{\mathbf{q}_p } } v)$ ] ) les rsultats prcdents montrent que @xmath888)^{\gamma_k}$ ] ( respectivement @xmath889)^{\gamma_k}$ ] ) est inclus dans @xmath54 ( respectivement dans @xmath53 ) .
montrons donc que @xmath891)^{\gamma_k}$ ] , et que @xmath892)^{\gamma_k}$ ] .
il suffit de soccuper du cas semi - stable car le cas cristallin en suit en faisant @xmath210 .
soit @xmath893 .
on peut supposer ( quitte tordre ) les poids de hodge - tate de @xmath1 ngatifs puisque lon a invers @xmath62 partout . on sait qualors
@xmath894 et de plus @xmath895 puisque @xmath896 a la bonne dimension . on en dduit que si lon choisit une base @xmath779 de
@xmath896 et @xmath897 une base de @xmath54 , alors la matrice @xmath898 dfinie par @xmath899 est de rang @xmath292 et vrifie @xmath900 o @xmath901 est la matrice de @xmath902 dans la base @xmath779 .
les oprateurs @xmath903 introduits au paragraphe [ rmdef ] sont @xmath719-linaires et commutent @xmath30 ( voir la proposition [ rm ] et la dfinition [ rmlog ] ) ce qui fait que @xmath904 . de plus , @xmath905 et si @xmath906 , alors @xmath907 .
soit @xmath908 . on a alors
@xmath909 et comme les actions de @xmath29 et @xmath30 commutent sur @xmath55 on a @xmath910 ( @xmath838 est la matrice de @xmath29 et est inversible car @xmath29 est surconvergent et @xmath42 est un corps ) ce qui fait que si @xmath911 , alors @xmath912 et donc @xmath913 .
la matrice @xmath914 dtermine @xmath292 lments de @xmath849 qui sont fixs par @xmath902 .
il reste montrer que ces lments sont libres sur @xmath14 quand @xmath710 est assez grand .
mais comme @xmath905 , la matrice @xmath914 va tre de rang @xmath292 pour @xmath710 assez grand ( puisque @xmath850 est de rang @xmath292 ) et donc dterminer un sous - module libre de rang @xmath292 de @xmath849 .
a fortiori le @xmath14-espace vectoriel engendr par les lments dtermins par @xmath914 va tre de dimension @xmath292 et donc gal @xmath54 .
[ isomcomp2 ] on a les isomorphismes de comparaison suivants : 1 .
si @xmath1 est une reprsentation semi - stable , alors @xmath915 = { \mathbf{d}_{\mathrm{st}}}(v ) \otimes_f { \mathbf{b}^{\dagger } _ { \mathrm{log } , k}}[1/t]\ ] ] 2 .
si @xmath1 est une reprsentation cristalline , alors @xmath61 = { \mathbf{d}_{\mathrm{cris}}}(v ) \otimes_f { \mathbf{b}^{\dagger } _ { \mathrm{rig } , k}}[1/t]\ ] ]
si de plus @xmath1 a ses poids de hodge - tate ngatifs , alors @xmath916 .
encore une fois on ne soccupe que du cas semi - stable , le cas cristallin sobtenant en faisant @xmath210 .
on peut supposer que @xmath1 a ses poids de hodge - tate ngatifs car cela revient multiplier par une puissance de @xmath62 le terme de gauche . on sait qualors
@xmath917 et que @xmath918 \otimes_{{\mathbf{b}^{\dagger } } _ k } { \mathbf{d}^{\dagger } } ( v)= { \widetilde{\mathbf{b}}^{\dagger } _ { \mathrm{log } , k}}[1/t ]
\otimes_f { \mathbf{d}_{\mathrm{st}}}(v)\ ] ] ce qui montre que si lon choisit des bases @xmath919 de @xmath54 et @xmath779 de @xmath896 , alors @xmath920 avec @xmath921)$ ] ; la proposition [ isomcomp ] implique dautre part que @xmath922 avec @xmath923)$ ] ; de plus @xmath924 . on peut alors
appliquer loprateur @xmath925 qui est @xmath926$]-linaire pour trouver @xmath927 ce qui fait que @xmath928 et que @xmath103 a donc ses coefficients dans @xmath926 $ ] et @xmath929)$ ] .
ceci permet de conclure .
les coefficients dune matrice de lisomorphisme de comparaison sont trs lis lexponentielle de perrin - riou @xcite .
[ detstdag ] soit @xmath1 une reprsentation semi - stable et soit @xmath850 la matrice de passage dune base
de @xmath54 une base de @xmath896 .
alors il existe @xmath930 et @xmath931 tels que @xmath932 .
le dterminant de la matrice de passage est gal au coefficient de la matrice de passage pour le dterminant de @xmath1 et il suffit donc de montrer lassertion en dimension @xmath69 .
une reprsentation semi - stable de dimension
@xmath69 est cristalline et est donc de la forme @xmath933 o @xmath145 est un caractre non - ramifi et @xmath934 est le caractre cyclotomique .
la priode de @xmath145 est alors un lment @xmath935 , ce qui fait que @xmath936 et @xmath937 do le rsultat .
dans ce paragraphe , on se place dans le cas @xmath36 et @xmath1 est une reprsentation cristalline de @xmath67 . on dit quune reprsentation @xmath0-adique @xmath1 de @xmath67 est de hauteur finie si @xmath101 possde une base sur @xmath938 forme dlments de @xmath939 .
un rsultat de fontaine @xcite ( voir aussi ( * ? ?
* iii.2 ) ) montre que @xmath1 est de hauteur finie si et seulement si @xmath101 possde un sous-@xmath940-module libre de type fini stable par @xmath29 de rang gal @xmath180 .
lobjet de ce paragraphe est de dmontrer le rsultat suivant : [ hf ] si @xmath1 est une reprsentation cristalline de @xmath67 , alors @xmath1 est de hauteur finie .
fixons une base @xmath779 de @xmath896 ainsi quune base @xmath897 de @xmath53 , et soit @xmath941 la matrice de passage de lune lautre cest - - dire que @xmath942 .
la matrice @xmath941 est coefficients dans @xmath943 $ ] . de plus
si lon remplace @xmath1 par @xmath944 , alors on peut remplacer @xmath945 par @xmath946 et donc @xmath941 par @xmath947 ce qui fait que quitte tordre suffisamment @xmath1 on peut supposer que @xmath941 est coefficients dans @xmath948 , ce que lon fait maintenant .
soit @xmath949 lanneau des sries formelles @xmath950 o @xmath51 et @xmath951 est de rayon de convergence @xmath69 ( cest - - dire quelles convergent sur le disque ouvert de rayon @xmath69 ) .
nous aurons besoin dun rsultat de kedlaya @xcite ( la formulation originale de kedlaya est @xmath952 mais on sy ramne en transposant ) : si @xmath941 est une matrice coefficients dans @xmath948 , alors il existe @xmath1 dans @xmath953
et @xmath954 dans @xmath955 telles que @xmath956 .
remarquons quon a ncessairement @xmath957 et comme @xmath958 est un corps cela implique que @xmath954 est inversible .
soit @xmath838 la matrice de @xmath29 dans la base @xmath779 ( qui a ses coefficients surconvergents ) , @xmath47 la matrice de @xmath29 dans la base @xmath897 ( qui est donc coefficients dans @xmath14 ) et @xmath102 la matrice de @xmath959 ( un gnrateur de @xmath960 ) dans la base @xmath779 ( elle est aussi surconvergente ) .
un petit calcul montre que , si @xmath961 est la base de @xmath896 dduite de @xmath779 par @xmath962 , alors : @xmath963 les coefficients de la matrice @xmath964 sont dans @xmath965 .
dautre part les coefficients de
@xmath966 sont dans @xmath958 . on en dduit que
dans la base @xmath961 , la matrice de @xmath29 a ses coefficients dans @xmath965 dune part et dans @xmath958 dautre part .
cest donc ( * ? ? ?
* ii.12 ) que @xmath967 .
de plus si @xmath968 , alors @xmath969 et comme @xmath970 on a @xmath971 . les coefficients de @xmath972 nont donc pas de ples en zro ce qui fait quil existe @xmath973 non - divisible par @xmath974 tel que @xmath975 .
pour les mme raisons , @xmath976 .
soit @xmath47 le @xmath940-module engendr par les @xmath977 .
on vient de voir que @xmath978 o @xmath973 nest pas divisible par @xmath974 et que le @xmath979-module engendr par @xmath47 est stable par @xmath960 . si @xmath47 est un @xmath940-module libre de type fini tel que le @xmath979-module engendr par @xmath47 est stable par @xmath960 et @xmath980 est tel que @xmath978 , alors il existe @xmath981 un sous - module libre
de type fini stable par @xmath29 et @xmath960 qui est de rang maximal .
pour montrer que la reprsentation @xmath1 est de hauteur finie il suffit donc , grce au rsultat de fontaine , de montrer le lemme ce que nous faisons maintenant .
le lemme est dmontr dans @xcite ( cest la runion des noncs iii.8 iii.15 ) mais lune des tapes utilise de manire cruciale que @xmath13 est fini et il nous faut la contourner .
un @xmath982-module libre de rang fini muni dune action de @xmath983 tel que le @xmath984-module @xmath985 soit muni dune action de @xmath29 commutant celle de @xmath983 et telle quil existe @xmath986 tel que lon ait @xmath987 .
alors @xmath988 . _ _ ''
cest cet nonc qui est dmontr et utilis dans la suite de @xcite . ] on va dabord montrer que lon peut supposer que @xmath989 .
soit @xmath990 ; on renvoie la consquence du corollaire iii.15 de @xcite pour la construction , sous lhypothse que le @xmath979-module engendr par @xmath47 est stable par @xmath960 , de @xmath991 tels que pour tout @xmath992 on ait @xmath993 .
comme @xmath994 est inversible dans @xmath940 , on peut supposer que @xmath974 ne divise pas @xmath995 ; en effet , linclusion @xmath996 implique que si
@xmath974 divise @xmath995 , alors @xmath974 divise @xmath997 .
soit alors @xmath998 le sous-@xmath940-module de @xmath999 engendr par les @xmath1000 o @xmath992 , @xmath1001 . comme @xmath940 est noetherien ( car principal ) , comme @xmath999 est de type fini , et comme @xmath998 est runion croissante de sous - modules de type fini de @xmath999 , cest que @xmath998 est en fait runion dun nombre fini de @xmath1002 . il existe donc @xmath1003 et @xmath1004 tels que @xmath1005
. de plus @xmath998 est stable par @xmath102 et donc par continuit il est stable par @xmath960 .
ensuite comme @xmath978 on a @xmath1006 .
soit @xmath1007 , on voit que @xmath1008 pour tout @xmath790 et donc que @xmath1009 et que @xmath974 ne divise pas @xmath1010 .
ceci montre que lon peut se ramener au cas o @xmath47 est stable par @xmath960 . on suppose donc maintenant que @xmath47 est un @xmath940-module libre de type fini stable par @xmath960 et que @xmath980 est tel que @xmath978 .
lidal @xmath309 des @xmath1011 tels que @xmath1012 est stable par @xmath960 et par ( * ? ? ?
* iii.8 ) il est de la forme @xmath1013 . de plus comme @xmath1014 et que @xmath974 ne divise pas @xmath1015 cest que @xmath1016 .
soit alors @xmath1017 .
un petit calcul montre que @xmath1018 et que @xmath1019 est stable par @xmath960 .
on peut donc prendre @xmath1020 et ceci achve la dmonstration du lemme et donc du thorme [ hf ] . lobjet de ce paragraphe est de montrer quil existe un anneau @xmath1021 , tel que @xmath1022 . on pose @xmath1023 .
le reste de ce paragraphe est consacr la preuve de quelques proprits de @xmath1021 , et du fait que @xmath1022 .
ces rsultats ne sont pas utiliss dans la suite de larticle . on a @xmath1024 .
voir la dmonstration du lemme 4.1 de @xcite . on a @xmath1025 .
on a @xmath1026 .
comme @xmath958 est un corps , il existe une base @xmath1027 de @xmath948 sur @xmath958 .
soit @xmath1028 , il scrit de manire unique @xmath1029 , avec @xmath1030 .
si @xmath476 est fixe par @xmath113 , cest donc ( comme les @xmath1031 sont fixes par @xmath113 ) que @xmath1032 .
le lemme en rsulte .
lapplication naturelle @xmath1033 dans @xmath10 est injective .
soit @xmath1034 , ayant la proprit suivante : cest le plus petit entier tel quil existe @xmath1035 , et @xmath1036 , tels que @xmath1037 dans @xmath10 , avec @xmath1038 .
alors , si @xmath1039 , on a @xmath1040 , et donc , @xmath1041 .
par lhypothse de minimalit sur @xmath13 , @xmath1042 , et donc @xmath1043 . on a alors @xmath1044 , et donc finalement , @xmath1045 .
lapplication @xmath1046 dans @xmath10 est donc injective . on a un isomorphisme @xmath1022 .
comme @xmath1047 , il suffit de montrer que @xmath1048 .
comme @xmath896 est un @xmath42-module libre de rang @xmath22 , on a : @xmath1049 si lon pose @xmath1050 $ ] , alors @xmath1051 .
lanneau @xmath50 est isomorphe ( canoniquement si @xmath36 ) lanneau de robba utilis dans la thorie des quations diffrentielles @xmath0-adiques .
ce chapitre est consacr la dmonstration de quelques unes de ses proprits relatives laction de @xmath30 : on dfinit aussi des oprateurs diffrentiels qui seront utiles pour la suite .
enfin , on montre un thorme de structure pour les modules sur lanneau de robba .
dans ce paragraphe , @xmath68 est un lment de @xmath30 et @xmath1052 . on suppose que @xmath1053 et que @xmath271 .
[ unmoinsgampetit ] si @xmath1054 $ ] est un intervalle ferm de @xmath521 , alors il existe @xmath1055 , tel que pour @xmath1056 , on ait @xmath1057 ds que @xmath1058 .
par densit et linarit on se ramne au cas o @xmath1059 avec @xmath1060 .
alors si @xmath250 @xmath1061 et sinon @xmath1062 comme on a @xmath1063 ( voir la remarque [ valanno ] ) , le lemme rsulte du fait que @xmath1064 si @xmath1065 est assez grand ; en effet , @xmath1066 tend vers @xmath31 quand @xmath68 tend vers @xmath69 .
si @xmath1054 $ ] , soit @xmath1067 lanneau des sries formelles en @xmath40 qui convergent sur la couronne de rayons intrieurs et extrieurs @xmath1068 et @xmath1069 .
le lemme prcdent montre que si @xmath68 est assez proche de @xmath69 , la srie doprateurs @xmath1070 converge vers un oprateur continu @xmath1071 , et un petit argument de sries formelles montre que @xmath82 ne dpend pas du choix de @xmath68 .
notamment @xmath1072 si @xmath1073 $ ] . on en dduit
que la valeur commune des @xmath1074 appartient @xmath12 , et que loprateur @xmath1075 , o @xmath1076 est la valeur commune des @xmath1074 , est continu pour la topologie de frchet .
de mme , si @xmath68 est assez proche de @xmath69 , alors la srie doprateurs @xmath1077 converge vers un oprateur continu @xmath1078 , et un petit argument de sries formelles montre que @xmath1079 ne dpend pas du choix de @xmath68 et dfinit aussi un oprateur de @xmath12 dans lui - mme , continu pour la topologie de frchet .
il est clair que @xmath1080 la restriction de @xmath71 @xmath12 vrifie @xmath1081 o @xmath1082 . attention au fait que ces notations ne sont pas compatibles avec @xcite .
ce qui est not @xmath1083 chez nous est not @xmath71 dans @xcite .
dans @xcite il est dmontr que limage de @xmath1084 par @xmath1085 dans @xmath172 est contenue dans @xmath1086 $ ] si @xmath1087 .
limage par @xmath445 de @xmath73 a la mme proprit par continuit .
comme @xmath445 est injectif et que @xmath71 , @xmath1088 commutent @xmath29 il suffit de montrer que @xmath1089 est nul sur @xmath1086 $ ] ce qui est vident .
montrons que @xmath1083 est presque surjective : [ invconn ] la connexion @xmath1083 ralise une surjection 1 .
de @xmath1090 dans @xmath50 ; 2 .
de @xmath719 dans lui - mme .
tout dabord si @xmath777 est lindice de ramification de @xmath1091 , alors @xmath1092 est une srie surconvergente ce qui fait que @xmath1093 $ ] .
rappelons que @xmath42 est un corps , et on peut donc se fixer @xmath271 tel que @xmath1094 et @xmath1095 appartiennent @xmath267
. soit @xmath1096 .
posons @xmath1097 , alors @xmath1098 ce qui montre que @xmath1099 avec @xmath1100 pour @xmath1101 .
ensuite la formule @xmath1102 montre que le ( 1 ) de la proposition implique le ( 2 ) par une rcurrence immdiate .
si @xmath1103 et @xmath1104 , alors on ne peut pas dire que @xmath1105 mais en revanche @xmath1106 pour tout @xmath1101 . les trois lemmes suivants seront utiles par la suite . [ divt ]
soit @xmath1056 tel que pour tout @xmath556 on ait @xmath1107 .
alors @xmath1108 .
soit @xmath1109 tel que @xmath1110 pour tout @xmath1111 .
rappelons que @xmath1112 sidentifie ( non canoniquement ) un anneau de sries formelles en @xmath40 .
soit @xmath1113 . on a @xmath1114 car une srie est divisible @xmath1115 si et seulement si les zros de la premire ont un ordre @xmath1116 ceux de la seconde , ce que lon peut vrifier localement . de mme , comme les @xmath1117 sont premiers entre eux , lhypothse du lemme est quivalente au fait que , pour tout @xmath556 , on ait @xmath1118 puisque @xmath1119 est inversible dans @xmath1112 . on peut donc crire @xmath1120 et on va montrer que la suite @xmath1121
converge dans @xmath1112 .
les @xmath1122 vrifient la relation : @xmath1123 et donc @xmath1124 on en dduit que si @xmath309 est un intervalle contenu dans @xmath1125 , alors @xmath1126 on se fixe @xmath309 un intervalle compact et comme @xmath1127 sur toute couronne du type @xmath1128 , on peut supposer que @xmath1129 pour @xmath77 assez grand .
alors lingalit ci - dessus montre que @xmath1130 et donc que @xmath1131 .
les @xmath1132 ont donc tous la mme valuation pour @xmath77 assez grand .
enfin lingalit ci - dessus et le fait que @xmath1133 impliquent que @xmath1134 et donc que la suite @xmath1132 converge .
ceci tant vrai pour tout @xmath309 , la suite @xmath1132 converge pour la topologie de frchet vers une limite @xmath1135 et un calcul immdiat montre que @xmath1136 . on peut aussi
dire que @xmath1137 est une fonction mromorphe sans ples , et cela la force tre holomorphe , cest - - dire que @xmath1138 .
cela suit de rsultats gnraux de lazard @xcite sur les fonctions analytiques , que lon trouvera dans le paragraphe suivant .
[ kertheta ] soit @xmath234 et @xmath1139 .
alors @xmath1140 @xmath1141 le premier point est une consquence immdiate de [ kertheta1 ] .
pour le deuxime si @xmath1142 est tel que @xmath1143 , alors @xmath1144 avec @xmath1145 et alors @xmath1146 ce qui fait que @xmath1147 et donc que @xmath1148 .
[ iotasurj ] si @xmath234 et @xmath1149 , alors lapplication injective @xmath1150 est une bijection .
tant donn que @xmath1151 , lextension de corps @xmath1152 sur @xmath1153 est de degr @xmath1154 $ ] .
la proposition est triviale dans le cas @xmath36 ( on a dj @xmath1155 ) , et dans le cas gnral , elle rsulte alors du fait que @xmath1152 est une extension de @xmath1156 de degr @xmath1157 $ ] pour @xmath556 .
on tablit ici des rsultats techniques qui serviront dmontrer un rsultat sur la structure de @xmath55 quand @xmath1 est de de rham .
soit @xmath1158 , tel que @xmath1159 .
lanneau @xmath1112 est muni de sa topologie de frchet , dfinie par les @xmath1160 , et si @xmath850 est un @xmath1112-module libre de rang @xmath22 , alors le choix dun isomorphisme @xmath1161 permet de munir @xmath850 dune topologie .
le thorme de limage ouverte pour les espaces de frchet montre que cette topologie ne dpend pas du choix dune base de @xmath850 .
lobjet de ce paragraphe est de montrer le thorme suivant et son corollaire : [ libre ] soit @xmath850 un @xmath1112-module libre de rang fini @xmath22 , et @xmath1162 un sous - module ferm pour la topologie de frchet de @xmath850 , tel que @xmath1163 alors @xmath66 est libre de rang @xmath22 .
[ libre2 ] soit @xmath850 un @xmath1112-module libre de rang fini @xmath22 , et @xmath1162 un sous - module ferm pour la topologie de frchet de @xmath850 .
alors @xmath66 est libre de rang @xmath1164 , o @xmath1165 ces deux rsultats sont , apparamment , bien connus des experts , mais nous en reproduisons la dmonstration pour la commodit du lecteur .
la dmonstration va ncessiter quelques rsultats prparatoires sur la structure de @xmath1112 .
ces rsultats se trouvent dans @xcite , moyennant une identification de lanneau @xmath1112 lanneau des sries formelles convergeant sur la couronne @xmath1166 , avec @xmath1167 .
si @xmath309 est un
intervalle de @xmath1168 , soit @xmath1067 lanneau des sries de laurent coefficients dans @xmath14 , qui convergent sur la couronne @xmath1169 .
soit @xmath309 un intervalle de @xmath1168 .
lanneau @xmath1067 a alors les proprits suivantes : 1 .
si @xmath309 est un intervalle compact , alors @xmath1067 est un anneau principal ; 2 .
tout idal de type fini de @xmath1067 est principal ( cest donc un anneau de bezout ) ; 3 .
tout idal ferm de @xmath1067 est principal ; 4 .
toute fonction mromorphe @xmath1170 , qui na pas de ples , est en fait holomorphe , cest - - dire que @xmath1171 ; 5 .
toute matrice @xmath22 lignes et @xmath1172 colonnes , @xmath1173 , peut scrire @xmath1174 o @xmath1175 , @xmath1176 et @xmath1177 est nulle sauf sur la diagonale , et @xmath1178 . en dautres
termes , @xmath1067 admet une thorie des diviseurs lmentaires .
pour les 4 premiers points , on se reportera larticle de lazard @xcite .
pour le cinquime point : on commence par montrer que @xmath1067 est un anneau _ adquat _ , cest - - dire que @xmath1067 est un anneau intgre o tout idal de type fini est principal , et qui vrifie la condition technique suivante : pour tous @xmath1179 , on peut crire @xmath1180 avec @xmath1181 et @xmath1182 pour tout @xmath1183 qui divise @xmath1184 ( et qui nest pas une unit ) .
cest une consquence des rsultats de @xcite , qui permettent de construire des fonctions holomorphes do nt lensemble des zros est fix lavance .
le ( 5 ) est alors une consquence
directe de ( * ? ? ?
* theorem 3 ) .
enfin , on aura besoin du lemme suivant : [ sousbez ] soit @xmath850 un @xmath1067-module , libre de rang @xmath22 , et @xmath1185 un sous - module de type fini .
alors @xmath1186 est libre de rang @xmath870 .
si @xmath1187 , alors @xmath1186 est libre de rang @xmath22 .
si lon exprime les coordonnes dune famille gnratrice de @xmath1186 dans une base de @xmath850 , cest une consquence immdiate de la thorie des diviseurs lmentaires .
la dmonstration est largement inspire de @xcite .
soit donc @xmath850 un @xmath1188-module libre de type fini , et @xmath1189 un sous - module ferm .
soit @xmath1190 .
si @xmath309 est un
intervalle compact , soit @xmath1191 lensemble des @xmath1192 qui engendrent le @xmath1067-module @xmath1193 .
on remarquera que si @xmath309 est ferm , alors @xmath1067 est principal , et donc @xmath1194 est un @xmath1067-module libre de rang fini @xmath870 .
comme @xmath1195 alors @xmath1194 est en fait un @xmath1067-module libre de rang gal @xmath22 .
montrons premirement que @xmath1191 est non - vide : comme lanneau @xmath1067 est principal , il existe @xmath1196 qui forment une base de @xmath1194 . on a @xmath1197 o @xmath1198 , et @xmath1199 est une famille de @xmath66 .
le @xmath1112-module engendr par les @xmath1200 est un sous - module de type fini de @xmath850 , il est donc libre de rang @xmath22 , de base @xmath1201 .
il nest pas difficile de voir que les @xmath1202 forment une base de @xmath1194 sur @xmath1067 et donc que @xmath1203 est non - vide .
ceci implique que @xmath1191 est non - vide pour @xmath1204 .
montrons ensuite que @xmath1191 est un ouvert de @xmath1205 .
comme @xmath1203 est non - vide , on peut se fixer @xmath1206 .
soit @xmath1207 , et @xmath838 la matrice des @xmath1200 dans la base @xmath1208 .
comme @xmath1112 admet la thorie des diviseurs lmentaires , on peut crire @xmath1209 , avec @xmath1210 , @xmath1211 , et @xmath47 nulle hors de la diagonale .
la matrice @xmath301 induit un homomorphisme de @xmath1205 .
quitte modifier la base @xmath1212 par @xmath1213 , et faire le changement de coordonnes sur @xmath1205 induit par @xmath301 , on se ramne supposer que @xmath1214 et @xmath1215 si @xmath1216 .
soit @xmath1217 la matrice @xmath1218 des @xmath1219 avec @xmath1220 .
par hypothse , @xmath1221 . or lensemble des @xmath1222 do nt le dterminant vrifie @xmath1223 est ouvert ( le dterminant est une application continue et @xmath1224 est ouvert dans @xmath1067 , puisque , @xmath309 tant compact , @xmath1067 est une algbre de banach ) et contient @xmath1217 .
lensemble des @xmath1225 obtenus via @xmath1226 est ouvert et contient @xmath1227 ,
si lon prend pour @xmath838 toutes les matrices @xmath1228 telles que : 1 .
le dterminant de la matrice @xmath1218 des @xmath22 premires lignes de @xmath838 est inversible dans @xmath1067 ; 2 . la matrice @xmath1229 des @xmath1230 dernires lignes de @xmath838 est quelconque .
alors lensemble des @xmath1225 obtenus via @xmath1226 est ouvert et contient @xmath1227 .
ceci montre que @xmath1191 est ouvert .
montrons enfin que si @xmath1231 , alors @xmath1191 est dense dans @xmath1205 .
soient @xmath1206 , et @xmath1232 .
comme auparavant , la thorie des diviseurs lmentaires permet de supposer que @xmath1214 et @xmath1215 si @xmath1233 .
soit @xmath1234 , et @xmath1235 pour @xmath1236 , @xmath1237 pour @xmath1238 , et @xmath1239 si @xmath1240 .
la famille @xmath1241 est une famille gnratrice de @xmath1242 , aussi proche de @xmath1243 que lon veut ( le choix de @xmath77 est libre ) .
lensemble @xmath1244 est donc un ouvert dense de @xmath1245 , et comme @xmath66 est un ferm dun frchet , cest un mtrique complet : il a la proprit de baire .
lintersection @xmath1246)\ ] ] est donc dense dans @xmath1245 , et notamment non - vide .
dans lintersection .
le sous @xmath1112-module de @xmath66 engendr par les @xmath1248 est de type fini et @xmath1249 , il est donc libre de rang @xmath22 , engendr par @xmath1250 .
on voit que ces @xmath1212 sont une base de @xmath1194 pour tout @xmath309 .
soit @xmath1251 .
on peut donc crire @xmath1252 avec @xmath1253 .
par unicit , on a @xmath1254 .
le module @xmath66 est donc libre de rang @xmath22 , tant engendr par les @xmath1212 .
soit @xmath1255 .
montrons que cest un @xmath1112-module libre de rang @xmath1164 . on a une suite exacte @xmath1256 , et par construction , @xmath1257 est sans - torsion : cest un module de type fini et sans torsion , donc cohrent , ce qui fait ( * ? ? ? * i.2 exercice 11 ) que @xmath1217 lui - mme est cohrent .
comme il est de type fini , le lemme [ sousbez ] montre quil est libre de rang fini @xmath1172 .
ensuite , il est immdiat que @xmath1258 , et on peut alors appliquer le thorme [ libre ] .
dans le chapitre prcdent on a fait ltude de @xmath1259 . ce chapitre est consacr la construction de lquation diffrentielle associe une reprsentation @xmath1 , qui est un module connexion au - dessus de @xmath1259 . on donne
ensuite des applications aux reprsentations semi - stables , la thorie de sen , et la caractrisation des reprsentations de de rham .
comme consquence de cela , on dmontre la conjecture de monodromie @xmath0-adique .
dans tout ce paragraphe , @xmath1 est une reprsentation @xmath0-adique de @xmath23 .
rappelons que lon a pos @xmath1260 et notamment que @xmath1261 . le lemme de rgularisation par le frobenius montre quon a @xmath1262 et donc que lon peut rcuprer @xmath1 par la formule @xmath1263 on choisit une base @xmath779 de @xmath896 et on prolonge les oprateurs @xmath903 @xmath1264 par @xmath1265 .
est @xmath42-linaire , le prolongement ne dpend pas du choix de la base @xmath779 .
nous aurons besoin du rsultat suivant qui est un corollaire immdiat de la proposition [ kertheta1 ] : [ kerthetadhol ] soit @xmath234 et @xmath1139
. alors @xmath1266 la proposition [ kertheta1 ] montre que si @xmath1143 , alors @xmath1144 avec @xmath1267 . on a alors
@xmath1146 et donc @xmath1268 .
enfin pour le deuxime point il suffit dutiliser le fait quun lment de @xmath1084 divisible par @xmath1117 dans @xmath73 lest dans @xmath1084 . on va munir @xmath55 dune connexion @xmath83 au - dessus de @xmath1259 , cest - - dire dun oprateur @xmath83 tel que @xmath1269 . on se fixe une base @xmath779 de @xmath896 et si @xmath1270 scrit @xmath1271 avec @xmath1272 , alors on pose ( si @xmath309 est tel que @xmath1273 ) : @xmath1274
la dfinition dpend de la base mais pas la topologie dfinie par les @xmath385 qui est quivalente la topologie induite sur @xmath55 par @xmath1275 , et @xmath1276 est complet .
dans cette partie @xmath68 est un lment de @xmath30 et @xmath1052 . on suppose que @xmath1053 et que @xmath271 .
si @xmath1054 $ ] est un intervalle ferm de @xmath521 , alors il existe @xmath1277 , tel que pour @xmath1270 , on ait @xmath1057 ds que @xmath1278 .
si @xmath1279 , alors @xmath1280 , quand @xmath1065 est assez grand , puisque laction de @xmath23 est continue . on se fixe @xmath1065 tel que @xmath1281
pour tout @xmath790 et le lemme rsulte alors du lemme [ unmoinsgampetit ] ( le cas de la reprsentation triviale ) puisque @xmath1282 si @xmath1054 $ ] , soit @xmath1067 lanneau des sries formelles en @xmath40 qui convergent sur la couronne de rayons intrieurs et extrieurs @xmath1068 et @xmath1069 .
le lemme prcdent montre que si @xmath68 est assez proche de @xmath69 , la srie doprateurs @xmath1070 converge vers un oprateur continu @xmath1283 , et un petit argument de sries formelles montre que @xmath82 ne dpend pas du choix de @xmath68 .
notamment @xmath1284 si @xmath1073 $ ] . on en dduit
que la valeur commune des @xmath1285 appartient @xmath1276 , et que loprateur @xmath1286 , o @xmath1287 est la valeur commune des @xmath1285 , est continu pour la topologie de frchet .
ensuite , un argument standard montre que si @xmath1288 , alors @xmath1289 ce qui fait que , comme @xmath1290 , @xmath83 vrifie @xmath1291 . cest donc une connexion au - dessus de loprateur @xmath1292
. un calcul facile montre par exemple que @xmath1293 sur @xmath1294 .
[ divtv ] soit @xmath476 un lment de @xmath1276 , tel que pour tout @xmath556 , on ait @xmath1295 .
alors @xmath1296 .
aprs le choix dune base de @xmath55 cela suit immdiatement de [ divt ] .
un cristal sur @xmath50 est un @xmath50-module libre muni dun frobenius et dune connexion qui commutent , la connexion tant au - dessus de @xmath71 .
commenons par dfinir ce quest un cristal unipotent .
on remarquera que cela ne dpend pas de lventuelle structure de frobenius .
soit donc @xmath1297 un @xmath50-module libre de rang fini @xmath22 muni dune connexion @xmath1298 au - dessus de @xmath71 .
les proprits suivantes sont quivalentes : 1 .
@xmath1298 est triviale sur @xmath1299 , cest - - dire quil existe @xmath1300 tels que @xmath1301 et @xmath1302 = { m}\otimes_{{\mathbf{b}^{\dagger } _
{ \mathrm{rig } , k}}}{\mathbf{b}^{\dagger } _ { \mathrm{log } , k } } [ 1/t]$ ] ; 2 .
il existe @xmath22 lments @xmath1303 de @xmath1297 qui forment une base de @xmath1304 $ ] sur @xmath1305 $ ] et tels que @xmath1306 o @xmath1307 dnote le @xmath50-module engendr .
commenons par montrer que ( 1 ) implique ( 2 ) . loprateur de monodromie @xmath1308 laisse stable le @xmath14-espace vectoriel engendr par les @xmath786 , car le ( 1 ) implique que les @xmath786 engendrent le noyau de @xmath1298 agissant sur @xmath1299 , et un calcul direct montre que @xmath66 commute @xmath1298 ( car @xmath66 commute laction de @xmath67 ) et donc stabilise son noyau : on peut alors supposer que @xmath1309 car @xmath66 est nilpotent . on peut
crire de manire unique @xmath1310 .
montrons que @xmath1311 est une famille qui satisfait la condition de ( 2 ) . le fait que @xmath1312 et @xmath1309 implique respectivement que @xmath1313 ce qui fait que @xmath1314 .
montrons enfin que les @xmath977 engendrent @xmath1315 $ ] sur @xmath1305 $ ] .
soit @xmath1316 $ ] , par hypothse on peut crire @xmath1317 et donc @xmath1318 o @xmath1319 est le terme constant de @xmath1320 et comme @xmath1316 $ ] on a ncssairement @xmath1321 .
montrons maintenant limplication rciproque . on va montrer par rcurrence que lon peut prendre @xmath1322 avec @xmath1323 et @xmath1324 .
en rang @xmath69 il ny a rien montrer . en rang @xmath22 , @xmath1298 induit une connexion sur @xmath1325 qui satisfait les mmes conditions et il existe donc @xmath1326 tels que @xmath1327 .
un calcul immdiat montre que @xmath1328 et donc quil existe @xmath1329 tel que @xmath1330 ( par la proposition [ invconn ] , il existe @xmath1329 tel que @xmath1331 ) . on pose alors
@xmath1332 et @xmath1333 , ce qui achve la rcurrence .
la matrice de passage des @xmath977 aux @xmath786 est triangulaire avec des @xmath69 sur la diagonale et on en dduit que les @xmath786 engendrent bien @xmath1334 $ ] sur @xmath926 $ ] . un cristal qui satisfait les conditions de la proposition prcdente est dit unipotent .
[ conilp ] soit @xmath1 une reprsentation @xmath0-adique et @xmath55 le cristal quon lui a associ .
alors il existe @xmath77 tel que la restriction de @xmath1 @xmath1335 est semi - stable ( respectivement cristalline ) poids ngatifs si et seulement si @xmath55 est unipotent ( respectivement trivial ) . on a vu que @xmath1 est une reprsentation semi - stable de @xmath1335 si et seulement si @xmath1336\ ] ] a une base dlments stables par @xmath1337 .
si @xmath1 est semi - stable , alors @xmath83 est donc triviale , cest - - dire que si @xmath1338 est une base de @xmath54 , alors ils satisfont le ( 1 ) de la proposition prcdente en raison du thorme de comparaison : @xmath1339= { \mathbf{d}^{\dagger } _ { \mathrm{rig}}}(v)\otimes_{{\mathbf{b}^{\dagger } _
{ \mathrm{rig } , k } } } { \mathbf{b}^{\dagger } _ { \mathrm{log } , k}}[1/t]\ ] ] rciproquement si @xmath55 est unipotent , alors les @xmath1338 engendrent un @xmath14-espace vectoriel sur lequel @xmath1340 agit trivialement ce qui fait que @xmath30 agit travers un quotient fini et donc que les @xmath786 sont stables par @xmath1341 pour @xmath77 assez grand et forment alors une base de @xmath1342 si @xmath1343 est la restriction de @xmath1 @xmath1335 . on dit aussi dans ce cas que la connexion est unipotente .
de plus @xmath1 est cristalline si et seulement si on peut choisir les @xmath977 tels que @xmath1344 et la connexion est alors triviale .
rappelons que sen a montr @xcite que si @xmath1 est une reprsentation @xmath0-adique , alors lensemble des sous-@xmath96-espaces vectoriels de dimension finie de @xmath1347 stables par @xmath30 admet un plus grand lment @xmath1345 et que @xmath1348 . de plus si @xmath1349 est suffisament proche de @xmath69 , alors la srie qui dfinit @xmath1340 converge en tant que srie doprateurs @xmath21-linaires de @xmath1345 et loprateur @xmath1350 est un oprateur @xmath96-linaire qui ne dpend pas de @xmath68 .
dautre part fontaine a montr dans @xcite que lensemble des sous-@xmath1351$]-modules libres de type fini de @xmath1352 stables par @xmath30 contient un plus grand lment ; nous le noterons @xmath1353 dans cet article et on a @xmath1354}{\mathbf{d}_{\mathrm{dif}}}^+(v)= { \mathbf{b}_{\mathrm{dr}}}^+ \otimes_{{\mathbf{q}_p } } v$ ] . de plus si @xmath1355 est suffisament proche de @xmath69 , alors la srie qui dfinit @xmath1340 converge en tant que srie doprateurs @xmath21-linaires de @xmath1353 et loprateur @xmath1356 est un oprateur qui ne dpend pas de @xmath68 et qui vrifie @xmath1357 ce qui montre que @xmath83 est une connexion sur @xmath1353 ; on ltend @xmath1358 } { \mathbf{d}_{\mathrm{dif}}}^+(v)$ ] . on peut retrouver
@xmath1359 partir de @xmath1360 via lapplication @xmath1361 comme nous le verrons ci - dessous .
rappelons @xcite que lapplication @xmath1085 envoie @xmath1084 dans @xmath1086 \subset { \mathbf{b}_{\mathrm{dr}}}^+$ ] et envoie donc @xmath1362 dans un sous @xmath1086$]-module de @xmath1353 . [ phiconn ] lapplication de @xmath1363 dans @xmath1346 dduite de @xmath493 est un isomorphisme de @xmath1364-modules avec connexion , si @xmath77 est assez grand . on prend
@xmath1111 tel que @xmath1365 $ ] .
il est clair qualors @xmath1366 \otimes_{\iota_n({\mathbf{b}^{\dagger , r_n}}_k ) } \iota_n({\mathbf{d}^{\dagger , r_n}}(v))\ ] ] est un sous @xmath1351$]-module libre de type fini de @xmath1367 stable par @xmath30 .
montrons que cest @xmath1353 pour @xmath556 . on a une application @xmath1368
et @xmath1345 est un @xmath96-espace vectoriel de dimension @xmath180 . on en dduit une application
@xmath1369 do nt le noyau est @xmath1370 par le lemme [ kerthetadhol ] ce qui fait que @xmath1371 ralise une injection dun @xmath1372 module de rang @xmath22 dans @xmath1345 et son image est un @xmath97-espace vectoriel @xmath1343 de dimension @xmath22 stable par @xmath30 .
un petit calcul montre que des lments de @xmath1343 qui sont lis dans @xmath1345 par une relation coefficients dans @xmath96 , sont lis par une relation coefficients dans @xmath97 , et donc que lapplication naturelle de @xmath1373 dans @xmath1345 est injective ; son image est de dimension maximale @xmath22 , et est donc gale @xmath1345 .
cela implique que le dterminant de linjection @xmath1366 \otimes_{\iota_n({\mathbf{b}^{\dagger , r_n}}_k ) } \iota_n({\mathbf{d}^{\dagger , r_n}}(v ) ) \hookrightarrow { \mathbf{d}_{\mathrm{dif}}}^+(v)\ ] ] nest pas divisible par @xmath62 et donc ( comme @xmath1351 $ ] est un anneau local didal maximal @xmath1374 ) par le lemme de nakayama que linjection ci - dessus est en fait un isomorphisme .
par extension des scalaires on en dduit que lapplication @xmath1375 dduite de @xmath493 est aussi un isomorphisme de @xmath1364-modules avec connexion , si @xmath77 est assez grand .
grce ces calculs on peut retrouver @xmath58 partir de @xmath896 ; la proposition suivante se trouve dans ( * ? ? ?
3.25 ) .
[ kernabla ] si @xmath1 est une reprsentation @xmath0-adique de @xmath23 , alors @xmath1376 est le noyau de la connexion @xmath83 oprant sur @xmath1346 . en particulier , @xmath1 est de de rham si et seulement si @xmath83 est la connexion triviale . laction de @xmath83 sur @xmath1376 est triviale et de plus si @xmath1377 , alors @xmath1378 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] est un sous-@xmath1351$]-module de @xmath1367 libre de type fini et stable par @xmath30 ce qui montre que @xmath1376 est bien inclus dans le noyau de @xmath83 agissant sur @xmath1346 .
la thorie gnrale des modules connexion montre que le noyau de @xmath83 sur @xmath1346 est un @xmath96-espace vectoriel de dimension finie au plus @xmath1379 ; il est aussi stable par @xmath30 et provient donc par extension des scalaires dun @xmath97-espace vectoriel stable par @xmath30 pour @xmath556 .
laction de @xmath30 sur ce @xmath97-espace vectoriel est discrte car lalgbre de lie de @xmath30 agit trivialement , puisque @xmath1380 ( cf @xcite ) , et donc quitte augmenter @xmath77 laction de @xmath1381 est triviale et cet espace est donc inclus dans @xmath1382 o @xmath1343 est la restriction @xmath1335 de @xmath1 , ce qui fait que le noyau de la connexion agissant sur @xmath1346 est inclus dans @xmath1376 . on a donc bien que @xmath1376 est le noyau de @xmath83 sur @xmath1346 .
soit @xmath92 .
lobjet de ce paragraphe est de dmontrer le thorme suivant , et sa rciproque , la proposition [ recipndr ] : [ ndr ] soit @xmath1 une reprsentation de de rham de @xmath23 . on suppose que les poids de hodge - tate de @xmath1 sont ngatifs .
alors il existe un unique sous @xmath50-module libre de rang @xmath22 de @xmath55 stable par @xmath94 , @xmath93 .
il vrifie de plus les proprits suivantes : 1 .
@xmath93 est stable par @xmath29 et par @xmath30 ; 2 . si @xmath293 est assez grand , alors il existe @xmath1383 , libre de rang @xmath22 , stable par @xmath94 et @xmath30 , tel que @xmath1384 . on va montrer
que @xmath1385 , o @xmath1386 lensemble des @xmath1387 , tels que pour tout @xmath1388 , on ait @xmath1389 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] .
le module @xmath93 que lon construit ainsi est lanalogue du module @xmath66 construit par n. wach @xcite dans le cas dune reprsentation absolument cristalline , ce qui explique la notation .
la dmonstration de ce rsultat va ncessiter plusieurs tapes intermdiaires , consacres la construction dlments de @xmath55 .
pour linstant , @xmath1 est une reprsentation @xmath0-adique quelconque .
on commence par le cas de la reprsentation triviale : [ dense ] soient @xmath1158 , et @xmath866 , tels que @xmath1390 , et @xmath1391 .
si @xmath1392 $ ] , alors il existe @xmath1393 tel que @xmath1394 $ ] . on cherche
montrer que @xmath1395 est dense dans @xmath1086 $ ] pour la topologie @xmath62-adique .
comme @xmath1396 , il suffit de montrer que lapplication naturelle @xmath1397/(t)$ ] est surjective . or , cette application concide avec @xmath1398 , qui est surjective par le lemme [ iotasurj ] .
[ partunit ] soit @xmath1399 . on a @xmath1400 si @xmath1401 , alors @xmath1402 sinon , si @xmath1403 @xmath1404 et si @xmath1405 @xmath1406 puisque @xmath1407 .
soit @xmath779 une base de @xmath896 . on suppose dans
la suite que @xmath293 est un rel tel que @xmath1408 , et que la matrice de @xmath29 dans cette base a ses coefficients dans @xmath1112 .
enfin , on suppose aussi que @xmath1094 est inversible dans @xmath1112 .
bien entendu , toutes ces conditions sont vrifies si @xmath293 est assez grand .
si @xmath1409 et @xmath1391 , alors il existe @xmath1410 , et @xmath1411 tels que : @xmath1412 . on a vu en [ phiconn ] que si @xmath77 est assez grand , alors @xmath1366 \otimes_{\iota_n({\mathbf{b}^{\dagger , r_n}_{\mathrm{rig } , k } } ) } \iota_n({\mathbf{d}^{\dagger , r_n}_{\mathrm{rig}}}(v ) ) = { \mathbf{d}_{\mathrm{dif}}}^+(v).\ ] ] si @xmath77 vrifie de plus @xmath1413 , alors comme on a suppos que @xmath1414 , on a @xmath1351 \otimes_{\iota_n({\mathbf{b}^{\dagger , s}_{\mathrm{rig } , k } } ) } \iota_n({\mathbf{d}^{\dagger , s}_{\mathrm{rig}}}(v ) ) = { \mathbf{d}_{\mathrm{dif}}}^+(v)$ ] . on peut donc crire @xmath1415 avec
@xmath1416 $ ] .
soit @xmath1417 , tel que pour tout @xmath790 , @xmath1418+t^w k_{\infty}[[t]]$ ] .
soit @xmath1419 la matrice de @xmath1420 dans la base @xmath786 ( si @xmath1421 est la matrice de @xmath29 , on a @xmath1422 et si @xmath1423 , alors @xmath1424 ) . on a @xmath1425 o @xmath1426+t^w k_{\infty}[[t]]$ ] , comme on le vrifie rapidement . par le lemme [ dense ] , il existe @xmath1427 , tels que @xmath1428 $ ] . on peut alors
poser @xmath1429 . on conserve les notations du lemme prcdent .
[ constrdrig ] si @xmath1430 , alors pour tout @xmath1111 , il existe @xmath1431 tel que @xmath1432 .
soit @xmath1433 llment construit dans le lemme prcdent .
il satisfait donc : @xmath1412 .
notamment , @xmath1434 .
soit @xmath1435 , alors @xmath1436 , et donc @xmath1437 avec @xmath1438 et @xmath1439 .
le lemme [ dense ] fournit des @xmath1440 tels que @xmath1441 $ ] . on pose alors
@xmath1442 . on suppose maintenant que la reprsentation @xmath1 est de de rham , et que ses poids de hodge - tate sont ngatifs , ce qui fait quil existe @xmath1391 tel que @xmath1443 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v ) \subset { \mathbf{d}_{\mathrm{dif}}}^+(v)\ ] ] ensuite , il existe @xmath1444 tel que , si @xmath1445 , alors @xmath1446 .
soient @xmath1447 une base du @xmath15-espace vectoriel @xmath58 , et @xmath1448 .
soit @xmath1386 lensemble des @xmath1387 , tels que pour tout @xmath1388 , on ait @xmath1389 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] .
alors @xmath1386 est un @xmath1112-module libre de rang @xmath22 , et @xmath1449 .
commenons par remarquer que @xmath1386 contient @xmath1450 , en particulier il est non - vide , et il contient un sous @xmath1112-module libre de rang @xmath22 .
ensuite , il est ferm pour la topologie de frchet de @xmath1451 , car @xmath493 est continu pour tout @xmath1452 .
le thorme [ libre ] montre alors que @xmath1386 est libre de rang @xmath22 .
il en existe donc une base @xmath1453 .
pour @xmath1111 , la famille @xmath1454 forme une base de @xmath1086 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] : en effet , il suffit de vrifier que @xmath1455 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] , cest dire de montrer que pour tout @xmath790 , @xmath1456 .
la proposition [ constrdrig ] , applique @xmath1457 , fournit @xmath1458 , tel que @xmath1459 .
soit @xmath1460 .
le lemme [ partunit ] montre quil existe @xmath1461 $ ] , tel que @xmath1462 , et que si @xmath1463 , alors @xmath1464 . on en dduit que @xmath1465 ,
et donc que @xmath1456 .
ensuite , il est clair que @xmath1466 , car @xmath83 commute @xmath493 , et @xmath1086 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] est stable par @xmath83 .
soit @xmath1467 la matrice de la connexion @xmath83 dans la base @xmath961 .
le fait que @xmath1468 \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v ) ) \subset t k_n[[t ] ] \otimes_k { \mathbf{d}_{\mathrm{dr}}}(v)$ ] , montre que pour tout @xmath1111 , @xmath1469)$ ] , et donc par le lemme [ kertheta1 ] , @xmath1470 .
ceci tant valable pour tout @xmath1111 , le lemme [ divt ] montre que @xmath1471 , et donc que @xmath1472
. montrons maintenant quun sous - module de @xmath1451 , do nt le rang est maximal , et qui est stable par @xmath94 , est unique .
[ drdet ] soient @xmath1 une reprsentation de de rham , do nt les poids de hodge - tate sont ngatifs , et @xmath1473 un @xmath1112-module libre de rang @xmath22 , inclus dans @xmath1451 .
si @xmath1474 , alors @xmath1475 , o @xmath1476 .
la reprsentation @xmath1477 est de de rham , et donc @xcite de la forme @xmath1478 ( @xmath1476 , cest loppos de la somme des poids de hodge - tate de @xmath1 ) , o la restriction de @xmath145 @xmath1479 ( linertie de @xmath23 ) est finie .
notamment , il existe une base @xmath1172 de @xmath1480 , telle que laction de @xmath30 sur @xmath1172 est donne par @xmath1481 , o @xmath1482 est dordre fini . on a donc @xmath1483 , si @xmath68 est assez proche de @xmath69 .
il existe @xmath1484 , tel que @xmath1485 , et cest un sous - module de @xmath1486 de rang @xmath69 , stable par @xmath94 .
on doit avoir : @xmath1487 , avec @xmath1488 et donc @xmath1489 .
comme @xmath1490 , cela montre que @xmath62 divise @xmath1015 , si @xmath1491 .
une rcurrence facile , sur @xmath292 , montre quil existe @xmath1393 , tel que @xmath1492 . on doit alors
avoir @xmath1493 , avec @xmath1494 .
identifions cet anneau avec lanneau des sries en @xmath33 , holomorphes sur la couronne @xmath1495 .
comme on a suppos que @xmath1094 est inversible dans @xmath1112 , si @xmath1496 , alors @xmath1497 , et par rcurrence @xmath1498 pour tout @xmath13 .
si @xmath1010 sannule quelque part , cest donc que @xmath1499 . on en conclut que @xmath1010 est inversible et donc que @xmath1500 .
@xmath1501 et @xmath1502 sont deux sous - modules libres de rang @xmath22 de @xmath1451 , stables par @xmath94 , alors @xmath1503 .
le module @xmath1504 vrifie les mmes hypothses , et @xmath1505 .
le lemme [ drdet ] montre que @xmath1506 , et donc que @xmath1507 .
le lemme prcdent montre que si @xmath1508 , alors linclusion @xmath1509 est en fait un isomorphisme
. posons @xmath1510 : on vient de voir que cette dfinition ne dpend pas du choix de @xmath293 . de plus , @xmath1511 est un sous - module libre de rang @xmath22 de @xmath1512 , stable par @xmath94 , et par unicit on a donc @xmath1513 .
par suite : @xmath1514 et @xmath93 est donc stable par @xmath29 .
il est clair que @xmath1386 , et donc @xmath93 , est stable par @xmath30 ; par construction il est stable par @xmath94 .
enfin il nest pas difficile de voir que @xmath93 dtermine @xmath1386 , qui est unique , et donc @xmath93 lui - mme est unique .
montrons maintenant la rciproque du thorme [ ndr ] : [ recipndr ] soit @xmath1 une reprsentation @xmath0-adique de @xmath23 . on suppose quil existe @xmath1383 , libre de rang @xmath22 et stable par @xmath94 .
alors @xmath1 est de de rham , et ses poids de hodge - tate sont ngatifs .
si @xmath77 est assez grand , alors @xmath1515 \otimes_{\iota_n({\mathbf{b}^{\dagger , s}_{\mathrm{rig } , k } } ) } \iota_n(n_s)$ ] est un sous @xmath1086$]-module de @xmath1353 , et @xmath1516 \otimes_{k_n[[t ] ] } m_n$ ] est un sous @xmath1351$]-module de @xmath1353 , de rang @xmath870 .
de plus , comme @xmath1517 @xmath850 est un @xmath1351$]-module libre de rang @xmath22 .
il est stable par @xmath94 , soit @xmath1213 la matrice de @xmath94 dans une base de @xmath850 .
alors lquation diffrentielle @xmath1518 a une solution @xmath1519)$ ] , qui dtermine une famille libre @xmath1520 de @xmath850 , telle que @xmath1521 .
la proposition [ kernabla ] montre que @xmath1 est de de rham , et comme @xmath1522 , cest que ses poids de hodge - tate sont ngatifs .
lobjet de ce paragraphe est de dmontrer le rsultat suivant , connu sous le nom de `` conjecture de monodromie @xmath0-adique '' .
cette conjecture , formule par fontaine @xcite , est lanalogue pour les reprsentations @xmath0-adiques du thorme de monodromie @xmath1523-adique de grothendieck . [ monofont ] si @xmath1 est une reprsentation @xmath0-adique de de rham , alors @xmath1 est potentiellement semi - stable .
les deux ingrdients principaux pour la dmonstration de ce thorme sont dune part la construction de @xmath93 , et dautre part un profond thorme dy .
andr sur la structure des quations diffrentielles @xmath0-adiques .
avant de dmontrer le thorme [ monofont ] , faisons quelques brefs rappels sur les quations diffrentielles @xmath0-adiques , afin de fixer les notations et dnoncer les rsultats de @xcite . supposons dans la suite que @xmath13 est algbriquement clos .
soit @xmath1524 lanneau de robba coefficients dans @xmath14 .
cet anneau est muni dun oprateur de frobenius @xmath29 et dune drivation @xmath1083 , qui proviennent tous les deux doprateurs dfinis sur @xmath1525 .
de plus , @xmath1525 est un anneau de sries formelles en @xmath974 , et @xmath1526 .
parmi les extensions finies de lanneau de robba , on sintresse celles qui proviennent dune extension finie sparable de @xmath958 , cest - - dire celles de la forme @xmath1527 .
elles sont canoniquement munies du frobenius de @xmath50 , et de la drivation qui tend @xmath1083 . via la thorie
du corps de normes , on voit que ces extensions sont celles qui proviennent dune extension finie sparable de @xmath1528 .
un @xmath29-module @xmath850 sur @xmath1529 est un @xmath1529-module libre de rang fini , muni dune application @xmath1530 semi - linaire , telle que @xmath1531 ( cest - - dire que le @xmath1529-module engendr par @xmath1532 est gal @xmath850 ) .
une quation diffrentielle @xmath0-adique est un @xmath1529-module de prsentation finie muni dune connexion @xmath1533 , ou , ce qui revient au mme car @xmath1534 , dune application @xmath1535 , qui tend la drivation @xmath1536 donne sur @xmath1537 par @xmath1538 , et qui satisfait la rgle de leibniz .
une quation diffrentielle @xmath0-adique @xmath850 est dite tre munie dune structure de frobenius , si @xmath850 est un
@xmath29-module et si le diagramme suivant est commutatif : @xmath1539 tant donn que sur @xmath1537 , on a @xmath1540 cela revient demander que @xmath1541 . les quations diffrentielles @xmath0-adiques , munies ou non dune structure de frobenius , ont t tudies en dtail par christol et mebkhout ; nous renvoyons @xcite pour un survol des rsultats et de quelques applications .
en utilisant ces rsultats , andr a russi dmontrer @xcite le thorme [ monoandre ] ci - dessous , lui aussi connu sous le nom de `` conjecture de monodromie @xmath0-adique '' .
que kedlaya a plus rcemment obtenu une dmonstration de ce mme rsultat dans @xcite , par des mthodes trs diffrentes ( plus proches de lesprit du prsent article ) .
enfin mebkhout a aussi anonc une dmonstration @xcite .
[ monoandre ] tout module diffrentiel @xmath850 de prsentation finie sur @xmath1529 admettant une structure de frobenius possde une base de solutions dans @xmath1542 $ ] o @xmath1543 est lextension finie tale de @xmath1529 issue dune extension finie sparable convenable de @xmath1544 , cest - - dire que @xmath850 possde une base de solutions dans @xmath1545 $ ] , pour une extension finie @xmath1546 de @xmath15 .
on est maintenant en mesure de dmontrer le thorme [ monofont ] .
soit @xmath1 une reprsentation @xmath0-adique de @xmath23 qui est de de rham .
on cherche montrer que @xmath1 est potentiellement semi - stable , et on peut supposer que 1 . les poids de hodge - tate de @xmath1 sont ngatifs , en effet @xmath1 est potentiellement semi - stable si et seulement si @xmath1547 lest pour @xmath930
@xmath13 est algbriquement clos , car @xmath1 est de rham ou potentiellement semi - stable si et seulement si sa restriction
linertie lest ( voir @xcite pour une dmonstration de cela dans le cas semi - stable .
dans les autres cas , la dmonstration est compltement identique ) . on peut alors
associer @xmath1 le module @xmath93 , qui est une quation diffrentielle @xmath0-adique munie dune structure de frobenius .
le seul point vrifier est que @xmath1548 .
mais @xmath1549 est un sous - module libre de rang @xmath22 de @xmath55 qui est stable par @xmath94 : par le thorme [ ndr ] , un tel objet est ncessairement gal @xmath93 .
par le thorme dandr , et la remarque qui le suit , il existe une extension finie @xmath1550 telle que @xmath1551 est un @xmath1552-espace vectoriel de dimension @xmath22 .
comme lalgbre de lie de @xmath1553 agit trivialement sur celui - ci , il existe un sous groupe ouvert de @xmath1553 qui agit trvialement dessus ( * ? ? ?
* chap v ) , et donc quitte remplacer @xmath1546 par @xmath1554 , avec @xmath1555 , on peut supposer que @xmath1553 agit trivialement sur @xmath1551 .
comme @xmath1556 , o @xmath1557 est @xmath1 vue comme reprsentation de @xmath1558 , on en dduit que @xmath1559 est un @xmath1552-espace vectoriel de dimension @xmath22 .
par le thorme [ isomcomp ] , @xmath1557 est semi - stable .
ceci implique que @xmath1 est potentiellement semi - stable .
pour illustrer les mthodes que nous avons dveloppes , nous donnons dans ce paragraphe une nouvelle dmonstration dun thorme de sen sur les reprsentations @xmath18-admissibles : ce sont celles do nt la restriction linertie est potentiellement triviale .
tout dabord , la thorie de sen permet de caractriser facilement les reprsentations @xmath18-admissibles en termes de loprateur @xmath1560 : si @xmath1 est une reprsentation @xmath0-adique , alors @xmath1 est @xmath18-admissible si et seulement si @xmath1561 . si @xmath1 est @xmath18-admissible , alors on a @xmath1562 , il contient le @xmath96-espace vectoriel @xmath1563 , qui est stable par @xmath30 et de dimension @xmath22 : cest donc @xmath1345 .
laction de @xmath83 est bien triviale sur cet espace .
rciproquement , si @xmath1345 est muni de la connexion triviale , alors laction de @xmath30 est discrte sur @xmath1564 , et par hilbert 90 , @xmath1 est @xmath18-admissible .
la proposition suivante est due sen @xcite et est un cas particulier de rsultats assez gnraux sur la caractrisation de lalgbre de lie de limage de @xmath1479 pour une reprsentation @xmath0-adique en terme de loprateur @xmath1560 : [ sen ] soit @xmath1 une reprsentation @xmath0-adique @xmath18-admissible .
alors la restriction de @xmath1 @xmath1479 est potentiellement triviale .
le but de ce paragraphe est de donner une dmonstration de cette proposition qui repose sur la thorie des quations diffrentielles @xmath0-adiques .
comme on ne sintresse qu la restriction de @xmath1 @xmath1479 , on suppose dans tout ce paragraphe que @xmath13 est algbriquement clos .
soient @xmath1565 ce qui fait que @xmath1566 o @xmath1567 .
nous aurons besoin du rsultat suivant de tsuzuki @xcite : [ tsutheo ] si @xmath13 est algbriquement clos , et si on se donne deux matrices @xmath1568 et @xmath1569 , telles que @xmath1570 , telles que le @xmath99-module correspondant @xmath850 est tale , alors il existe une extension finie @xmath1550 , et une matrice @xmath1571 , telle que @xmath1572 et @xmath1573 .
en dautres termes , les @xmath14-isocristaux de pente nulle sont potentiellement triviaux . on utilise ce thorme pour montrer le rsultat suivant : [ tsutsu ] soit @xmath1574 dfini par @xmath1575 .
si @xmath1 est une reprsentation @xmath0-adique de @xmath23 et @xmath1576 est un oprateur diffrentiel au - dessus de @xmath1577 tel que @xmath1578 .
alors le sous - groupe dinertie de @xmath113 agit travers un quotient fini sur @xmath1 .
soient @xmath301 la matrice de @xmath29 et @xmath1579 .
la relation de commutation entre @xmath29 et @xmath94 se traduit par @xmath1580 .
on peut donc appliquer le thorme de tsuzuki , et en dduire quil existe @xmath1550 finie telle que @xmath1581 est trivial : la matrice @xmath1582 nous donne une base @xmath1583 de @xmath1581 , telle que @xmath1584 ce qui fait que la restriction de @xmath1 @xmath1585 est triviale et donc que la restriction de @xmath1 ( comme reprsentation de @xmath23 ) linertie de @xmath113 est potentiellement triviale . [ controle ] si la connexion @xmath83 de @xmath55 vrifie @xmath1586 , alors loprateur @xmath1587 vrifie @xmath1588 .
soit @xmath1589 une base de @xmath896 sur @xmath42 , telle que la matrice de @xmath29 vrifie @xmath1590 .
cest aussi une base de @xmath55 sur @xmath50 .
soit @xmath1591 .
on remarquera que @xmath1592 puisque @xmath1083 prserve @xmath42 .
le fait que @xmath83 et @xmath29 commutent montre que @xmath1593 .
soit @xmath1594 .
lquation prcedente scrit alors @xmath1595 et @xmath1596 .
le rsultat que lon cherche tablir est que @xmath1597 et donc @xmath47 a ses coefficients dans @xmath42 .
il suffit donc de montrer le fait suivant : si @xmath1598 vrifie @xmath1599 , alors en fait @xmath1600 , ce que nous faisons maintenant .
soient @xmath1601 les normes correspondant aux valuations @xmath385 ; on les tend @xmath1602 en dcidant que @xmath1603 . un lment @xmath476 de @xmath12 est dans @xmath267 si et seulement si la suite des @xmath1604 ( avec @xmath1605 ) est borne , et il en est donc de mme pour les matrices coefficients dans ces anneaux .
comme @xmath1606 , on a pour toute matrice @xmath850 , @xmath1607 .
fixons @xmath292 tel que @xmath1217 et @xmath1608 , et @xmath1609 .
soit @xmath1610 $ ] .
dans notre cas @xmath1611 et il existe une constante @xmath1213 telle que pour tout @xmath309 , on ait : @xmath1612 .
comme on a @xmath1613 , alors pour @xmath1614 : @xmath1615 ce qui fait que @xmath1616 , pour tout @xmath809 , et donc que @xmath1617 .
[ conntriv ] la connexion @xmath1560 sur @xmath1345 est triviale si et seulement si @xmath1618 .
si @xmath1618 , alors comme @xmath1359 est limage par @xmath1619 de @xmath1620 , cest que @xmath1561 .
rciproquement si @xmath1561 , alors pour tout @xmath556 on a @xmath1621 par le lemme [ kerthetadhol ] , et le lemme [ divtv ] permet de conclure .
comme on sintresse la restriction de @xmath1 linertie , on peut supposer que le corps rsiduel @xmath13 de @xmath15 est algbriquement clos , ce que lon fait maintenant . on a vu quune reprsentation @xmath1 est @xmath18-admissible si et seulement si le module @xmath1345 qui lui est associ par la thorie de sen est muni de la connexion triviale et on a montr que cette connexion est limage par @xmath465 de celle qui existe sur @xmath1353 .
le lemme [ conntriv ] montre que dans ce cas @xmath1618 .
le lemme [ controle ] montre alors que @xmath1622 est une connexion surconvergente sur @xmath896 .
le lemme [ tsutsu ] montre ensuite quil existe une extension finie @xmath1550 telle que laction de @xmath1585 sur @xmath1 est triviale sur @xmath1 .
comme @xmath1623 , laction de @xmath1553 est finie , ce qui fait quil existe @xmath296 tel que @xmath1624 agit trivialement sur @xmath1 , et donc que la restriction de @xmath1 @xmath1479 est potentiellement triviale .
une reprsentation @xmath18-admissible @xmath1 est de de rham , et on a @xmath98 dans ce cas particulier .
[ remasentsu ] en fait , le thorme de sen implique le thorme de tsuzuki ( et lui est donc quivalent ) .
indiquons brivement de quoi il sagit : si @xmath66 est un isocristal de pente nulle , on peut `` exponentier '' la connexion pour construire une matrice que lon utilise pour faire agir un sous - groupe ouvert de @xmath30 . le fait que @xmath29 est de pente nulle assure que cette matrice est surconvergente . on a donc construit
un @xmath25-module tale , cest - - dire une reprsentation @xmath0-adique @xmath1 .
il nest pas dur de montrer que @xmath1 est @xmath18-admissible , et den conclure , via le thorme de sen , que @xmath1 , et donc @xmath66 , devient trivial aprs une extension finie .
dans ce chapitre , on utilise le thorme de monodromie @xmath0-adique , pour montrer que si @xmath1625 est une reprsentation de de rham qui est une extension de deux reprsentations semi - stables , alors @xmath1625 est semi - stable . on va montrer que : 1 .
toute reprsentation ordinaire de @xmath23 est semi - stable ; 2 .
une extension de deux reprsentations semi - stables qui est de de rham est elle - mme semi - stable ; 3 .
si @xmath1 est une reprsentation semi - stable do nt les poids de hodge - tate sont tous @xmath8 , alors @xmath100 est un isomorphisme .
ces trois rsultats taient connus dans le cas dun corps rsiduel fini , o on peut les dduire de calculs de dimension de groupes de cohomologie galoisienne ( ce qui nest plus possible si le corps rsiduel nest pas fini ) . le ( 1 ) avait t dmontr dans ce cas l par perrin - riou @xcite comme corollaire des calculs de bloch et kato @xcite , le ( 2 ) par hyodo @xcite , et le ( 3 ) par bloch et kato . dans une version antrieure de cet article , on utilisait le calcul de la cohomologie galoisienne via les @xmath25-modules pour dmontrer ces rsultats .
lobjet de ce paragraphe est de montrer le thorme suivant : [ extsst ] si @xmath1625 est une reprsentation de de rham , qui est une extension de @xmath954 par @xmath1 , o @xmath1 et @xmath954 sont semi - stables , alors @xmath1625 est semi - stable .
et son corollaire : [ ordisst ] toute reprsentation ordinaire de @xmath23 est semi - stable .
commenons par montrer la proposition [ extsst ] ; par le thorme [ monofont ] , il suffit de montrer le lemme suivant : si @xmath1625 est une reprsentation potentiellement semi - stable de @xmath23 , qui est une extension de @xmath954 par @xmath1 , o @xmath1 et @xmath954 sont deux reprsentations semi - stables de @xmath23 , alors @xmath1625 est semi - stable .
soit @xmath1546 une extension finie
de @xmath15 , telle que la restriction de @xmath1625 @xmath1558 est semi - stable .
soit @xmath1626 . on a une suite exacte : @xmath1627 et en prenant les invariants par @xmath1628 , on en dduit : @xmath1629
pour des raisons de dimension , il suffit de montrer que @xmath1630 ce qui rsulte du fait que @xmath1631 est un @xmath21-espace vectoriel , et donc en particulier que @xmath1632 y est inversible .
pour montrer le corollaire [ ordisst ] , on va tout dabord montrer le lemme suivant ; si @xmath1 est une reprsentation de de rham , @xmath1633 est lensemble des classes dextensions qui sont de de rham .
[ posdr ] si @xmath1 est une reprsentation de de rham do nt les poids de hodge - tate sont tous @xmath1634 , alors @xmath1635 .
soit @xmath1625 une extension de @xmath21 par @xmath1 : il suffit de montrer que @xmath1625 est de de rham .
lhypothse sur les poids de hodge - tate de @xmath1 implique que @xmath1636 ( ceci suit de la dcomposition de hodge - tate de @xmath1637 , du thorme de tate : @xmath1638 si @xmath1639 , et des suites exactes @xmath1640 , que lon utilise pour @xmath331 ) . on a donc
un morceau de suite exacte @xmath1641 , ce qui fait que @xmath1642 se relve en un lment de @xmath1643 , qui nest pas dans @xmath1644 . on en dduit immdiatement
que la dimension de @xmath1643 est la bonne , et donc que @xmath1625 est de de rham .
rappelons @xcite quune reprsentation @xmath1 est ordinaire , sil existe une filtration @xmath1645 de @xmath1 , dcroissante exhaustive et spare , par des sous - espaces @xmath1645 stables par @xmath23 , et telle que le groupe dinertie @xmath1479 agit sur @xmath1646 par @xmath852 .
une rcurrence immdiate sur la longueur de la filtration montre que quune reprsentation ordinaire est de de rham , et comme elle est extension successive de reprsentations semi - stables ( par hypothse de rcurrence ) , elle est elle - mme semi - stable .
dans ce paragraphe , @xmath1 est une reprsentation semi - stable .
rappelons que les anneaux @xmath190 et @xmath171 sont relis par la suite exacte fondamentale ( cf ( * ? ? ?
* iii.5 ) et @xcite ) : @xmath1647 en tensorisant avec @xmath1 et en prenant les invariants par @xmath23 on obtient un dbut de suite exacte longue : @xmath1648 soit @xmath1649 . on dduit
de la suite exacte ci - dessus une application de @xmath58 dans @xmath1650 appele exponentielle de bloch - kato , et note @xmath9 .
dautre part comme @xmath1 est de de rham on a @xmath1651 et ( * ? ?
* lemma 3.8.1 ) limage de @xmath9 est @xmath1650 tout entier .
dans le cas o @xmath13 est fini des calculs de dimension montrent que si @xmath1652 , alors @xmath1653 et donc que @xmath1654 est un isomorphisme ( on remarquera que si @xmath1652 , alors @xmath1655 ) .
le but de ce paragraphe est de montrer , sans condition sur @xmath13 , que si tous les poids de hodge - tate de @xmath1 sont @xmath8 , alors @xmath1656 est un isomorphisme .
si @xmath1 est une reprsentation semi - stable telle que @xmath1657 est surjectif et @xmath1658 , alors @xmath1659 .
soit @xmath1660 et @xmath1625 lextension de @xmath21 par @xmath1 qui correspond @xmath640 .
le thorme [ extsst ] montre quil existe @xmath1661 tel que @xmath1662 .
comme @xmath1663 on a @xmath1664 et donc @xmath1665 .
il existe donc @xmath1666 tel que @xmath1667 ce qui revient @xmath1668 et quitte remplacer @xmath476 par @xmath1669 on peut donc supposer que @xmath1670 .
comme @xmath66 commute @xmath23 on a @xmath1671 et donc @xmath1672
. de plus comme @xmath1670 on doit avoir @xmath1673 et donc @xmath1674 ce qui montre que @xmath1675 . si @xmath1 a tous ses poids de hodge - tate @xmath8 , alors @xmath1 satisfait les conditions du lemme prcdent .
rappelons que le frobenius @xmath29 sur @xmath54 est bijectif . de plus , si @xmath954 est une reprsentation semi - stable do nt les poids de hodge - tate sont positifs , alors il existe un rseau @xmath1676 de @xmath1677 tel que @xmath1678 ( cest le dual de lassertion que si @xmath1679 est semi - stable positive , il existe un rseau de @xmath1680 stable par @xmath29 ) . en tordant , on voit
donc quil existe un rseau @xmath850 de @xmath54 , tel que @xmath1681 .
il est alors vident que @xmath1682 est surjectif , et que @xmath1658 .
[ expbk ] si @xmath1 est une reprsentation semi - stable de @xmath23 , do nt les poids de hodge - tate sont @xmath8 , alors lexponentielle de bloch - kato @xmath100 est un isomorphisme .
les deux lemmes prcdents , et le lemme [ posdr ] , montrent que lon a @xmath1683 et @xmath9 est donc surjective .
dautre part , on a @xmath1655 , ce qui montre que @xmath9 est injective .
cest donc un isomorphisme .
le diagramme ci - dessous rcapitule les relations entre les diffrents anneaux de priodes .
les flches qui se terminent par @xmath1684&}$ ] sont surjectives , la flche en pointills @xmath1685&}$ ] est une limite inductive de morphismes dfinis sur des sous - anneaux ( @xmath1686 ) , et toutes les autres sont injectives .
@xmath1687^{\theta } & \\ & { \widetilde{\mathbf{b}}^{\dagger } _ { \mathrm{log } } } \ar@{.>}[urrr ] & { \widetilde{\mathbf{b}}^{+}_{\mathrm{log } } } \ar[rr ] \ar[l ] & & { \mathbf{b}_{\mathrm{st}}}^+ \ar[u ] & \\ & { \widetilde{\mathbf{b}}^{\dagger } _ { \mathrm{rig } } } \ar[u ] & { \widetilde{\mathbf{b}}^{+}_{\mathrm{rig } } } \ar[u ] \ar[l ] \ar[rr ] & & { \mathbf{b}_{\mathrm{max}}}^+ \ar[u ] & \\ { \widetilde{\mathbf{b } } } & { \widetilde{\mathbf{b}}^{\dagger } } \ar[l ] \ar[u ] & { \widetilde{\mathbf{b}}^+}\ar[l ] \ar[u ] \ar@{->>}[r]^{\theta } & { \mathbf{c}_p } & & \\ { \widetilde{\mathbf{a}}}\ar[u ] \ar@{->>}[d ] & { \widetilde{\mathbf{a}}^{\dagger } } \ar[l ] \ar[u ] & { \widetilde{\mathbf{a}}^+}\ar[l ] \ar[u ] \ar@{->>}[d ] \ar@{->>}[r]^{\theta } & { \mathcal{o}}_{{\mathbf{c}_p } } \ar[u ] \ar@{->>}[d ] & & \\ { \widetilde{\mathbf{e } } } & & { \widetilde{\mathbf{e}}^+}\ar[ll ] \ar@{->>}[r]^{\theta } & \overline{k } & & \\ } \ ] ] tous les anneaux qui ont des tildes ( @xmath1688 ) ont aussi des versions sans tilde : on passe de cette dernire la version avec tilde en rendant frobenius inversible , et en compltant .
par exemple , @xmath108 est le complt de la perfection de @xmath152 .
les trois anneaux de la colonne de gauche ( du moins leurs versions sans tildes ) sont relis la thorie des @xmath25-modules .
les trois anneaux de la colonne de droite sont relis la thorie de hodge @xmath0-adique .
pour faire le lien entre ces deux thories , on passe dun ct un autre via tous les anneaux intermdiaires .
la situation idale est quand on peut se placer dans la colonne du milieu ( par exemple , de haut en bas : les reprsentations semi - stables , cristallines , ou de hauteur finie ) .
rcapitulons les diffrents anneaux de sries formelles qui interviennent ; soit @xmath1689 la couronne @xmath1690 . on a alors : @xmath1691 } \\ { \mathbf{a}_f^+ } & = { { \mathcal{o}}_f[[t ] ] } \\ { \mathbf{b}_f^+ } & = { f \otimes_{{\mathcal{o}}_f } { \mathcal{o}}_f[[t ] ] } \\ { } & { } \\ { \mathbf{e}_f } & = { k((t ) ) } \\ { \mathbf{a}_f } & = { { \widehat}{{\mathcal{o}}_f[[t]][t^{-1 } ] } } \\ { \mathbf{b}_f } & = { f \otimes_{{\mathcal{o}}_f } { \widehat}{{\mathcal{o}}_f[[t]][t^{-1 } ] } } \\ { } & { } \\ { { \mathbf{a}^{\dagger , r}}_f } & = { \{\text{s{\'e}ries de laurent $ f(t)$ , qui convergent sur $ c[r;1[$ , } } \\ & { \text{et y sont born{\'e}es par $ 1$}\ } } \\ { { \mathbf{b}^{\dagger , r}}_f } & = { \ { \text{s{\'e}ries de laurent $ f(t)$ , qui convergent sur $ c[r;1[$ , } } \\ & { \text{et y sont born{\'e}es}\ } } \\ { } & { } \\ { { \mathbf{b}^{\dagger , r}_{\mathrm{rig } , f } } } & = { \{\text{s{\'e}ries de laurent $ f(t)$ , qui convergent sur $ c[r;1[$}\ } } \\ { { \mathbf{b}^{\dagger , r}_{\mathrm{log } , f } } } & = { { \mathbf{b}^{\dagger , r}_{\mathrm{rig } , f}}[\log(t ) ] } \\ { } & { } \\ { { \mathbf{b}^+_{\mathrm{rig } , f } } } & = { \ { f(t ) \in
f[[t]],\ \text{$f(t)$ converge sur le disque ouvert $ d[0;1[$ } \ } } \\ { { \mathbf{b}^+_{\mathrm{log } , f } } } & = { { \mathbf{b}^+_{\mathrm{rig } , f}}[\log(t)]}\end{aligned}\ ] ] kato k. : _ lectures on the approach to iwasawa theory for hasse - weil @xmath1546-functions via @xmath171 . _ arithmetic algebraic geometry , lecture notes in mathematics 1553 , springer - verlag , berlin , 1993 , 5063 .
serre j - p .
abelian @xmath1523-adic representations and elliptic curves .
_ with the collaboration of willem kuyk and john labute .
second edition . advanced book classics .
addison - wesley publishing company , advanced book program , redwood city , ca , 1989 .
xxiv+184 pp .
|
in this paper , we associate to every @xmath0-adic representation @xmath1 a @xmath0-adic differential equation @xmath2 , that is to say a module with a connection over the robba ring .
we do this via the theory of fontaine s @xmath3-modules .
this construction enables us to relate the theory of @xmath3-modules to @xmath0-adic hodge theory .
we explain how to construct @xmath4 and @xmath5 from @xmath2 , which allows us to recognize semi - stable or crystalline representations ; the connection is then either unipotent or trivial on @xmath6 $ ] .
in general , the connection has an infinite number of regular singularities , but we show that @xmath1 is de rham if and only if those are apparent singularities .
a structure theorem for modules over the robba ring allows us to get rid of all singularities at once , and to obtain a `` classical '' differential equation , with a frobenius structure .
a recent theorem of y. andr gives a complete description of the structure of such an object .
this allows us to prove fontaine s @xmath0-adic monodromy conjecture : every de rham representation is potentially semi - stable . as an application
, we can extend to the case of arbitrary perfect residue fields some results of hyodo ( @xmath7 ) , of perrin - riou ( the semi - stability of ordinary representations ) , of colmez ( absolutely crystalline representations are of finite height ) , and of bloch and kato ( if the weights of @xmath1 are @xmath8 , then bloch - kato s exponential @xmath9 is an isomorphism ) .
|
the conversion of gas into stars is one of the most fundamental processes in astronomy . yet , despite decades of effort , a simple prescription of star formation ( sf ) that successfully describes all observations of galaxies across a range of halo masses has remained elusive . in broad terms , more massive star - forming galaxies will have larger gas reservoirs ( both atomic and molecular ) and higher global star formation rates ( sfr ) than less massive systems ( see , e.g. , * ? ? ?
however , the gas mass fractions in star forming galaxies tend to increase with decreasing mass ( e.g. , * ? ? ?
* ) . empirical correlations between gas properties and various tracers of instantaneous (
using h@xmath2 emission , with a characteristic timescale of @xmath8 10 myr ) or ongoing ( using fuv emission , with a characteristic timescale of @xmath8 100 myr ) sf have been numerous in the literature .
the most common parameterization relates a sfr surface density to a gas surface density : @xmath9 with the sfr surface density ( @xmath10 ) in units of m@xmath11yr@xmath1kpc@xmath12 , the gas surface density ( @xmath13 ) in units of m@xmath11pc@xmath12 , and @xmath5 being a positive number .
@xcite found that @xmath14 , and similar indices have been derived numerous times over the last half - century ( see elmegreen 2011 for a recent review ) .
for example , @xcite found @xmath15 [email protected] for 61 large spiral galaxies when relating the h@xmath2-based sfr to the total gas surface density ( via both and co observations , where the co is used as a tracer for molecular gas ) . in a study of this relation on sub - kpc scales
, @xcite found that the relationship between the total gas surface density and the sfr surface density varied dramatically among and within individual spiral galaxies , and that most of the sample showed little or no correlation between @xmath4 and @xmath3 . in an associated paper , @xcite found a molecular schmidt power - law slope of @xmath16 in 18 nearby spiral galaxies ; similarly , @xcite found @xmath15 1.3 - 1.8 for 10 nearby spiral galaxies . in a subsequent study ,
@xcite find no clear evidence for sf thresholds and emphasize that it may not be realistic to expect them .
molecular gas in low mass galaxies remains largely undetectable via traditional co tracers , thus studying the relationship between atomic gas and sf is especially important in these metal - poor low - mass systems ( e.g. , * ? ? ?
in fact , very few detections of co gas exist at metallicities less than @xmath010% z@xmath11 , even in systems that are actively forming stars ( see , e.g. , taylor etal .
1998 , schruba etal .
2012 , warren etal .
2015 ) . for low mass galaxies
, this implies that studying the relationship between @xmath10 and @xmath17 as a function of galaxy mass is necessarily constrained to only the atomic gas component .
it is thus interesting to note that a correlation between @xmath18 and @xmath10 appears to hold in some galaxies and not in others .
for example , @xcite find that @xmath18 and @xmath10 are not related in individual spiral disks .
this can be compared to the results in @xcite , which show that 10@xmath19 cm@xmath12 ( corresponding to 7.9 m@xmath11 pc@xmath12 or 10.6 m@xmath11 pc@xmath12 when accounting for helium ) represents a requisite threshold surface density for massive sf .
similarly , @xcite examined 19 low surface - brightness galaxies and found an apparent threshold in the gas surface density in the range 3 @xmath21 10 m@xmath11 pc@xmath12 below which very little sf ( traced by h@xmath2 ) is observed . extending to even lower masses , @xcite and @xcite
find that all gas in their galaxies with @xmath4 @xmath22 10 m@xmath11 pc@xmath12 ( @xmath6 1.2@xmath2310@xmath19 cm@xmath12 ) have associated sf , but there is no threshold below which sf is not observed ( that is , sf is observed in regions with columns @xmath24
10@xmath19 cm@xmath12 ) .
most recently , @xcite found that the @xmath18 @xmath21 @xmath10 relation for a set of dwarf irregular galaxies was nearly linear .
@xcite find consistency across a range of scales ( 400pc and 1kpc ) and galaxy types , including both low - mass galaxies and more massive spiral disks .
based on the above results , it is not trivial to anticipate where in a galaxy one will observe ongoing sf - regardless of how massive that galaxy is .
knowledge of the properties alone is often insufficient to predict where in a given galaxy the conditions are ripe for sf .
for example , @xcite suggests that molecular gas is a better predictor of sf than the neutral ism .
we examine this issue from the perspective in section [ s4.2 ] .
in addition to the @xmath17 vs. @xmath10 analysis , we also study the star formation efficiency ( sfe ) , which is a useful metric when discussing where in a galaxy sf is occurring @xcite .
several ways of describing the sfe exist , but for consistency with other surveys we use sfe = @xmath25 with units of yr@xmath1 .
the sfe is more useful than sfr alone to identify where conditions are conducive to sf because it is normalized by the gas mass surface density .
thus , it quantifies the local physical properties in regions where the gas is being turned into stars efficiently : regions of elevated gas surface density which have no young stars associated with them are inefficient , while regions of elevated gas surface density that show co - spatiality with young stars are efficient . conveniently , the inverse of the sfe is the gas depletion time , which is the time required for sf to consume the gas reservoir at the present - day sfr . for a sample of low - mass galaxies ,
@xcite found gas depletion timescales of @xmath010@xmath26 yr , an order of magnitude lower than is estimated for the outer regions of large spiral galaxies @xcite .
we calculate gas consumption times for our sample of galaxies in section [ s4.2 ] .
the alfalfa survey @xcite has produced one of the largest and most statistically useful catalogs of nearby , gas - rich galaxies to date .
the final alfalfa database will include source parameters for more than 30,000 systems .
with the acquisition of data for alfalfa now complete , a unique database exists to facilitate the study of fundamental galaxy properties across an unprecedented range of physical parameters .
one particularly rich area of exploration that has been enabled by alfalfa is a robustly - populated faint end of the mass function @xcite .
specifically , it is possible to identify a complete sample of gas - bearing , low - mass galaxies by matching the alfalfa database to existing optical survey data .
as introduced in @xcite , the `` survey of in extremely low - mass dwarfs '' ( hereafter , shield ) is a multi - wavelength , detailed study of the properties of alfalfa - discovered or cataloged systems with extremely small mass reservoirs ( see section [ s2 ] for detailed discussion of the sample selection ) .
subsequent works have established the distances @xcite , the nebular abundances @xcite , and the qualities of sf based on spatially resolved hubble space telescope ( hst ) imaging @xcite .
the primary goals of shield are to 1 ) characterize the nature of sf in very low - mass galaxies and to 2 ) determine what fraction of the mass in these low - mass galaxies is baryonic . in this work
, we undertake a comparative study of the gas and various tracers of recent sf in order to address goal # 1 .
a companion paper by @xcite explores the dynamical properties of our sample galaxies in order to address goal # 2 .
shield is one of a number of recent surveys of dwarf galaxies using interferometric data .
this list includes whisp ( the westerbork survey of irregular and spiral galaxies ; swaters 2002 ) , figgs ( faint irregular galaxies gmrt survey ; begum etal .
2008 ) , vla - angst ( very large array survey of acs nearby galaxy survey treasury galaxies ; ott etal .
2012 ) , little - things ( local irregulars that trace luminosity extremes in the nearby galaxy survey ; hunter etal .
2012 ) , and lvhis ( the local volume survey ; kirby 2012 ) .
these studies have yielded valuable insights into a total of nearly two dozen systems with m@xmath27 @xmath8 @xmath28 m@xmath11 .
shield adds to this relatively understudied region of parameter space by significantly increasing the number of sources with resolved imaging .
shield is a multi - wavelength survey of 12 low - mass galaxies in the local volume .
sample members were selected from the first @xmath010% of the alfalfa - detected galaxies on the basis of estimated mass ( m@xmath27 @xmath24 10@xmath29 , based on flow model distances using the prescription of masters 2005 ) . the w@xmath30 condition ( line width at 50% of peak @xmath24 65 kms@xmath1 , with no correction for inclination which would increase rotational velocity ) discriminates against massive but -poor galaxies and identifies truly low - mass galaxies . following the presentation of early shield results in @xcite ,
the sources were observed with hst to derive their distances via the tip of the red giant branch ( trgb ) method .
the details are given in @xcite ; all sources moved to somewhat higher distances than the flow model predictions and so the masses of the systems are slightly increased .
even with the updated distances , all but one of the galaxies has m@xmath27 @xmath24 10@xmath31 .
the median distance , mass , and line width for the sample are 7.86 mpc , 10@xmath32 m@xmath11 , and 25 kms@xmath1 , respectively .
table [ t1 ] provides a summary of physical characteristics of the shield galaxies .
the data for the survey were obtained using the karl g. janksy vla in multiple array configurations for programs vla/10b-187 ( legacy code ac990 ; p.i .
cannon ) and vla/13a-027 ( legacy code ac1115 ; p.i .
our observational strategy ( 9 , 4 , and 2 hours of observation per source in the b , c , and d arrays of the vla , respectively , with typical calibration overheads of 25% ) achieves high spatial ( @xmath06 synthesized beam at full resolution ) and spectral resolution ( @xmath00.824 kms@xmath1 ch@xmath1 ) , while retaining sensitivity to extended structure .
the widar correlator is used to provide a single 1 mhz sub - band with 2 polarization products and 256 channels each , covering 211 kms@xmath1 of frequency space at 3.906 khz ch@xmath1 , which is the setup for the b- and c - configuration observations . for the d - configuration observations ,
the frequency coverage was widened to 4 mhz ( corresponding to 1024 channels and 844 kms@xmath1 at 3.906 khz ch@xmath1 ) .
vla data acquisition for shield began in 2010 october , and was completed in 2013 .
all of the sample members were observed in the b - configuration except for agc111164 , agc111977 , and agc112521 .
the vla data reduction techniques are standard and were done using the common astronomy software application ( casa ; mcmullin etal .
radio frequency interference was excised by hand .
calibrations were then derived for antenna position , antenna - based phase delays , atmospheric opacity , and bandpass frequency response . after continuum subtraction in the @xmath33-plane
, the b , c , and d - configuration measurement sets were gridded together using casa s clean algorithm to produce data cubes .
we generated a boolean mask based on the dirty cube for use in our deep cleans ( threshold set at 2.5@xmath34 ) .
we also performed residual flux rescaling on our data cubes .
deconvolution of the dirty map induces a difference area between the ` dirty beam ' of the residual map and the ` clean beam ' of the deconvolved map and thus the resultant flux density in the final ` combined ' image ; in order to increase the accuracy of the flux scaling in the final image ( which is the linear sum of the residual map and clean map ) the residual must be rescaling by some factor proportional to the difference in the ( frequency dependent ) ` dirty ' and ` clean beam ' areas .
this is a higher - order correction , as the error on the flux scaling of these observations is primarily limited by the calibration model uncertainties .
the details and motivation for this procedure are discussed in @xcite . following these corrections
, we then implemented a correction for the primary beam attenuation .
the robust data parameters are summarized in table [ t2 ] . to produce two - dimensional images from the three - dimensional cubes
, we implemented a threshold mask followed by a manual inspection and hand - blanking of the cubes .
our final data products include 4 cubes and 4 moment-0 ( integrated intensity ) maps for each of the 12 galaxies .
the cubes are either natural - weighted or robust - weighted ( explicitly , weighting@xmath35``briggs '' and [email protected] ) , and at either native spectral resolution ( 0.82 kms@xmath1 ch@xmath1 ) or spectrally smoothed by a factor of 3 ( 2.46 kms@xmath1 ch@xmath1 ) .
we estimate the uncertainty on our flux densities to be a minimum of 10% , and propagate this throughout the calculation of masses , column densities , and @xmath17 .
this estimate accounts for random noise in our maps as well as errors in flux calibration .
the flux densities of the maps are converted to masses using the standard transformation for optically thin emission : @xmath36 where m@xmath27 is the mass in m@xmath11 , d is the distance in mpc , and the integral @xmath37@xmath38 is the line flux of the source in jy km s@xmath1 @xcite .
one of the shield sources , agc111164 , presented unique challenges in the data processing .
agc111164 is located @xmath015 east of the more massive , gas - rich galaxy ngc784 .
both sources overlap in velocity space ; ngc784 lies almost exactly one half - power beam - width away from agc111164 , and thus the sidelobes from ngc784 are co - spatial with the shield source and are challenging to clean properly .
we used the _ outlierfile _
parameter for casa s clean task , which proved to be helpful in extracting the larger source from our field .
we also shifted the phasecenter of the clean to lie directly between the two sources .
galex is a 50 cm diameter uv telescope that images the sky simultaneously in both a far ultraviolet ( fuv ) and a near ultraviolet ( nuv ) band , with effective wavelengths of 1528 and 2271 , respectively @xcite .
the field of view of galex is approximately circular with a diameter of 1.25 degrees , with an intrinsic angular resolution of 4.2 and 5.3 fwhm in the fuv and nuv , respectively .
the ultraviolet data presented here are derived from three galex programs : the guest investigator program ( gi ) , the medium imaging survey ( mis ) , and the all - sky imaging survey ( ais ) .
all sources had exposure times between @xmath01600 and @xmath02800 seconds except for agc174605 , which only has ais - depth imaging ( 120 seconds of integration ) .
agc749237 has no fuv data because its 2010 observation followed the 2009 suspension of fuv operations due to an electrical overcurrent ; all the other sources have fuv data .
all 12 of the sources in the survey have nuv data .
the galex data were processed with a pipeline which performed calibration and background subtraction .
details of the galex detectors , pipeline , calibration , and source extraction can be found in @xcite .
galex pipeline products have an assumed 10% flux calibration error .
the fuv and nuv images were cropped to a region centered on the galaxy and excluding most foreground and background sources . by comparison with the hst images ,
remaining contaminants were excised by hand .
photometry was then extracted using standard techniques , and a conversion from raw counts per second in the galex images to magnitudes was performed .
the galex website and associated papers @xcite provide the standard prescriptions to convert from background - subtracted images to ab magnitudes . as @xcite and
others have shown , accounting for severe dust attenuation and extinction of light from extragalactic sources is a difficult problem to address . @xcite and @xcite found that in nearby star - forming galaxies , dust attenuation in the uv regime can vary from zero to several magnitudes .
the galaxies in our sample are low - metallicity dust - poor dwarfs ( see haurberg etal .
2015 for details ) , and so we expect and assume neglible internal dust attenuation . while we would prefer to use an energy balance approach with fuv and total infrared ( tir ) emission to probe the dust - free luminosity ( e.g. , mcquinn etal .
2015b ) , our sources were not imaged at 24 @xmath39 m or longer wavelengths by _
spitzer_. even if we did have _ spitzer _
mips imaging and were able to use this tir+fuv flux method , it is likely that the extinction corrections derived in this way would be small ; for example , the median a@xmath40 found by @xcite for a sample of low - metallicity dwarf galaxies using this method was 0.76 mag . galactic extinction along the line - of - sight can be significant in the fuv and nuv . we applied the method of using the dust maps from @xcite to account for the effects of galactic extinciton .
the extinction at fuv wavelengths is calculated using the following prescription from @xcite : @xmath41 the ultraviolet extinction values for the sample members were all below 1 mag and are included in table [ t1 ] . after correcting the measured fuv magnitudes for galactic extinction
, we converted to fuv luminosities using standard prescriptions and the trgb distances in table [ t1 ] .
other issues affecting the data were minor .
agc111946 resides less than 200 from the edge of the frame in both the fuv and nuv so vignetting is significant .
although the background subtraction performed on this image via the galex pipeline is adequate , our background subtraction uncertainty has still increased to account for the vignetting . for the fuv observation of agc174585 ,
the pipeline background subtraction was unsatisfactory due to a bright foreground source in the northwest corner of the image , so a manual background subtraction was performed using an average sky value from a region in the image which was unaffected by the contaminating source .
ground - based optical images were obtained using the mini - mosaic imager on the wiyn 3.5 m telescope at kitt peak national observatory ( kpno ) .
the observations were performed over four nights : 2010 october 7 - 8 and 2011 march 29 - 30 .
the fall 2010 nights had good seeing ( @xmath6 0.5 ) and were done under photometric conditions ; the spring 2011 nights had degraded seeing ( @xmath6 1.2 ) and the conditions were not photometric .
all sources were observed in four separate filters : broadband johnson b , v , and r , and narrowband w036 h@xmath2 ( fwhm = 60 ) .
exposure times for the fall images were 900 , 720 , 600 , and 900 s for the b , v , r , and h@xmath2 .
the spring images were 1200 , 720 , 900 , and 1200 s exposures .
our treatment of the data used standard prescriptions in iraf . a complete description of the wiyn 3.5 m datasets and imaging procedures can be found in @xcite and haurberg et al .
( in prep ) .
note that agc748778 and agc749241 are non - detections in h@xmath2 , and the latter is a non - detection in the _ spitzer _ images as well . in table
[ t3 ] we include the final h@xmath2 luminosities for all sample members ; related quantities are presented in table [ t4 ] .
sfrs have uncertainties which include both the distance and photometric errors .
imaging of the 12 shield galaxies was obtained for program go-12658 ( p.i .
the hst observations were conducted with the advanced camera for surveys ( acs ) .
the f606w and f814w filters were used , with average total exposure times of 1000 s and 1200 s , respectively .
final 3-color images were produced by creating a third `` green '' image , which is the linear average of the blue and red images ; figure [ ellipses ] contains these final images .
a complete description of the data handling and analysis is given in @xcite ; that manuscript also derives the distances of the shield galaxies shown in table [ t1 ]
. further analysis of the recent sf histories of the shield galaxies , as well as a discussion of the birthrate parameter ( @xmath42 ) in the context of distance to neighboring systems , are given in @xcite .
a discussion of the uncertainty in the trgb distances , which are involved in many calculations and are propagated through in quadrature , may be found in @xcite .
observations of the shield galaxies with _ spitzer _ were acquired in 2011 - 2012 for program go-80222 ( p.i .
cannon ) . near - infrared images at 3.6 and 4.5
@xmath39 m were acquired with the infrared array camera ( irac ) operating during the `` warm phase '' @xcite .
the average exposure time was 96 seconds before mosaicking .
data were acquired from the _ spitzer _ heritage archive and corrected for the effects of column pull - down from bright , saturated foreground objects using the irac instrument team software .
the mosaicker and point source extractor ( mopex ) was then employed to produce flux calibrated , artifact - corrected , mosaicked science images for each galaxy in each filter @xcite . in this paper ,
the _ spitzer _
images are included both for completeness and for straightforward visual comparison of the locations of the old stellar populations of the systems ( seen in 3.6 @xmath39 m in figures [ fig110482 ] - [ fig749241 ] ) to the younger gaseous and stellar regions imaged with other telescopes .
@xcite used the _
data to derive stellar masses of the shield systems ; the companion paper to this work , @xcite , shows the 4.5 @xmath39 m images .
since the shield galaxies have highly irrgular morphologies , determinations of basic galaxy parameters such as inclination are non - trivial .
thus , we derived parameters for photometric analysis using the custom idl program cleangalaxy @xcite , which fits surface brightness contours as a function of galactocentric radius to the hst f606w images of each source .
the position angles , ellipticities , semi - major axes , and inclinations derived for the ellipses are included in table [ t5 ] .
a representative ellipse is overlaid on each of the hst 3-color images in figure [ ellipses ] .
dependent values in the plots and tables of this paper are corrected for inclination effects following the prescriptions described in @xcite . in figures
[ fig110482 ] through [ fig749241 ] , we show a six - panel mosaic for each shield galaxy .
panel ( a ) shows the image used in our sf analysis in greyscale ( created using the robust - weighted , spectrally - smoothed data products described in section [ s2.2.1 ] ) and with column density contours overlaid to demonstrate dynamic range .
the same contours are shown on all remaining panels : ( b ) shows a greyscale representation of the hst f606w image ; ( c ) shows the wiyn 3.5 m
b - band image ; ( d ) shows the spizter 3.6@xmath39 m image ; ( e ) shows the wiyn 3.5 m continuum - subtracted h@xmath2 image ; ( f ) shows the galex fuv image .
all images are regridded to the hst image pixel scale and orientation .
note that all _ spitzer _ images are shown with identified foreground and background objects removed , except for agc749241 , which is a non - detection .
these mosaics facilitate an important visual assessment of the degree to which the gas in these galaxies correlates with regions of ongoing sf .
the fuv and h@xmath2 regions are sometimes co - spatial with the knots ( e.g. , agc110482 , agc 111946 ) , but some systems show the opposite : sf regions appear where minima occur ( e.g. , agc111977 , agc749241 ) .
an in - depth discussion of the situation for each individual galaxy follows in section [ s4.1 ] . in order to compare photometric measurements at multiple wavelengths ,
the images must be registered to the same coordinate grid . for the surface brightness profile analysis
, we preserved the original pixel scale of the individual images in order to faithfully represent the fluxes contained within the elliptical annuli .
the final images ( figures [ fig110482 ] - [ fig749241 ] ) were regridded to the hst fields using the miriad task regrid . for the pixel correlation procedures discussed in section [ s4 ] , the fuv data were smoothed with a gaussian kernel equal to the size of the restoring beam in the maps and then regridded to those maps .
radial profiles were produced by integrating over concentric elliptical annuli in the resulting , h@xmath2 , and fuv images .
observations of the shield sources in the far - ultraviolet ( fuv ) continuum and in the h@xmath2 emission line provide constraints on the sfrs of the galaxies .
both of these tracers are attributable to the formation of young stars @xcite .
fuv radiation has been used as a tracer of spatially and temporally extended sf , revealing populations of relatively young stars ( lifetimes @xmath010@xmath43 yr , masses @xmath44 6 m@xmath11 ; e.g. , salim etal . 2007 ) .
note that usage of any fuv scaling relation to infer a sfr assumes that sf has been constant over a @xmath0100 myr timescale .
near - ultraviolet ( nuv ) emission is also occasionally converted to a sfr , although fuv is generally preferred over nuv : 1 ) nuv images are more contaminated by foreground stellar sources than fuv images ; 2 ) the nuv emission is generally less reliable for tracing the recent sf since the flux at the redder nuv wavelengths will have a greater contribution from stars with lifetimes @xmath44 10@xmath43 yr @xcite .
while the ultraviolet continuum probes sf over the most recent @xmath0100 myr , h@xmath2 line emission provides an almost instantaneous snapshot of the formation of massive stars .
since only stars more massive than @xmath017 m@xmath11 can produce significant numbers of photons capable of ionizing neutral hydrogen @xcite , and these stars have very short main sequence lifetimes , the presence of significant recombination line emission requires the presence of such short - lived stars .
due to this direct coupling between the nebular emission and massive sf , various works have used h@xmath2 emission to probe the ongoing sf in spiral and irregular galaxies ( e.g. , kennicutt 1983 , kennicutt 1998 ) .
even though fuv and h@xmath2 emission have different characteristic timescales , both sf indicators are expected to agree in systems with fully populated initial mass functions ( imf ) and constant sfrs . indeed ,
@xcite find a constant ratio of h@xmath2 to fuv emission in systems with global sfrs larger than @xmath00.1 m@xmath11yr@xmath1 .
however , as the integrated sfr falls , the h@xmath2-based sfr no longer tracks the fuv - based sfr .
possible reasons are numerous , including non - constant sf histories , stochasticity in the upper imf , and leakage of ionizing photons , among others @xcite .
the shield galaxies provide a unique perspective on this issue ; as figures [ fig110482 ] - [ fig749241 ] demonstrate , and as discussed below , some of the shield galaxies harbor only single regions .
the effects of stochasticity are expected to be significant .
many conversions from fuv luminosity to sfr exist in the literature .
most assume a fully populated imf and/or a solar metallicity @xcite .
as discussed in @xcite and @xcite , both of these assumptions fail for the shield galaxies .
a preferable approach takes into account a varying and stochastically populated imf .
we thus apply the recent empirical calibration of @xcite , which is based on a randomly populated salpeter imf with mass limits between 0.1 and 120 m@xmath11 for low - metallicity dwarf galaxies with sfr@xmath45 between 10@xmath46 and 10@xmath1 m@xmath11 yr@xmath1 : @xmath47 where sfr@xmath45 is in m@xmath11 yr@xmath1 and l@xmath48 is in erg s@xmath1 hz@xmath1 .
this prescription yields somewhat higher sfr@xmath45 values ( by @xmath49 ) than the other prescriptions in the works cited above .
the final global fuv - derived sfrs are included in table [ t3 ] .
note that no fuv data are available for agc749237 , and so we have calculated an approximate value for the sfr@xmath45 based on its nuv emission .
we empirically determined the correlation between the fuv and nuv counts per second for each galaxy and use the resulting line - of - best fit equation to estimate what the fuv counts for agc749237 would be , given its observed nuv counts per second .
we then correct for extinction and carry this through to a sfr@xmath40 .
in figure [ sfrvshimass ] we plot the total mass versus the fuv - based sfr for the shield galaxies and for the members of other nearby galaxy surveys .
note that we did not revise the methods used to calculate sfr@xmath45 for the galaxies in the other surveys ; the scatter in sfr@xmath45 may be larger as a result .
this figure demonstrates that the observed sfr@xmath45 of our sample members is comparable to members of some of the other low - mass dwarf galaxy surveys .
as expected , there is a correlation between the total @xmath50 of the systems and the sfr@xmath45 . treating each survey s sample of galaxies independently , the plot also fits a linear regression to each group ; the shield and figgs galaxies , selected in similar ways ( i.e. , had to meet line width criteria ) , show the closest agreement in slope . note that except for a small fraction of the little - things galaxies , all systems which have log(m@xmath51 ) @xmath8 7.5 m@xmath11 have log(sfr@xmath45 ) @xmath8
@xmath212.0 m@xmath11 yr@xmath1 . in figure [ sfrvshimass - composite ]
we treat all of the galaxy surveys uniformly and fit the composite sample with a single linear regression . when no errorbars for archival data were available , we assumed a 10% uncertainty on either variable .
the slope of the resulting line is [email protected] ; if the more massive things galaxies @xcite are excluded the slope becomes slightly more shallow ( [email protected] ) .
we note that these formal uncertainties are likely underestimates , due to the different sfr metrics used by different comparison studies . with this caveat in mind , from figure [ sfrvshimass - composite ] we conclude that the hi mass and uv - based sfr are strongly correlated ; the sfe is essentially constant over five orders of magnitude .
the most widely - used conversion from h@xmath2 luminosity to an instantaneous sfr is given by @xcite : @xmath52 where the sfr@xmath53 is the h@xmath2-based sfr in units of m@xmath11 yr@xmath1 and l@xmath53 is the h@xmath2 luminosity in units of ergs@xmath1 .
this prescription assumes a fully populated salpeter imf ( i.e. , there is a sufficient number of stars throughout the entire mass range , in this case 0 - 100 m@xmath11 ) .
based on the very low fuv - based sfrs , stochasticity in the upper imf ( the high - mass star regime ) is expected to be significant in the shield galaxies . however , in the absence of an alternative calibration of h@xmath2 luminosities into sfrs in the extreme low sfr regime , we adopt the @xcite calibration for purposes of direct comparison .
the global sfr@xmath53 for the shield galaxies are included in table [ t3 ] .
note that agc748778 and agc749241 are h@xmath2 non - detections , and so we provide upper limits for the h@xmath2-based sfr .
the range of sfr@xmath53 of the sample is 10@xmath54 m@xmath11 yr@xmath1 ( agc749237 ) to 10@xmath55 m@xmath11 yr@xmath1 ( agc112521 ) . in figure [ fuvvshasfrs ]
we plot sfr@xmath53 as a function of sfr@xmath45 for the shield galaxies and for comparison samples drawn from the 11hugs @xcite , little - things @xcite , and figgs @xcite samples .
the overall trend of increasing scatter at decreasing sfr is apparent in all survey samples .
the shield galaxies are some of the most extremely low - sfr galaxies in these surveys .
in addition to increased scatter at the faint end of figure [ fuvvshasfrs ] , the sfr@xmath53/sfr@xmath45 ratio becomes progressively smaller as a function of decreasing sfr .
this trend is demonstrated clearly in table [ t3 ] and figure [ sfrratios ] , where the same samples of galaxies are shown in the sfr@xmath53/sfr@xmath45 vs. sfr@xmath53 plane .
the most quiescent systems detected in h@xmath2 are also the most quiescent systems detected in the fuv .
as seen in multiple previous works , the decreased h@xmath2 luminosity relative to fuv is an expected effect of stochasticity at the upper end of the imf in the low sf regime @xcite . @xcite
all noted discrepancies between uv and h@xmath2 sfrs in the sense that h@xmath2 sfrs are systematically lower with decreasing galaxy luminosity .
this was originally interpreted as potentially due to a change in the upper imf with decreasing galaxy mass / metallicity . however , fumagalli etal .
( 2011 ) , da silva etal .
( 2012 ) , and eldridge etal .
( 2012 ) all showed that the observed trend could be explained by a stochastically sampled cluster and stellar mass function scenario .
@xcite demonstrated that the observed trend was a natural result of fluctuating sf rates @xmath21 exactly what is seen in low mass galaxies .
we now examine the complex relationships between the spatial distribution of gas and the spatial distribution of the sf tracers .
we perform this analysis using several different strategies , each of which is discussed in detail below .
first , and most simplistically , we visually inspect the images of each galaxy ( figures [ fig110482 ] - [ fig749241 ] ) at each of the wavelengths available .
analyzing the morphology of the galaxies allows us to identify parts of a system in which the hi gas and sf tracers are co - spatial , as well as areas in a system which have elevated sf but are devoid of hi gas .
it is also interesting to explore the reasons why this co - spatiality does or does not occur ; the visual arrangement of the sf regions and gas within a galaxy can inform us about how much the sf events disrupt the gas and how long it takes for sf to clear out clumps of atomic gas .
second , we calculate and examine radial profiles of the gas surface density ( @xmath4 , units of m@xmath11pc@xmath12 ) , the h@xmath2-based sfr surface density ( @xmath56 , units of m@xmath11yr@xmath1kpc@xmath12 ) , and the fuv - based sfr surface density ( @xmath57 , units of m@xmath11yr@xmath1kpc@xmath12 ) as functions of radius within each galaxy . in order to be systematic and uniform , all profiles are centered on the fitted center of the hst f606w images as determined by cleangalaxy @xcite ; these positions are the centers of the white ellipses shown in figure [ ellipses ] .
they are presented as surface densities in concentric annuli , so the axes in these plots are radius ( in kpc ) and @xmath10 or @xmath18 .
figure [ radprofs ] shows these profiles for all 12 shield galaxies . in general , the profiles show smooth curves while the h@xmath2 and fuv profiles are more choppy and reach the level of noise at smaller radii . in some cases the variations seen in the h@xmath2 curves are reflected in the fuv curves ( e.g. , agc111164 , agc174585 ) , but in other cases the profiles do not mimic each other s shape ( e.g. , agc110482 , agc174605 ) .
the reason for these matchups or discrepancies could be associated with the temporal nature of the h@xmath2 emission compared to fuv : a recently - depleted h@xmath2 region could certainly still be bright in fuv .
third , we plot @xmath10 against @xmath4 in a variety of ways .
one method uses the fuv , h@xmath2 , and emssion within each annulus described above ; the surface densities within the concentric annuli are plotted and a slope is derived from the points with values @xmath44 5@xmath34 .
we present plots of @xmath58 vs. @xmath4 ( using the fuv sfr prescription from mcquinn etal .
2015b ) and @xmath59 vs. @xmath4 ( using the kennicutt 1998 h@xmath2 sfr prescription ) in figures [ fuvsfe ] and [ k98sfe ] , respectively .
because the h@xmath2 data is often clumpy , not radially extended , and in some cases does not have a good s / n ratio , figure [ k98sfe ] is difficult to interpret with confidence ; thus , we do not consider this plot in our final assessment of the average k - s slope for the sample .
figure [ fuvsfe ] includes data for every source in the sample , and because the emission is more smoothly distributed than the h@xmath2 , we have more confidence in using elliptical annuli to find the slope value .
the slopes derived in each of these figures and the sample average are included in table [ t6 ] ; note that the uncertainties quoted are formal uncertainties only .
fourth , we examine @xmath57 vs. @xmath4 on a per - pixel basis . for this analysis , we use a gaussian kernel to smooth the fuv images to the fwhm of the respective images .
the fuv images are regridded to the fields of view and then both are cropped to a 64@xmath2364 pixel grid ( sufficient to include all detected emission in all galaxies ) .
the resulting plots are shown in figure [ sfepixcorr ] ; the best - fit regression line to all pixels is shown in red for each galaxy , and the slope @xmath5 of the @xmath10 @xmath60 @xmath61 relation is shown . if @xmath57 and @xmath4 are both high in the same pixels @xmath21 that is , the fuv and emission are co - spatial @xmath21 and they are both relatively low in the same pixels , then we expect a positive slope of the best - fit line . if the @xmath57 falls off more quickly , then the slope is steeper , but if the @xmath4 falls off more quickly , then the slope is shallower .
an inspection of the images shown in figures [ fig110482 ] through [ fig749241 ] reveals that both the gas and the fuv emission have resolved structure ; further , some sources have significant offsets of versus fuv .
thus , the slopes of the lines are sometimes negative ( e.g. , agc748778 and agc749241 ) , indicating an anti - correlation between the atomic gas and the areas of elevated sf activity .
the slopes of the k - s relations seen in figure [ sfepixcorr ] and the average slope for the sample are included in table [ t6 ] .
finally , we show a series diagnostic plots which serve two functions : comparing the and sf properties , and demonstrating potential effects of the physical resolution on a system s peak hi column density and k - s slope .
the first of these , figure [ coldentrend ] , compares the peak column density of each galaxy to its sfr ( derived from both fuv and from h@xmath2 ) .
this allows us to search for trends related to the maximum surface density detected in a given source .
the second , figure [ globalsigmas ] , compares the global and sfr surface densities of the shield sample against those in other relevant studies . for the shield galaxies , the global values
are calculated by summing all fuv emission and all emission in concentric annuli that encompass all of the fuv flux ; regions of extended gas are not included in the calculation of the areas .
this plot allows us to contextualize the global sf properties of the shield sample with respect to sources with significantly higher and lower sfr surface densities ; however , we note that the derivation of the sfr is different in each of the included surveys ( kennicutt 1998a , wyder etal . 2009 , roychowdhury etal . 2014 )
. figures [ resolutionvscolden ] and [ slopevalsvsresolution ] compare the physical resolution sizes of the shield sources against their observed peak hi column density and derived k - s @xmath5 value . using the above diagnostics , we examine causes for the appearances and trends and compare our composite sample to other low - mass galaxy surveys . in section [ s4.1 ]
we discuss each galaxy individually , focusing on the morphology , physical reasons for the derived k - s slopes , and local environment surrounding each source . in section [ s4.2 ]
we describe the results for the sample as a whole and contextualize our results with regard to other relevant studies . because we have several different metrics for determining the slope for the kennicutt - schmidt ( herefafter , k - s ) relation
, we will discuss the relative strengths and weaknesses of each method .
note that error bars are shown in the figures , but are often unseen because they are smaller than the points themselves .
all lines - of - best - fit that are fitted to data in the plots have been weighted by their uncertainties .
average values for the sample , such as the average consumption time , are weighted averages . as figures [ fig110482 ] through [ fig749241 ]
demonstrate , some galaxies in the shield sample have ongoing sf in regions of relatively high column density while others do not .
the strength of the correlation between the sf tracers and the locations of elevated column densities varies dramatically among the galaxies . as shown in table [ t2 ] , only 4 of galaxies ( agc110482 , agc174585 , agc731457 , and agc749237 ) have column densities that exceed 10@xmath19 cm@xmath12 ( shown by an orange contour in those respective figures ) , yet several of the other sources with `` sub - critical '' peak column densities have co - spatial knots and sf regions as traced by both fuv and h@xmath2 .
we now briefly discuss each of the 12 galaxies in turn .
: there is excellent agreement between the two prominent h@xmath2 regions and the fuv peaks .
the southeastern peak of the ( @xmath01.7@xmath2310@xmath19 cm@xmath12 ) is exactly co - spatial with the h@xmath2 and fuv peaks in that area .
there is also a second northwestern maximum in the sf tracers which is very close to the n@xmath27 @xmath35 10@xmath19 cm@xmath12 contour .
based on these qualities , we expect the smooth radial profile of surface density and a visible bump in the h@xmath2 sfr surface density profile seen in figure [ radprofs ] . the k - s slope is steeper in figure [ fuvsfe ] than in figure [ k98sfe ] .
the k - s index is [email protected] from the pixel - by - pixel correlation method ( figure [ sfepixcorr ] ) .
@xcite notes that agc110482 is @xmath22 0.65 mpc from other members of the ngc672 group .
: the gas is distributed almost circularly ( but note that agc111164 is one of the three galaxies that does not have b - configuration imaging , so the resolution element is comparatively coarse ; however , because this galaxy lies at a relatively small distance , the linear resolution is comparable to other sources ) . however ,
the stellar component of the galaxy is elongated in the hst and b - band , and although the _ spitzer _ image shows similar overall structure , there is a line of bright infrared emission perpendicular to the major axis of the galaxy which is not highlighted in the other panels . the fuv and h@xmath2 maxima are co - spatial with b - band and maxima .
additionally , the fuv emission is somewhat extended but the h@xmath2 morphology is consistent with a single region ; the radial profiles shown in figure [ radprofs ] highlight this structure .
the peak column density is among the lowest in the sample ( again , the coarse resolution element likely affects this ) , yet massive sf is occurring .
the slopes of the fuv and the h@xmath2 surface density profiles differ . using the fuv images , we determine @xmath15 [email protected] on a pixel - by - pixel basis .
@xcite notes that this galaxy lies only @xmath00.14 mpc from ngc784 and is part of a linear structure of galaxies associated with this larger dwarf starburst galaxy . :
the column density does not reach the 10@xmath19 cm@xmath12 level . however , the gas is generally co - spatial with the stellar component in the hst and wiyn optical images , and the two h@xmath2 and fuv peaks lie precisely within the peaks .
there are observable bumps in the radial profiles ( figure [ radprofs ] ) due to the morphology of the sf tracers and the .
the slope derived in figure [ sfepixcorr ] is [email protected] .
@xcite finds that this galaxy is also located in the ngc784 group , but is @xmath01 mpc away from its nearest neighbor , suggesting that the recent sf activity is not driven by gravitational interactions .
: this is one of the most enigmatic shield galaxies .
the peak column density is low ( although this source was not observed in the b configuration and so the beam size is coarse , and the relatively larger beam size yields a poorer linear resolution ) .
there is a significant diffuse southern h@xmath2 region that is prominent in the fuv as well .
interestingly , the column density maximum is on the other side of the disk from these sf regions . because of the large offset between and sf peaks , the slopes of the radial profile plots are difficult to interpret .
the pixel - by - pixel method shows a relatively shallow slope with significant dispersion in the @xmath10 dimension ; the average slope is [email protected] .
@xcite finds that agc111977 is separated by @xmath00.65 mpc from agc112521 . :
the peak column density is only half of the canonical threshold value ( although this source was not observed in the b configuration and so the beam size is coarse , resulting in a poorer linear resolution ) , and there is very little sf currently ; the fuv and h@xmath2 luminosities are very low .
the single region in the north is coincident with the fuv and b - band maximum .
the h@xmath2 and fuv radial profiles are dominated by small number statistics .
the pixel - by - pixel method shows the smallest scatter of any of the shield galaxies in the @xmath10 dimension ; the average slope is [email protected] .
@xcite finds that agc112521 is @xmath00.65 mpc from several neighbors , including agc110482 , agc111977 , and three members of the ngc672 group .
: this galaxy has a small peak that exceeds the 10@xmath19 cm@xmath12 column density threshold ; this peak is not co - spatial with the sf regions in the source , which are extended spatially in both the h@xmath2 and the fuv images .
the radial profiles quantify what is clear from the images in figure [ fig174585 ] : extended sf is associated with gas of a range of mass surface densities . as expected , the slope of the pixel - by - pixel method is among the most shallow in the sample ( @xmath15 [email protected] ) .
@xcite notes that agc174585 is located @xmath00.9 mpc away from its nearest neighbors . :
although the column densities are low , the peak is tightly correlated with the h@xmath2 and fuv peaks ( well within the beam ) .
the fuv image is the only one in the sample at ais depth , and the noise is high .
similarly , the h@xmath2 image has a pronounced gradient .
the radial profiles show somewhat steeper slopes than the pixel - by - pixel method ( @xmath15 [email protected] ) .
@xcite finds that agc174605 is truly isolated with no known neighbors in a 1 mpc radius .
: this galaxy has the lowest peak column density of the shield galaxies , and yet it is relatively luminous in both h@xmath2 and the fuv .
note that both the h@xmath2 and the fuv emission are spatially extended .
there is a noticable offset between the hi peak and the hst , b - band , and fuv peaks , and it is even more apparent when compared to h@xmath2 .
agc182595 is the only source in the sample with a ratio of sfr@xmath53/sfr@xmath45 @xmath44 1 , perhaps indicating a modest starburst episode .
@xcite notes that it has the lowest gas - to - stellar mass ratio of the sample and has the second - shortest gas consumption timescale .
figures [ fuvsfe ] , [ k98sfe ] , and [ sfepixcorr ] all show smooth slopes consistent with trends seen in larger galaxies .
although the @xmath5 values range considerably in these plots , the slopes are all well - defined ; the pixel - by - pixel average slope is [email protected] .
@xcite finds that agc182595 is truly isolated with no known neighbors in a 1 mpc radius . :
this source eclipses the 10@xmath19 cm@xmath12 column density threshold and has one of the highest sfrs in the sample using both h@xmath2 and fuv metrics .
there are several luminous h@xmath2 regions ; some of these are co - spatial with the extended fuv emission , but there are regions that are h@xmath2 bright and fuv dim and vice - versa .
the peak is slightly offset not only from the sf regions but also from the optical counterpart of the source as shown by the hst image .
a cursory inspection of figure [ fig731457 ] shows that sf regions are associated with gas at a broad range of observed mass surface densities . as in agc182595
, we see steeper slopes for the radial sfr surface density plots than for the pixel - by - pixel correlation ; agc731457 has the shallowest ( positive ) slope of any of the shield galaxies via this metric ( n @xmath35 [email protected] ) .
@xcite notes that agc731457 is located @xmath00.9 mpc away from its nearest neighbors .
: this source has a dramatic morphology that is highly extended relative to the stellar population , extending more than 1 kpc to the south .
two maxima are observed , only one of which overlaps spatially with the stars in the galaxy .
the h@xmath2 emission is very weak in this source , so much so that it was a non - detection in our observations .
interestingly , the source is detected with significance in the fuv image .
the recent sf traced by this fuv flux is associated with a range of mass surface densities , and some fuv flux contains no associated gas at our current level of sensitivity .
@xcite finds that agc748778 is truly isolated with no known neighbors in a 1 mpc radius , suggesting that it is unlikely that gravitational interactions have driven the recent sf activity .
: this is the largest galaxy in the sample by mass , optical and diameter , and it has the highest sfr ( in both fuv and h@xmath2 ) .
it has the second highest peak column density in the sample .
the most intriguing aspect of the galaxy is that two of the h@xmath2 and fuv knots are co - spatial with two of the peaks , but not the largest and densest one .
figure [ fig749237 ] reveals that the highest mass surface densities are associated with the outer disk of the system and contain no h@xmath2 and weak fuv emission ( compare the h@xmath2-based sfr density profile in figure [ k98sfe ] with the fuv - based profile in figure [ fuvsfe ] ) . in figure [ sfepixcorr ] , we see an intriguing effect : the large spread in fuv points at the peak values occurs because there are both very high and very low fuv values associated with the peak pixels ; this trend is seen clearly in figure [ fig749237 ] .
nonetheless , agc749237 stands out in the pixel - by - pixel diagram as having the steepest k - s index ( @xmath15 [email protected] ) .
@xcite finds that this galaxy is truly isolated with no known neighbors in a 1 mpc radius , so it is likely that its recent sf activity is internally regulated .
: like agc748778 , this source has a highly unusual morphology , where the bulk of the gas is not co - spatial with the stellar component .
the source is a non - detection in both the h@xmath2 and infrared images . like agc748778
, it is detected with confidence in the fuv , although it does harbor the lowest sfr@xmath45 in the sample , and shows an obvious offset in the gas and fuv emission . the @xmath10 radial profile and the pixel - by - pixel
diagnostic both suggest a negative k - s index for this source . as for agc748778 ,
the offset of and fuv is the cause . for a galaxy to have such inefficient sf combined with a highly offset component
is certainly an interesting scenario .
the question of whether some kind of tidal interaction has separated the so visibly from the stars has been addressed : @xcite investigates the environment surrounding this source and finds that agc749241 lies in populated region with 9 galaxies which form a linear structure 1.6 mpc from end to end .
while it is possible that gravitational interactions among the systems in this structure have contributed to the offset of gas and stars visible in agc749241 , the low recent sf activity found in the galaxy indicates that any gravitational interaction has not had a dramatic impact on the star - forming properties over the past 200 myr .
the discussion above highlights the difficulty of determining a sf `` law '' in extremely low - mass galaxies .
using h@xmath2 images presents challenges ; in a given galaxy , regions are few in number ( at most 3 well - defined clumps in a single shield source ) , faint ( low total luminosity ) , and sometimes offset from the optical centers of the sources . while the use of fuv images has certain advantages ( longer sf timescales , more uniform spatial coverage ) , the physical sizes of the galaxies themselves ( of order 1 kpc or smaller ) present fundamental limitations that are not encountered when undertaking similar studies of more massive galaxy disks .
further , the sources are relatively faint
. neither of these problems would exist for more massive galaxies at the same distances . despite these challenges , we produce three sets of plots in order to decipher the k - s relation for each of the shield galaxies ( figures [ fuvsfe ] , [ k98sfe ] , and [ sfepixcorr ] ) .
because two different methodologies were used and two different sf tracers were involved , the slope values derived in the plots vary substantially .
for these three figures , we derive average slopes of n @xmath6 [email protected] , n @xmath6 [email protected] , and n @xmath6 [email protected] , respectively .
note that the inclusion of agc748778 and agc749241 , both of which have low or negative slopes in the fuv - based figures and are absent from the h@xmath2-based figure , may skew the average slopes to low values .
further , the images used to create all three figures are smoothed to a small degree , effectively smearing out the flux from the peak and possibly making the slopes less dramatic . comparing the modes of analysis noted above
, we conclude that the pixel - by - pixel correlation method ( figure [ sfepixcorr ] ) provides the most holistic and reliable representation of the comparative qualities of and recent sf in the shield galaxies .
the uncertainties for the slopes are lower for the pixel - by - pixel method than for the radial profile methods ( figures [ fuvsfe ] and [ k98sfe ] ) , and the power - law slopes from the pixel - by - pixel method more closely match what would be expected from visual inspection of figures [ fig110482 ] - [ fig749241 ] .
we also note that the inner annuli used to derive the surface densities and produce figures [ fuvsfe ] and [ k98sfe ] are sensitive to small changes in the precise positioning of the ellipse centers ; moving in a given direction by a single pixel can include or exclude significant h@xmath2 or uv flux ( moving the aperture by an entire beam size can result in flux changes of 50% or more ) , thus changing the shape of the resulting power - law slope . using figure [ sfepixcorr ] as our metric
, the k - s slope values of the 12 galaxies break down as follows : 2 sources have @xmath62 , 6 have @xmath63 , and 4 have @xmath64 .
we must note here that our technique for analyzing the k - s relation on sub - kpc scales differs in an important way from other recent studies .
because the galaxies in a given sample all reside at different distances , the physical size scale for each source will be different .
the native resolution of the hi beam will also contribute directly to this size scale : a beam of finer resolution will yield a smaller physical resolution .
these other studies ( such as roychowdhury etal .
2015 , roychowdhury etal .
2009 , and bigiel etal .
2008 ) smooth the data for each galaxy in their sample to a common physical resolution ( e.g. , @xmath0400 pc or 1 kpc ) in order to analyze the relationship of gas and stars on the same physical scale .
we chose not to perform this smoothing , instead keeping the shield galaxies at their native physical resolutions ( see table [ t2 ] ) .
there are a number of reasons for this decision : first , the minimum common resolution we would have had to smooth to would have been @xmath0600 - 650 pc ( limited unfortunately by those sources which lack vla b - configuration hi data ) .
while this does not pose a serious problem for the things galaxies ( larger spiral galaxies ) or the figgs galaxies ( dwarf galaxies with a mean sample distance of 4.7 mpc ) , the shield survey works with galaxies which are of similar size to figgs but span a factor of two in distance .
our sources are at a mean distance of 8.1 mpc with none closer than 5.1 mpc .
this range in distances means that we can fit fewer resolution elements across a galaxy@xmath21for example , even agc749241 ( at a relatively small distance of 5.62 mpc ) is only @xmath02100 pc across in its hi map , meaning we could fit less than four smoothed resolution elements across the galaxy , effectively smearing out any useful spatial information .
second , the high - resolution data gives us insight into the exact locations of the hi peaks and also lets us see the highest column density hi gas ; if we smoothed all galaxies to a coarser common resolution , our analysis of the hi column density threshold would be less secure .
third , our analysis of the shield galaxies on a variety of fine resolutions demonstrates the breakdown of the k - s relation on small scales , as found in previous studies , such as @xcite .
situations such as agc111977 and agc749241 reveal certain advantages of not fixing the physical resolution : both of these sources have interesting morphologies , with hi peaks that are distinctly offset from the sf tracers .
both of these sources are at a distance of @xmath06 mpc , but agc111977 lacks b - configuration data and thus has a resolution about four times as coarse as agc749241 .
unfortunately , this makes comparing their k - s relations challenging , and it also begs the question : if higher - resolution data was available for agc111977 , what would we see ?
instead of having a slope of @xmath65 [email protected] when the physical resolution is @xmath0690 pc , it might shift to a flat or negative slope , as is the case with agc749241 at a resolution of @xmath0170 pc . likewise , smoothing the data for agc749241 to a coarse resolution might reduce the strength of the anti - correlation we currently see in figure [ sfepixcorr ] . based on the analysis in section [ s4.1 ] and the pixel - by - pixel metric ( figure [ sfepixcorr ] ) , we divide our sample into three broad categories that share similar characteristics : 1 ) mainly co - spatial and sf regions , found in systems with highest peak column densities and highest total masses ; 2 ) systems which show a range of correlation strengths between and sf regions , and which also span a range of peak column densities ; and 3 ) obvious offsets between and sf peaks , found in systems with the lowest total masses .
these groupings are heterogeneous but useful for comparison purposes .
group 1 contains the systems that most closely resemble the `` expectations '' of active sf in regions of high column density : agc110482 , agc111946 , agc174585 , agc731457 , and agc749237 .
these sources demonstrate the strongest observed correlations between the and sf peaks , and each system reaches a relatively high peak column density ( @xmath22 9@xmath2310@xmath66 cm@xmath12 ) .
it is important to note that the comparative properties of and sf vary among and within these sources : the pixel - by - pixel @xmath5 indices of these sources vary between 0.59 and 2.04 , and agc110482 shows compact regions while agc731457 shows extended h@xmath2 emission and widely extended fuv emission .
these comparatively `` well - behaved '' sources are generally at the higher - mass end of the shield sample .
we include agc749237 in this first group , but note that while some of its ongoing sf is strongly correlated with high mass surface density neutral gas , the highest columns are devoid of sf altogether .
group 2 contains a variety of shield sources which span a range of properties : agc111164 , agc111977 , agc112521 , agc174605 , and agc182595 .
they have moderate peak column densities ( @xmath67 5@xmath2310@xmath66 cm@xmath12 ) and display a range of correlation strengths between these peaks and the regions of sf .
we must also note that the three sources which are lacking b - configuration data are included in this group ; it is likely that their low sensitivity to high column density hi gas contributes to the low hi column densities of those sources which have been grouped here .
visual inspection of the images of these galaxies illustrates some similar traits among these sources .
in general , higher than average columns are nearly co - spatial with sf tracers ( e.g. , agc111164 ) . again , notable issues exist amongst and within the members of this group : agc111977 , for example , harbors extended sf , but the peak is not co - spatial with the highest surface brightness emission . with the exception of agc111164 , these systems also represent the middle of the mass range for the sample . finally , group 3 contains two highly unusual galaxies : agc748778 and agc749241 .
these systems stand out compared to the other shield galaxies : they harbor the lowest stellar masses @xcite , the narrowest linewidths ( see table [ t2 ] ) , the weakest h@xmath2 emission , and among the weakest fuv emission .
further , their morphologies are highly unusual compared to the other sample members , in that the gas is significantly offset from the high surface brightness stellar population .
the negative k - s indices highlight the extreme nature of these sources ; with the present data we are unable to determine the origin of the offset of the and the stars , but two scenarios are possible .
one possibility is tidal effects from the local environment near the galaxy .
@xcite examined the environment immediately surrounding each of the shield sources in 3 dimensions using the nearest neighbor metric , and suggested that agc749241 has likely been influenced by the external gravitational perturbations from other systems .
another possible scenario is that both of these galaxies had strong sf events in the past , resulting in the prominent fuv but non - existant h@xmath2 emission seen today . due to their low mass ( both stellar and hi )
, these episodes might have disrupted the central hi gas and , in turn , shut down any further sf .
while the details are complex and depend on the coupling of mechanical feedback energy to the surrounding ism , over time , this could create an apparent offset between the gas and the stars .
while the offsets might appear more extreme in these two systems compared to the rest of the sample , the observed offsets in other galaxies in the sample ( e.g. , agc111977 ) could also be explained by a major sf event disrupting the nearby gas . in this scenario ,
the time elapsed since a major sf event matters greatly : regions would only be co - spatial with the hi gas peaks if the sf event is young or current and hasnt had time to disrupt the gas significantly .
figure [ resolutionvscolden ] plots the shield galaxies in these three distinct groups to identify trends based on the physical resolution and peak hi column density of each source .
this plot confirms that the peak hi column observed certainly depends on the fineness of the resolution element , indicated by the location of group 2 in comparison to group 1 .
however , group 3 defies this trend , and there exist members of group 1 which have resolutions barely better than members of group 2 yet which have significantly higher hi columns .
we also note that group 2 contains all three galaxies which lack b - configuration data ; this demonstrates the degree to which the high - resolution vla observations allow us to see higher column density hi gas .
figure [ slopevalsvsresolution ] yields a different perspective .
again , the three groups of galaxies do not mingle very much , and it is apparent that some correlation exists between the physical resolution element for a galaxy and its resulting k - s slope . at smaller resolutions ,
it is more likely that the hi gas will appear offset from the sf regions , resulting in flatter or more negative @xmath5 values .
this trend can also be seen in the images of the galaxies ( figures [ fig110482 ] - [ fig749241 ] ) and in figure [ sfepixcorr ] .
these plots together provide evidence for one of our key results : at small physical scales , the k - s relation is no longer valid . taken as a composite sample , the average of the pixel - by - pixel k - s indices for the shield galaxies is @xmath5 @xmath6 [email protected] ( based on the slopes derived in figure [ sfepixcorr ] ) . to put this average value and the individual galaxies values in perspective
, we compare to @xcite and @xcite , who find @xmath68 [email protected] for the molecular schmidt law ( only relating @xmath69 and @xmath10 ) in their sample of spiral galaxies .
they do not ultimately base their conclusions on the relation of @xmath4 to @xmath10 because most of their galaxies show little correlation between mass surface density and ongoing sf tracers .
our @xmath5 values are quite similar and our method intentionally parallels theirs ; however , it is important to remain mindful that this comparison is not one - to - one , since we use data exclusively and the studies of @xcite and @xcite account for the molecular phase .
those works conclude that , even for the spirals which are qualitatively different from the shield galaxies in fundamental ways , the sfe varies considerably across their sample and within their individual galaxies .
the conclusion of both studies is that the sfe is set by local environmental factors ( i.e. , sf is internally regulated ; see mcquinn et al .
2015a ) .
other recent work has extended the k - s analysis to other types of galaxies ; figure [ globalsigmas ] presents this comparison .
a study of low surface brightness ( lsb ) galaxies by @xcite finds that these sources tend to lie below the canonical composite ( + ) k - s relation from @xcite .
an extrapolation of the slope for these sources would appear even steeper than for the larger and higher surface brightness galaxies . however , the analysis by @xcite of figgs dwarf irregulars yields an @xmath68 [email protected] , which is in excellent agreement with our results .
both of these slopes are clearly shallower than the composite slope from @xcite , indicating a deviation from the canonical k - s law found for higher mass galaxies but in a different manner than found in @xcite .
each of these low - mass galaxy studies@xmath21shield , the lsb galaxy study , and figgs@xmath21assume the molecular component of the gas is negligible .
the authors of the figgs study suggest that their results favor a model of sf where thermal and pressure equilibrium in the ism regulate the rate at which sf occurs , where the thermal pressure in turn is set by supernova feedback .
our analysis leads us to a similar conclusion as those presented in these other works : sf in extremely low - mass galaxies is dominated by stochasticity and random fluctuations in their ism .
using the qualities of gas alone , it is very challenging to predict where a given system will show signs of recent sf .
similarly , knowing the distribution of h@xmath2 and/or fuv emission offers little insight into where we might expect to see high mass surface density .
we do not see strong evidence for a threshold column density above which we see signs of recent sf and below which we do not .
the shield sample contains multiple examples of both `` super - critical '' and `` sub - critical '' gas associated with h@xmath2 and with fuv emission .
as figure [ coldentrend ] demonstrates , the column density of an region correlates only loosely with its sfr in our sample .
while the differing hi beam dimensions and resulting physical resolutions certainly play a role in determining the highest column density gas detected , a sample - wide trend would not necessarily be borne out by improved resolution for the low column density sources ; agc748778 and agc749241 have only moderate hi column densities despite their fine physical resolutions .
overall , our data suggest that sf does not only occur in regions exceeding the suggested critical column density .
in fact , there does not even seem to be a lower threshold : members of group 2 ( see above ) all have relatively co - spatial ongoing sf and hi peaks , yet they have column densities @xmath67 5@xmath2310@xmath66 cm@xmath12 where the sf is occurring ; members of group 3 have significant fuv emission in regions that are in fact devoid of entirely down to the limits of our observations .
it is apparent from our study of these low - mass dwarf galaxies that is not a useful tracer of sf when used in the k - s relation ; numerous studies have come to similar conclusions @xcite .
additionally , recent studies by @xcite and @xcite have found that low - mass sf galaxies , despite having very low molecular gas fractions , are still able to form massive stars .
this idea is supportive of our results in that high - mass sf has occurred in the metal - poor shield galaxies ; the stochastic nature of sf is certainly a contributing factor to the variation in k - s relations seen in the sample .
we close by commenting on the characteristic gas consumption timescales ( gcts ) of the shield galaxies .
there are two methods of calculating the gct ( the time it would take for the galaxy , at its present - day sfr , to completely use up its neutral hydrogen gas reservoir ) ; the first requires taking the inverse of the total sfe . in this way
, we find an average gct for the sample of @xmath02@xmath2310@xmath70 yr from the global comparison of @xmath10 to @xmath71 .
this is nearly an order of magnitude lower than those found in @xcite for the figgs galaxies , and roughly two orders of magnitude below the timescales derived in the outer regions of spiral galaxies . @xcite and @xcite
find that molecular gcts for the things galaxies are of order @xmath02 @xmath23 10@xmath70 yr . the authors also suggest a mechanism that we conclude is likely at play in our sample as well : microphysics in the interstellar medium below the scales which we can observe ( in the form of random gas motions and stellar feedback ) which govern the formation of molecular gas from .
employing the second method , in which we simply divide the total hi gas mass by the total sfr of the galaxy , we get slightly different results .
this straightforward approach yields an average gct for the shield galaxies of @xmath010@xmath26 yr ; if we ignore the most extreme outlier in the sample ( agc112521 , with a derived gct of @xmath045@xmath2310@xmath70 yr ) , then the average drops to @xmath07@xmath2310@xmath70 yr .
these higher average values show consistency with results for a sample of alfalfa dwarfs , in which the authors found a gct of 8.9@xmath2310@xmath70 yr @xcite .
consumption times derived via both methods are included in table [ t4 ] .
note that the uncertainties for both modes of calculation are significant .
shield is a systematic investigation of a sample of extremely low - mass dwarf galaxies outside the local group . despite the low column densities observed in many systems , each shield galaxy has a significant fuv luminosity , and we detect h@xmath2 emission in all but two of them .
the ability to compare multi - configuration vla data with sf tracers in other wavelengths allows us to examine , on a local and global scale , the sf `` law '' in these systems .
we calculate @xmath4 and @xmath10 for all sample members and find the sfes .
we derive an index for the kennicutt - schmidt relation via several different methodologies .
the ensemble average index using the pixel correlation mehtod gives @xmath65 [email protected] ; this is in good agreement with other studies of low - mass dwarf galaxies , but is shallower than the canonical k - s relation from @xcite . by comparing the shield results to those from other major nearby galaxy surveys , we find that hi mass and uv - based sfr are strongly correlated over five orders of magnitude . we stress that any one galaxy in the sample is not representative , and that the values of @xmath5 vary considerably from system to system .
instead , we focus on the narratives of the individual galaxies and their distribution of gaseous and stellar components , which are complex and occasionally puzzling .
the average consumption time for the sample of @xmath02 - 10 gyr suggests that they will consume their gas reservoir over timescales similar to those found in other dwarf galaxy surveys . at the extremely faint end of the mass function
, these systems appear to be dominated by stochastic motions in their extreme ism .
the local microphysics within the ism of individual galaxies may govern whether or not they show signs of recent sf . based on our data ,
knowledge of the properties holds little predictive power in terms of the resulting sf characteristics .
similarly , we see ongoing or recent sf in unexpected regions of many of the shield galaxies . the observed offsets between gas and stars in the galaxies could originate from tidal interactions , or the offsets might appear based on how much time has elapsed since a major sf event disrupted the central gas component . if causal relationships between atomic gas and sf exist in galaxies in this extreme mass range ( 6.6 @xmath24 log(m@xmath27 ) @xmath24 7.8 ) , these relationships remain elusive with current data .
in addition to the analysis presented here , a companion paper by @xcite studies the hi gas kinematics and dynamics of the shield galaxies .
two- and three - dimensional analyses are used to constrain rotational velocities .
it is argued that the shield galaxies span an important mass range where galaxies transition from rotational to pressure support .
the sources are contextualized on the baryonic tully - fisher realtion .
the now - complete alfalfa catalog contains dozens of galaxies with and stellar properties comparable to those of the 12 shield galaxies studied in this work . observations similar to the ones presented here are underway to characterize the sf properties of this statistically robust sample .
the authors acknowledge the work of the entire alfalfa collaboration team in observing , flagging , and extracting the catalog of galaxies used to identify the shield sample .
the alfalfa team at cornell is supported by nsf grants ast-0607007 and ast-1107390 to rg and mph and by grants from the brinson foundation .
y.g.t , a.t.m . , and j.m.c
. are supported by nsf grant ast-1211683 .
is supported by top1ew.14.105 , which is financed by the netherlands organisation for scientific research ( nwo ) .
we gratefully acknowledge adam leroy for the idea of an idl code for mask - creation , and charlotte martinkus & ned molter for moral support .
this work would not have been possible without the incredible support for python provided by the online community at stackoverflow.com .
plots featured in this paper have been created with the python package matplotlib .
support for hubble space telescope data in this work was provided by nasa through grant go-12658 from the space telescope institute , which is operated by aura , inc .
, under nasa contract nas5 - 26555 .
the arecibo observatory is operated by sri international under a cooperative agreement with the national science foundation ( ast-1100968 ) , and in alliance with ana g. mndez - universidad metropolitana , and the universities space research association .
galaxy evolution explorer ( galex ) is a nasa small explorer , launched in 2003 april .
we gratefully acknowledge nasa s support for construction , operation , and science analysis for the galex mission , developed in cooperation with the centre national detudes spatiales of france and the korean ministry of science and technology .
this research made use of nasa s astrophysical data system , the nasa / ipac extragalactic database which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration , and montage , funded by the nasa s earth science technology office , computation technologies project , under cooperative agreement number ncc5 - 626 between nasa and the california institute of technology .
montage is maintained by the nasa / ipac infrared science archive .
_ facilities : hubble space telescope , galaxy evolution explorer , _ spitzer _ space telescope , kitt peak national observatory s wisconsin - indiana - yale - noao observatory 3.5 m telescope , expanded very large array .
_ ccccccccccc agc110482 & 01:42:17.4 & 26:22:00 & [email protected] & [email protected] & [email protected] & [email protected] & 7.74@xmath72 & 357@xmath71 & [email protected] & 0.75 + agc111164 & 02:00:10.1 & 28:49:52 & [email protected] & [email protected] & [email protected] & [email protected] & 7.00@xmath73 & 163@xmath73 & [email protected] & 0.45 + agc111946 & 01:46:42.2 & 26:48:05 & 9.02@xmath74 & [email protected] & [email protected] & [email protected] & 7.23@xmath75 & [email protected] & [email protected] & 0.68 + agc111977 & 01:55:20.2 & 27:57:14 & 5.96@xmath76 & [email protected] & [email protected] & [email protected] & 7.57@xmath77 & 207@xmath72 & [email protected] & 0.58 + agc112521 & 01:41:07.6 & 27:19:24 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & 0.51 + agc174585 & 07:36:10.3 & 09:59:11 & 7.89@xmath78 & [email protected] & [email protected] & [email protected] & 6.95@xmath79 & 356@xmath73 & ... & 0.32 + agc174605 & 07:50:21.7 & 07:47:40 & [email protected] & [email protected] & [email protected] & [email protected] & 7.45@xmath80 & 351@xmath71 & ... & 0.19 + agc182595 & 08:51:12.1 & 27:52:48 & [email protected] & [email protected] & [email protected] & [email protected] & 7.70@xmath81 & 398@xmath72 & [email protected] & 0.34 + agc731457 & 10:31:55.8 & 28:01:33 & 11.13@xmath82 & [email protected] & [email protected] & [email protected] & 7.81@xmath83 & 454@xmath73 & [email protected] & 0.23 + agc748778 & 00:06:34.3 & 15:30:39 & 6.46@xmath84 & [email protected] & [email protected] & [email protected] & 6.48@xmath85 & [email protected] & ... & 0.52 + agc749237 & 12:26:23.4 & 27:44:44 & 11.62@xmath82 & [email protected] & [email protected] & [email protected] & 7.72@xmath86 & 372@xmath71 & [email protected] & 0.16 + agc749241 & 12:40:01.7 & 26:19:19 & 5.62@xmath87 & [email protected] & [email protected] & [email protected] & 6.60@xmath88 & 451@xmath71 & ... & 0.12 + [ t1 ] ccccccc agc110482 & 1.0 & [email protected] & [email protected] & 450 & 30@xmath72 & [email protected] + agc111164 & 1.4 & [email protected] & [email protected] & 530 & 27@xmath76 & [email protected] + agc111946 & 1.1 & [email protected] & [email protected] & 450 & 21@xmath73 & [email protected] + agc111977 & 1.5 & [email protected] & [email protected] & 690 & 26@xmath74 & [email protected] + agc112521 & 1.3 & [email protected] & [email protected] & 700 & 26@xmath71 & [email protected] + agc174585 & 0.88 & [email protected] & [email protected] & 240 & 21@xmath76 & [email protected] + agc174605 & 0.51 & [email protected] & [email protected] & 620 & 24@xmath72 & [email protected] + agc182595 & 0.69 & [email protected] & [email protected] & 440 & 20@xmath74 & [email protected] + agc731457 & 0.88 & [email protected] & [email protected] & 330 & 36@xmath76 & [email protected] + agc748778 & 0.93 & [email protected] & [email protected] & 190 & 16@xmath73 & [email protected] + agc749237 & 0.79 & [email protected] & [email protected] & 350 & 65@xmath72 & [email protected] + agc749241 & 0.79 & [email protected] & [email protected] & 170 & 18@xmath72 & [email protected] + [ t2 ] ccccccc agc110482 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -2.27@xmath89 + agc111164 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -3.28@xmath90 + agc111946 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -2.46@xmath91 + agc111977 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -2.59@xmath92 + agc112521 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -3.44@xmath93 + agc174585 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -2.64@xmath94 + agc174605 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -2.37@xmath95 + agc182595 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -2.32@xmath96 + agc731457 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -1.85@xmath97 + agc748778 & [email protected] & @xmath2436.18 & [email protected] & @xmath24 - 4.92 & [email protected] & -3.17@xmath98 + agc749237 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & -1.89@xmath99 + agc749241 & [email protected] & @xmath2436.10 & [email protected] & @xmath24 - 5.00 & [email protected] & -3.49@xmath100 + [ t3 ] ccccccc agc110482 & [email protected] & [email protected] & [email protected] & 3@xmath71 & [email protected] + agc111164 & [email protected] & [email protected] & [email protected] & 2@xmath71 & [email protected] + agc111946 & [email protected] & [email protected] & [email protected] & 3@xmath71 & [email protected] + agc111977 & [email protected] & [email protected] & [email protected] & 1@xmath71 & [email protected] + agc112521 & [email protected] & [email protected] & [email protected] & 1@xmath77 & 45@xmath712 + agc174585 & [email protected] & [email protected] & [email protected] & 1@xmath71 & [email protected] + agc174605 & [email protected] & [email protected] & [email protected] & 4@xmath72 & [email protected] + agc182595 & [email protected] & [email protected] & [email protected] & 2@xmath71 & [email protected] + agc731457 & [email protected] & [email protected] & [email protected] & 1@xmath71 & [email protected] + agc748778 & [email protected] & [email protected] & [email protected] & 2@xmath71 & [email protected] + agc749237 & [email protected] & [email protected] & [email protected] & 3@xmath71 & [email protected] + agc749241 & [email protected] & [email protected] & [email protected] & 3@xmath71 & [email protected] + [ t4 ] ccccc agc110482 & 125 & 0.42 & 16 & 55 + agc111164 & 162 & 0.36 & 21 & 50 + agc111946 & 175 & 0.52 & 17 & 62 + agc111977 & 207 & 0.47 & 29 & 59 + agc112521 & 189 & 0.42 & 15 & 55 + agc174585 & 348 & 0.26 & 11 & 42 + agc174605 & 49 & 0.06 & 11 & 19 + agc182595 & 106 & 0.21 & 10 & 39 + agc731457 & 53 & 0.23 & 12 & 40 + agc748778 & 43 & 0.23 & 10 & 40 + agc749237 & 85 & 0.40 & 16 & 54 + agc749241 & 119 & 0.29 & 13 & 45 + [ t5 ] cccc agc110482 & [email protected] & [email protected] & [email protected] + agc111164 & [email protected] & @[email protected] & [email protected] + agc111946 & [email protected] & @[email protected] & [email protected] + agc111977 & [email protected] & [email protected] & [email protected] + agc112521 & @[email protected] & [email protected] & [email protected] + agc174585 & [email protected] & [email protected] & [email protected] + agc174605 & [email protected] & [email protected] & [email protected] + agc182595 & [email protected] & [email protected] & [email protected] + agc731457 & [email protected] & [email protected] & [email protected] + agc748778 & [email protected] & ... & @[email protected] + agc749237 & [email protected] & [email protected] & [email protected] + agc749241 & @[email protected] & ... & @[email protected] + average & [email protected] & [email protected] & [email protected] + [ t6 ]
|
we analyze the relationships between atomic , neutral hydrogen ( ) and star formation ( sf ) in the 12 low - mass shield galaxies .
we compare high spectral ( @xmath00.82kms@xmath1ch@xmath1 ) and spatial resolution ( physical resolutions of 170pc700pc ) imaging from the vla with h@xmath2 and far - ultraviolet imaging .
we quantify the degree of co - spatiality between star forming regions and regions of high column densities .
we calculate the global star formation efficiencies ( sfe , @xmath3/@xmath4 ) , and examine the relationships among the sfe and mass , column density , and star formation rate ( sfr ) .
the systems are consuming their cold neutral gas on timescales of order a few gyr . while we derive an index for the kennicutt - schmidt relation of @[email protected]@xmath70.04 for the shield sample as a whole , the values of @xmath5 vary considerably from system to system . by supplementing shield results with those from other surveys , we find that hi mass and uv - based sfr are strongly correlated over five orders of magnitude .
identification of patterns within the shield sample allows us to bin the galaxies into three general categories : 1 ) mainly co - spatial and sf regions , found in systems with highest peak column densities and highest total masses ; 2 ) moderately correlated and sf regions , found in systems with moderate column densities ; and 3 ) obvious offsets between and sf peaks , found in systems with the lowest total masses .
sf in these galaxies is dominated by stochasticity and random fluctuations in their ism .
|
recent studies revealed that a significant fraction of stars in the present - day galaxies were formed between redshifts 1 and 3 ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for example , results of deep near - infrared observations show the average stellar mass density at @xmath9 is @xmath10 of that in the present universe , while at @xmath11 it is just @xmath12 of the present value @xcite .
although such studies demonstrate the total star formation rate ( sfr ) density in the universe as a function of cosmic time , it is still uncertain which populations of galaxies contribute to the active star formation between redshift 1 and 3 .
the evolution of the stellar mass - sfr relation constrains theoretical views of how galaxies accumulate the stellar mass @xcite .
therefore , as a next step , in order to understand the build up of stellar mass in galaxies at this redshift range , it is crucial to examine the relationship of sfr to stellar mass .
sfrs are often estimated from photometric indicators that are closely related to the emission from massive stars @xcite .
massive stars have short lifetimes , so they represent stars which were born recently . the uv continuum ( 15002800 )
directly reflects the number of massive stars , and therefore can be used as an indicator of sfr for galaxies in this redshift range . additionally , far - infrared emission is another indicator of sfr ; the uv emission of young massive stars is absorbed by interstellar dust and reradiated in the far - infrared at @xmath1381000@xmath2 m .
fluxes at 24@xmath2 m wavelength taken with the multiband imaging photometer for _ spitzer _ ( mips ) are commonly used to estimate far - infrared emission of galaxies at high redshifts , because observations in longer wavelengths are too shallow to detect such galaxies .
@xcite showed that @xmath14-selected galaxies at @xmath7 have a tight correlation ( @xmath15dex ) between stellar mass and sfr estimated from uv and ir luminosities .
the sfrs of these galaxies are roughly proportional to their stellar masses .
however , the relation is not still conclusive .
other studies show different relationships between sfrs and stellar masses .
for example , rest - uv selected bm / bx galaxies also have a correlation between these values but the correlation is not proportional @xcite .
furthermore , the sfrs of 24@xmath2m - selected galaxies do not depend on their stellar masses @xcite .
we need to note that sfrs estimated from uv and ir luminosities have large uncertainties .
uv light is affected by dust extinction and stars with long lifetimes of @xmath16gyr . on the other hand
, 24@xmath2 m radiation corresponds to the rest - frame 8@xmath2 m wavelength for galaxies at @xmath7 .
it is known that the far - infrared luminosity estimated from rest-8@xmath2 m luminosity has a large uncertainty due to the complex sed at those wavelengths @xcite .
luminosity is one of the most reliable indicators of sfr .
first , the emission line is reemitted from hydrogen gas ionized by uv photons shorter than 912 , which are radiated by only the most massive stars .
these stars have shorter lifetimes than stars emitting the uv continuum above 1500 .
second , the emission line is less affected by dust extinction than uv emission .
the amount of dust extinction is larger at shorter wavelengths .
third , the h@xmath3 luminosity can be measured without the uncertainty inherent in sed estimates for the far - infrared luminosity .
however , it is difficult to measure h@xmath3 luminosities of galaxies at high redshift , because the h@xmath3 emission line at @xmath17 is out of the wavelength range of optical spectrographs . recently , near - infrared spectroscopic studies of sfrs based on the h@xmath3 luminosity for large samples of star - forming galaxies at @xmath7 have been conducted @xcite .
@xcite demonstrate a weak correlation between sfr estimated from h@xmath3 luminosity and stellar mass .
their sample galaxies are selected by @xmath18 color and have redshifts confirmed using a blue - sensitive optical spectrograph .
the observed galaxies are biased toward less attenuated galaxies .
since red galaxies account for the significant fraction of stellar mass density at @xmath7 @xcite , the galaxies at @xmath7 should be re - examined with a technique more sensitive to red galaxies ( e.g. , @xmath14-selected galaxies ) .
the @xmath14-selected star - forming galaxies observed by @xcite showed a flatter sfr - stellar mass correlation than that of @xcite , but the number of @xmath14-selected galaxies with h@xmath3 detection is still rather limited and observations for a larger sample are necessary . in this work ,
we study the sfr and stellar mass of a large number of @xmath14-selected star - forming galaxies at @xmath7 based on near - infrared observations with the multi - object infrared camera and spectrograph ( moircs ; * ? ? ?
moircs has the capability for multi - object near - infrared spectroscopy .
we also utilize the deep near - infrared imaging of the moircs deep survey ( mods ; * ? ? ?
* ) , which covers @xmath19 in the great observatories origins deep survey north ( goods - n ) region . the sample for the present work is selected from @xmath14-selected galaxies in the mods catalog and their properties are studied using mods and publicly available multi - wavelength data of goods - n . throughout the paper ,
we use the ab magnitude system and adopt cosmological parameters of @xmath20 , @xmath21 , @xmath22 .
in order to select star - forming galaxies at @xmath7 , we apply the bzk color criterion for star - forming galaxies ( sbzk ; @xmath23 ; @xcite ) to the @xmath14-selected galaxies with @xmath24 from the mods catalog , whose limiting magnitude is @xmath25 ( @xmath26 , @xmath27 ) .
the bzk diagram of the extracted galaxies is shown in figure [ fig : bzk ] .
@xmath28 and @xmath29 band magnitudes are measured on the f435w and f850lp images of the advanced camera for surveys ( acs ) on board the _ hubble space telescope _ ( hst ) in the course of the goods treasury program ( version 1.0 ; @xcite ) . in total ,
508 galaxies are identified as candidates .
we chose 37 sbzk galaxies from the candidates for the spectroscopic observation . the coordinates and photometry of the target galaxies
are listed in table [ tab : obscat ] .
although two of the sample galaxies ( mods12 - 0125 and mods41 - 0194 ) are not detected in the @xmath28-band , we include these galaxies because their @xmath30 colors suggest that they are sbzk or passive bzk ( pbzk ; @xmath31 ) , and their photometric redshifts are consistent with the redshift range of bzk galaxies .
we did not select on the basis of spectroscopic redshift , although spectroscopic redshifts were previously known for some galaxies in the present sample ( table [ tab : obscat ] ) . to preferentially select actively star - forming galaxies
, we put higher priority on the galaxies identified in the publicly available _ spitzer _ mips source catalog ( the first delivery of the _ spitzer _ data catalog from the goods legacy project ; chary et al . in preparation ) .
the flux limit of the catalog ( @xmath32 ) corresponds to @xmath33 at @xmath34 using the infrared sed templates by @xcite ; thus the galaxies in the mips source catalog are as luminous as ( ultra-)luminous infrared galaxies .
mips sources in the public catalog are identified with 127 of the sample galaxies ( we call them sbzk - mips galaxies , hereafter ) .
we selected 18 sbzk - mips galaxies .
furthermore , we added 19 sbzk galaxies not listed in the mips public catalog to the list of targets ( sbzk - non - mips galaxies , hereafter ) .
it is known that sbzk criterion picks out various types of star - forming galaxies and passively evolving galaxies at @xmath35 @xcite . in order to check that our sample represents the sbzk population
, we compare the @xmath36 color distribution of our target galaxies to that of sbzk - non - mips and sbzk - mips galaxies with @xmath37 in the mods field ( figure [ fig : hist_bzk ] ) .
the @xmath36 color reflects both the age and the amount of dust of a galaxy at this redshift range .
our sbzk - mips sample is reasonable representative of those in the field , according to both @xmath38-test ( 64.65% ) and kolmorov - smirnov ( ks ) test ( 92.46% ) . on the other hand ,
the sbzk - non - mips sample is possibly biased toward blue galaxies ( @xmath38-test : 35.35% , ks - test : 17.74% ) .
red sbzk - non - mips galaxies are relatively rare , so that the scaled number in the redder bin is too small ( @xmath39 ) to provide robust statistics given the present sample size .
we utilize publicly available multi - wavelength data of goods - n to exclude possible agn which may contaminate the optical star - formation emission .
the hard x - ray luminosity at the detection limit of the _ chandra _ deep field - north 2ms point - source catalogs @xcite at @xmath7 ( @xmath40 ) , suggests that the emission lines of the hard x - ray detected sbzk galaxies are powered by an agn rather than massive stars @xcite .
x - ray sources are identified with 62 of the 508 sbzk galaxies .
we exclude these sources from the present sample .
rest - frame infrared colors are often used to identify agn candidates using _ spitzer _ irac bands ( 3.6@xmath2 m , 4.5@xmath2 m , 5.8@xmath2 m , and 8.0@xmath2 m ) ( e.g. , * ? ? ?
* ; * ? ? ?
a galaxy with emission dominated by the stellar population has a strong bump in the sed at 1.6@xmath2 m @xcite .
we confirmed that all of the present non - x - ray galaxies have strong 1.6@xmath2 m bumps in the sed of the irac bands .
the spectroscopic observations for the selected galaxies were performed with moircs @xcite attached to the subaru telescope @xcite during 2007 march 2427 in the course of the near - infrared spectroscopic observation of x - ray sources ( s07a-083 , pi m. akiyama ) .
we used moircs in mos mode with the low - resolution grism ( hk500 ) , which covers 1.32.3@xmath2 m wavelengths .
the spectral resolution is @xmath41 ( @xmath42 ) with 08 slit width .
the 37 galaxies were observed with four slit masks as shown in table [ tab : obslog ] .
one slit mask was used for each night .
the total exposure time was 2.55hour for each mask .
the slit width was fixed to 08 for targets .
most of the slit lengths were 100 or longer , depending on target positions . to monitor the flux variation due to the atmospheric absorption and photon loss at the slit mask , a bright point source was observed simultaneously on each mask , except for mask cdfn3 .
the fwhms of seeing size in the @xmath1 band during the observations were typically @xmath43 . according to the atmospheric attenuation archived in the canada france hawaii telescope web site ,
the conditions were photometric for the first three nights .
for the forth night , thin cirrus covered the sky .
the observation log is summarized in table [ tab : obslog ] .
the observation followed the standard procedure for mos observation with moircs described in @xcite .
the target acquisition was done using 6 to 7 alignment stars distributed over each slit mask .
the telescope was dithered along the slit ( we call it an `` ab pattern '' ) by 30 , to achieve sufficient sky subtraction . the detector was read eight times in non - destructive readout mode to reduce the read out noise down to 16@xmath44 .
the exposure time was set so that most of sky emission lines , which were used later for the wavelength calibration , were not saturated .
the typical exposure times were 300900sec .
the relative position of the alignment star against the slit mask was checked , and then the telescope pointing was adjusted every hour or two without removing the slit masks from the focal plane . the systematic guide error in one hour was measured to be @xmath4504 . at the beginning of each observing night , an a0v type star ( hip53735 ; @xmath46 ; two micron all sky survey point source catalog ; @xcite ) was observed for flux calibration with the same instrument configuration as the targets .
the data reduction was made , using a semi - automatic reduction script ( moircs mos data pipelines ; mcsmdp ) , which follows the standard procedure of iraf : onedspec . at first
, we make a map of bad pixels ; bad pixels are identified as those that show non - linearity .
this includes dead ( always zero count ) and hot ( always high count ) pixels .
cosmic rays are detected by a moving block average filter using iraf : craverage .
each input frame for iraf : craverage is first divided by its dithered pair so that cosmic rays stand out behind the sky background . the cosmic - ray mask was combined with the bad pixel mask .
the values of bad pixels and cosmic - ray pixels are replaced with values interpolated linearly from the pixels along the slit .
sky emission lines were subtracted using the dithered pair frames ( we call the procedure `` a - b subtraction '' ) .
the dome - flat frames for flat fielding were taken with the same slit mask and grism configuration as used for the observation .
for the first - order distortion correction , the database for the imaging mode of moircs was applied to the spectral frames .
the tilt of the spectrum against the detector rows was corrected in the cases when it was measurable in the continuum of bright objects .
next , each spectrum was cut out based on a mask design file .
for wavelength calibration , oh sky emission lines were identified in each frame using a night - sky spectral atlas @xcite .
using the pixel coordinates of the sky emission lines , the sky - subtracted frames were transformed such that the sky lines aligned along the column of the frames and that the wavelength was a linear function of pixel coordinates .
background emission still remained after a - b subtraction due to time variation of oh emission during dithering observations .
the residuals were subtracted by fitting with a quadratic equation along the column . finally , all frames were weighted with the flux of the bright point source and stacked using the weighted average to obtain the final two - dimensional spectra .
the response spectra , including atmospheric transmission and the instrument efficiency , are calculated using the flux standard star to convert the observed count to an absolute flux .
the data of the flux standard star is reduced to a two - dimensional spectrum in the same way for galaxy targets and then the continuum is traced with a polynomial to extract a one - dimensional spectrum with iraf : apall .
we calculate a model spectrum of the a0v flux standard star using spectrum ( spectrum uses @xcite model ) .
the model was scaled to @xmath47 and the spectral resolution of the model was degraded to @xmath41 with a gaussian kernel , so as to give the same hydrogen line width as the observed spectra .
finally , the one - dimensional spectrum of the standard star was divided by the model spectrum to obtain the response spectrum .
emission lines are detected by visual inspection on the flux - calibrated two - dimensional spectra .
the signal - to - noise ratio evaluation described later shows that visual inspection corresponds to a larger than @xmath48 detection .
the detected emission lines were integrated along slit direction within an appropriate aperture giving the largest s / n to obtain one dimensional spectra .
we first consider the strongest emission line as h@xmath3 .
taking account of the wavelength range where we observed and the expected redshift range of sbzk galaxies , it is the most plausible candidate for a strong emission line .
the identification of h@xmath3 is confirmed in all cases when other emission lines , such as [ nii]@xmath496548,6583 , [ oiii]@xmath494959,5007 , and h@xmath50 are also detected at the expected wavelength .
otherwise , we check the consistency of the redshift of the h@xmath3 emission line with that of the sed model fitting using broad - band data ( see [ subsec : sedfit ] ) .
although one of the conventional methods to estimate random error is to obtain the rms of the scattering of background counts at both sides of the object , the slit length for each object is not long enough .
the readout noise is suppressed down to 16@xmath44 by non - destructive readout mode , so that the readout noise is about 1/5 of the background noise .
because the readout noise and the photon noise of objects are negligible , we estimated the background noise by assuming that the random error is dominated by sky - background noise , which follows poisson statistics . in parallel to the reduction of the object spectra , sky - background spectra were reduced without the background subtraction , and the poisson error of each pixel is estimated from them .
we found that the poisson error is systematically lager than the rms error by a factor of @xmath51 , though the scatter is large .
the rms errors can be underestimated , because counts in the two - dimensional spectra correlate with the neighboring pixels due to the sub - pixel shift in distortion correction and wavelength calibration . to measure the flux and central wavelength of the emission lines , we used specfit @xcite , which fits multiple emission lines and continuum simultaneously .
gaussian profiles were used for emission lines .
the continuum was assumed to be a linear function of wavelength .
the flux of [ nii]@xmath52 was fixed to be 1/3 of [ nii]@xmath53 flux . if h@xmath50 and/or [ oiii]@xmath494959,5007 emission lines were detected on the two - dimensional spectra , we measured the flux in the same way .
the random errors of the fluxes were estimated from the best - fit parameters of the model fitting based on the sky - background error described above . the emission lines detected with s / n@xmath543 are listed in table [ tab : obscat ] , and the spectra are shown in figures [ fig : haspec ] and [ fig : o3spec ] .
the measured flux and the redshift derived from the central wavelength of h@xmath3 emission line are shown in table [ tab : sfrcat ] .
it is difficult to accurately evaluate systematic flux error since part of the flux from the object will be lost by a slit mask ; the systematic error due to the flux loss is not negligible if we measure the total flux of emission lines from slit spectroscopy .
the flux loss can be caused by a variety of things , for example , a larger size of the object than the slit width , imperfect alignment of the slit mask to the target , and misalignment of the dithering direction along the slit .
first we estimated the flux loss for a point source . when a gaussian profile with fwhm = 06 , which is the typical seeing during the observations , and a slit width = 08 are assumed , fluxes passing through the slit are 88.4% of the integrated flux ; the aperture correction factor is @xmath55 .
we applied this factor to the point sources observed .
we found that , however , the fluxes of the point sources were still underestimated by up to @xmath56% , even if we also correct the effect of the seeing size variation and pointing error , which are measured from alignment stars .
this fact indicates that imperfect alignment of dithering and/or short - time variation of the seeing size during the observations can also add to the systematic error . on the other hand
, it is expected that the flux loss for extended sources varies with the size of the sources .
we calculated aperture correction factors from the ratio of total flux to that expected to come in through slits , assuming that the regions emitting stellar continuum and h@xmath3 had the same spatial distribution .
the flux through a slit was measured on the @xmath1-band image with a rectangle of the same aperture size as for extracting one - dimensional spectrum and the same position angle of the slit , while the total flux was obtained from mag_auto of sextractor .
the alignment error of the slit was estimated from the alignment stars on the slit mask .
we found the median of the aperture correction factors to be @xmath57 with little dependence on the size of the objects .
the rms scatter of the factor among the objects is @xmath58 .
we apply the median value to all objects to correct the flux loss .
we checked the above aperture correction method by comparing the continuum flux derived from the corrected spectrum with total magnitude , which was measured from @xmath1 or @xmath59-band imaging data according to the redshift of each galaxy .
we used 7 galaxies in the sample whose stellar continuum around the h@xmath3 emission line was detected with sufficient s / n .
we applied the aperture correction to the continuum flux , then contrasted the corrected flux with the total magnitude .
the comparison shows that both results are consistent and the uncertainty in the aperture correction is within @xmath60 .
we need to note that the assumptions on the light distribution of h@xmath3 emission line would also cause systematic errors .
the h@xmath3 and continuum emitting regions would not always be similar for star - forming galaxies at @xmath7 .
studies with the integral - field spectrograph show that h@xmath3 emission does not trace the stellar continuum in some galaxies @xcite . however , this kind of check on our slit - spectroscopy sample is difficult , because both the continuum and emission lines are so faint that the light distribution along the slit is not able to be measured . as a lower limit of the aperture correction , when the h@xmath3 emitting region is point like , the flux becomes smaller by @xmath61 .
the flux ratios of rest - optical emission lines such as [ nii]@xmath626583/h@xmath3 , [ oiii]@xmath625007/h@xmath50 , and h@xmath3/h@xmath50 provide us information on physical properties of ionized regions such as the energy sources for ionization ( [ nii]/h@xmath3 and [ oiii]/h@xmath50 ; @xcite ) , the amount of dust extinction ( balmer decrement , e.g. , h@xmath3/h@xmath50 ; @xcite ) , and gas - phase metallicity ( [ nii]/h@xmath3 ; @xcite )
. however , except for h@xmath3 , the emission lines of the present sample are so faint that we can not evaluate the line ratios of an individual galaxy .
therefore , we stacked the calibrated spectra in the [ nii]-h@xmath3 , and [ oiii]-h@xmath50 wavelength ranges to obtain high s / n spectra , and discuss the averaged physical properties
. the stacked spectrum @xmath63 was calculated with @xmath64 where @xmath65 and @xmath66 are the spectrum and the total h@xmath3 flux of @xmath67 object , respectively .
the spectra were re - sampled in the rest - frame wavelength scale ( @xmath68 ) .
the fluxes were scaled with total h@xmath3 flux , though the variety of fluxes among the samples are not large .
because our main purpose of stacking is to examine the average physical properties of galaxies from their line ratio , weighted mean ( by either inverse square of noise or luminosity ) is not applied ; the line ratios of weighted mean flux could be biased toward strong emission line galaxies .
we selected the objects whose spectra covered the wavelengths of redshifted h@xmath3 , [ nii ] , h@xmath50 and [ oiii ] .
the emission lines of the objects which had at least h@xmath50 or [ oiii ] with @xmath69 were stacked for the measurements of the line ratios of [ nii]/h@xmath3 and [ oiii]/h@xmath50 .
as a result , 9 galaxies were used for the stacking .
the stacked h@xmath3-[nii ] and h@xmath50-[oiii ] spectra are shown in figure [ fig : o3stack ] .
the emission lines of the stacked spectra were fitted with a combination of the same gaussian profiles and linear continuum measurements used in [ subsec : emission ] .
the best - fit spectra are also shown in figure [ fig : o3stack ] .
we also made stacked spectra including the objects with neither h@xmath50 nor [ oiii ] detection , but the resultant line ratio was not changed .
to evaluate the effect of galaxies which would happen to have a significantly different line ratio from the others , we used a bootstrap method .
we built 1000 bootstrap sets , whose galaxies were randomly selected , allowing duplicate sampling from all galaxies used in the stacking , and stacked the spectra of the galaxies in each set to derive the line ratios .
we regard the 16% and 84% quantile of the distribution as 1@xmath70 uncertainty in line ratio . to evaluate the properties of stellar population such as stellar mass , amount of dust extinction ( @xmath71 ) , and age , we use multiband photometric data from the mods @xmath72 images and publicly available images .
the publicly available images we use are : @xmath73(kpno / mosaic ; @xcite ) , @xmath74 ( acs / hst ; goods treasury program version 1.0 ; @xcite ) , 3.6@xmath2 m , and 4.5@xmath2 m ( irac / sst ; dr2 of the spitzer legacy science program for goods ; dickinson et al . in preparation ) .
these images are deep enough to detect our galaxies with sufficient s / n and to obtain the properties of stellar population .
the flux of the galaxies were obtained with @xmath75 for @xmath76 of the objects in all bands , except for @xmath28 ( 60% ) and @xmath73 ( 24% ) bands .
the aperture size for photometry is set to 15 , which is approximately @xmath77 of the psf size in @xmath1 band .
aperture corrections were applied to each band properly ( see * ? ? ? * ) .
the contribution of h@xmath3 emission to the flux in @xmath1 or @xmath59 band , which depends on redshift , was subtracted .
the corrections are 0.01 - 0.34mag with a median of @xmath78mag . we fitted the photometry data with stellar population sed models .
the sed libraries of galaxev @xcite are used for the fitting .
we adopted the libraries with metallicities of 1.0@xmath79 , 0.4@xmath80 , 0.2@xmath80 and a @xcite imf . exponentially declining star formation histories ( @xmath81-model )
are assumed : sfr @xmath82 with @xmath83 0.01 , 0.02 , 0.05 , 0.10 , 0.20 , 0.50 , 1.0 , 2.0 , and 5.0 gyr .
the starburst attenuation law of @xcite was applied to the model with 0.1 steps from @xmath71 = 0.0 to 3.0 .
estimations of the stellar mass are affected by the adopted imf .
many authors employ a @xcite imf , which is known to overestimate stellar mass due to the steep faint - end slope of the imf .
in addition to the @xcite imf , we fitted the seds to the model with the @xcite imf and measured stellar mass for comparison . for our sample ,
the stellar mass with the @xcite imf is larger by a factor of @xmath84 than that with the @xcite imf .
recent evolutionary population synthesis models , for example @xcite , which take account of contributions of thermal - pulsating asymptotic giant branch stars to seds , result in changes in the inferred sed properties . we also fitted the seds with @xcite model and found that the model made the stellar masses smaller by a factor of @xmath85 .
hyperz @xcite was used to determine best - fit sed models .
the redshift for the fitting was fixed to that derived from the h@xmath3 emission line .
we derived total stellar masses of galaxies , scaling the stellar masses and seds with the total magnitude ( mag_auto ) of the @xmath1 band .
the parameter matrix , which gives a @xmath38 value for each grid of sed models , was used to evaluate the uncertainty in the model parameters .
the uncertainty is obtained from the probability distributions of the parameters .
for example , the probability distribution as a function of stellar mass ( @xmath86 ) was calculated from a minimum @xmath38 value for each stellar mass ( @xmath87 ) .
we determined a 68% confidence level of the distribution as 1@xmath70 error of the parameter .
the stellar population parameters and the 1@xmath70 errors derived from the sed fitting are listed in table [ tab : sedcat ] .
some examples of the best fit seds and the probability distributions of stellar mass , @xmath71 , and age are shown in figure [ fig : sed ] .
h@xmath3 emission lines are detected from 23 out of the 37 galaxies ( 62% ) .
the detection rate does not depend on the mid - ir flux .
the rate for sbzk - mips and sbzk - non - mips galaxies are 11/18 ( 61% ) and 12/19 ( 63% ) , respectively .
the spectroscopic redshifts are consistent with their photometric redshifts in the mods catalog @xcite ; the median value of the relative errors is @xmath88 with a scatter of 0.035 , where @xmath89 . the spectroscopic redshifts are newly determined for 10 galaxies . out of the remaining 13 galaxies with spectroscopic redshifts in the literature @xcite , two galaxies have slightly different spectroscopic redshift .
mods32 - 0116 has @xmath90 ( @xmath91 ) and mods21 - 2321 has @xmath92 ( @xmath93 ) .
the left panel of figure [ fig : halum ] shows the apparent h@xmath3 luminosity and the redshift distribution of the h@xmath3-detected sbzk galaxies .
the galaxies lie at redshifts between 1.49 and 2.49 with a median of 2.12 .
the redshift distribution is consistent with that expected for bzk galaxies ( @xmath94 ; @xcite ) .
the lack of galaxies at @xmath95 is due to an atmospheric absorption band between @xmath59 and @xmath14 bands .
the luminosity distances of these objects are calculated with their redshifts measured using the h@xmath3 emission lines .
the fluxes of h@xmath3 emission lines ( @xmath96 ) range between 1 and 17 @xmath97 , which corresponds to an h@xmath3 luminosity ( @xmath98 ) of 0.5 to 7.0 @xmath99 . @xmath96 and @xmath98 are listed in table [ tab : sfrcat ] .
possible reasons for non - detections of emission lines from the remaining 14 galaxies are ; ( 1 ) their h@xmath3 emission lines are too faint for the present observations ; ( 2 ) the redshift of those galaxies are out of the @xmath100 range , where the grism ( hk500 ) we used covers the h@xmath3 wavelength ; ( 3 ) the galaxies lie at the redshift where h@xmath3 emission line falls on strong sky emission lines or atmospheric absorption bands .
because the photometric redshifts and the errors of the galaxies with emission lines undetected are at @xmath101 , the second possibility is unlikely except for mods21 - 2047 , whose @xmath102 is 0.999 . in order to check the third possibility
, we calculate the detection limit of h@xmath3 emission as a function of redshift using the sky background noise and the response spectrum .
a gaussian profile is assumed for h@xmath3 with @xmath103 , which corresponds to the wavelength resolution of the hk500 grism with a 08 slit width .
provided that h@xmath3 emitters have a uniform redshift distribution at @xmath104 and an apparent h@xmath3 luminosity of @xmath105 , which corresponds to the 84% quantiles of the distribution of apparent h@xmath3 luminosity , the expected detection rate of the h@xmath3 emission line with @xmath106 is 87.7% , which corresponds to @xmath107 objects : objects at @xmath108 are rarely detected .
therefore , for 5 of the observed samples , h@xmath3 emission lines are undetected due to the strong sky emission lines or atmospheric absorption bands , while the remaining 9 galaxies have smaller sfr and/or larger attenuation than the galaxies with emission line detection .
diagnostic diagrams based on optical emission - line ratios are commonly used to determine the energy source of ionization .
possible sources are massive stars , agns and energetic shocks @xcite . in this paper
, we use h@xmath3 luminosity as an indicator of sfr , assuming that its energy source is massive stars .
we excluded agn candidates from the sample using multi - wavelength data ( see [ subsec : sample ] ) .
in addition to these diagnostics , we use optical emission - line ratios to determine the energy sources of the present emission line objects . in figure [
fig : bpt ] , we plotted [ oiii]@xmath625007/h@xmath50 and [ nii]@xmath626583/h@xmath3 line ratios of the stacked spectra and the galaxy with the four emission lines detected ( mods31 - 0033 ) .
results for @xmath109 galaxies from sdss dr6 @xcite and local emission line galaxies @xcite are also shown . a classification line empirically determined from the sdss objects @xcite and a theoretically determined classification line @xcite are also shown .
emission - line galaxies located below and to the left of the lines are expected to have massive stars for their energy sources .
the distribution of local sdss galaxies exhibits the two expected sequences : the populations below ( star - forming galaxies ) and above ( agns ) the solid line . the emission line of the stacked sample place it on the theoretical line , while mods31 - 0033 is located to the lower left of the empirical line .
our results are located above the sequence of local star - forming galaxies , closer to the agn sequence .
in fact , a similar offset is seen in star - forming galaxies at @xmath7 @xcite .
one possible reason for the offset is a higher electron density and a harder ionizing spectrum than for local star - forming galaxies @xcite .
we note that agns can be also identified based on [ oiii]-h@xmath3 line ratios for objects without an h@xmath50 detection .
agns typically have [ oiii]/h@xmath50 @xmath110 on figure [ fig : bpt ] . on the other hand
since the h@xmath3/h@xmath50 line ratio must be @xmath111 ( the balmer decrement ) or larger ( if there is a significant dust extinction ) the [ oiii]-h@xmath3 line ratio of an agn candidate will be larger than 1 . only one object ( mods41 - 0194 ) has [ oiii]-h@xmath3 @xmath112 ( 1.67 ) .
when we exclude mods41 - 0194 from the stacking , the stacked line ratios slightly move below the empirical line ( figure [ fig : bpt ] ) .
we regard this object as a possible candidate of agn in the later discussion . in order to evaluate sfr from intrinsic h@xmath3 luminosity
, the extinction due to interstellar dust needs to be corrected .
moreover , the samples are expected to have large amounts of dust , because about half of the present galaxies are luminous infrared galaxies , in which strong infrared emission is thought to originate from dust thermal emission .
the extinction of emission lines in star - forming regions can be estimated with the reddening of the balmer decrement through an extinction curve .
the intrinsic line ratios of the balmer emission lines are nearly constant with respect to electron temperature of ionized gas .
the amount of reddening is estimated from the ratio of the attenuated - to - intrinsic balmer line ratio , @xmath113 where `` i '' and `` o '' denote intrinsic and observed flux , respectively , and @xmath114 is an extinction curve .
we assume @xcite law for @xmath114 .
the balmer decrement method can only be applied to the 14 galaxies at @xmath115 in the present sample , because the h@xmath50 wavelength is not covered by the hk500 grism for galaxies at lower redshift .
both h@xmath3 and
h@xmath50 are detected in four galaxies ( mods11 - 0390 , mods12 - 0255 , mods22 - 1411 , mods31 - 0033 ) , so that @xmath116 is estimated using the above equation . for the rest of the sample , lower limits on the @xmath116
are evaluated using h@xmath50 emission - line upper limits .
the resultant @xmath116 values are @xmath117 .
we also used the stacked spectrum to measure the average amount of reddening and found it to be @xmath118 .
we estimate the amount of reddening for the ionized gas from the sed fitting and apply these extinction corrections to the h@xmath3 luminosities for all of the galaxies .
figure [ fig : extinction ] shows the comparison between the amounts of reddening derived from the balmer decrement and sed fitting .
our result shows that the gas is more attenuated than the stars . from sed fitting ,
the amount of reddening for stars in the samples used for stacking are found to be @xmath119 , thus @xmath120 .
it should be noted that the galaxies which we used to determine the relation suffer relatively less attenuation since we would not be able to detect h@xmath50 emission in the more highly attenuated galaxies .
the hatched region in the figure also shows the lower limit of @xmath116 corresponding to the 3@xmath70 detection limit of h@xmath50 .
the height of the region calculated from the 68% distribution of the observed h@xmath3 fluxes . from studies of local starburst galaxies
, it is known that ionized gas is more attenuated than stars @xcite .
the difference is thought to be caused by changes in the geometry of the ionized gas and stars in a typical galaxy .
@xcite give the relation between the color excess observed for gas and stars : @xmath121 this relation is shown with the dashed line in figure [ fig : extinction ] .
our stacked result is consistent with the equation within errors . on the other hand ,
h@xmath50 detected galaxies are also consistent with the case that extinction for gas and stars is equal ( defined as `` equal extinction '' , hereafter ) , though these galaxies would lie at the lower end of reddening as noted above .
we , therefore , employ equation [ eq : ebv ] as an averaged nature of dust extinction to correct for the effect of dust extinction on the emission lines and compare these with that calculated assuming equal extinction .
we discuss the reliability of the application of these extinction correction further in the [ subsec : extcor ] .
the right panel of figure [ fig : halum ] shows the luminosity of the h@xmath3 emission lines of the galaxies after the extinction is corrected .
the extinction correction factor of our samples ranges from 1.2 to 87 , resulting in a wider spread in extinction corrected h@xmath3 luminosities ( @xmath122dex for the @xcite relation and @xmath123dex for equal extinction ) than in the apparent h@xmath3 luminosities ( @xmath124dex ) .
this result indicates that proper extinction correction is important to reliably calculate the intrinsic h@xmath3 luminosity . to convert the extinction - corrected h@xmath3 luminosity to sfr
, we derive a relation between them using stellar population synthesis models , following @xcite .
we employ galaxev models @xcite with the @xcite imf and the same parameters as used in
[ sec : ana ] to calculate ionizing photons ( @xmath125 ) .
@xmath125 is converted to @xmath126 based on a case - b recombination model @xcite .
if a constant sfr model and solar metallicity are assumed , the @xmath126-sfr relation with an age older than 20myr is given by , @xmath127 using this equation , we derived the sfr with extinction - corrected h@xmath3 luminosity .
the result is listed in table [ tab : sfrcat ] .
the sfr estimate is larger by a factor of @xmath51 , if the @xcite imf is adopted .
the stellar masses derived from the sed fitting are compared with the sfrs calculated using both extinction correction approaches in figure [ fig : sed_ms ] . for the extinction correction using the @xcite relation , the sfrs range from 10 to several hundreds @xmath128 .
some massive galaxies ( @xmath129 ) and a significant fraction of low mass galaxies show sfr larger than @xmath130 . as a result
, there is no clear correlation between sfr and stellar mass .
the present sample contains many sbzk - mips galaxies , whose sfrs are expected to be high .
although sbzk - mips galaxies lie at the massive and high sfr end of the distribution , there still exist low - mass and high - sfr galaxies among sbzk - non - mips galaxies .
the specific sfr ( ssfr ) , which is the sfr divided by stellar mass , is one of the indicators showing the level of star formation ; higher ssfr indicates that a galaxy is accumulating large stellar mass compared to its own stellar mass on a short timescale . in figure
[ fig : sed_ms ] , ssfr is depicted by dashed lines .
ssfr does not depend on mips detection but on stellar mass .
there exist actively star - forming galaxies ( @xmath131 ) among the less - massive galaxies , while the ssfr of massive galaxies ( @xmath132 ) is nearly constant ( @xmath133 ) .
it can imply that some low mass galaxies at @xmath7 are actively forming stars . on the other hand , for the sfrs corrected assuming the equal - extinction case , the sfrs and ssfrs of the high - ssfr galaxies become lower , but they still have higher ssfr when compared to other galaxies .
the ssfr of the original high - ssfr galaxies become @xmath134 . consequently
, there is still no correlation between them .
the distribution of sfrs is nearly constant ( @xmath135 ) .
the possible agn candidate ( mods41 - 0194 ; denoted by a open square ) shows lower sfr than other galaxies with similar stellar masses .
we note that this object does not change our findings of no correlation between sfr and stellar mass and low mass galaxies with active star formation .
the inverse of ssfr ( ssfr age ) gives the estimated age for a galaxy to build up all of its stellar mass with the current sfr . on the contrary , mass - weighted age , which is calculated by integrating the ages of stars weighted by sfr at that time ,
is an indicator of the average age of the stars in a galaxy .
the mass - weighted age is given from integration of the star formation history of galaxy as @xmath136 where @xmath137 is the sfr at age t. comparing the two ages , we can roughly check consistency between the sed fitting result and the sfr estimated from extinction corrected h@xmath3 luminosity . the comparison between the ssfr age for the both cases of extinction correction and mass - weighted age of our sample
is shown in figure [ fig : age_sts ] . for extinction correction with the @xcite relation ,
the ssfr ages agree with the mass - weighted ages to within 1@xmath70 error ; a galaxy with a higher ssfr has a younger stellar population .
this fact suggests that the majority of the stellar mass for the galaxies in our emission line samples were formed in a recent starburst .
on the other hand , for the extinction correction assuming the equal extinction case , the ages of stars are young compared to the sfr especially for high - ssfr galaxies . in this case
the galaxies must have experienced much more active star formation in the recent past and terminated the star formation in a very short timescale ( @xmath138myr ) .
the median value of the reddening derived from the sed fitting is @xmath139 ( @xmath140 ) for sbzk - mips and sbzk - non - mips galaxies , respectively .
both values are relatively larger than those for local galaxies ( e.g. , @xmath141 for @xmath142-band selected sdss galaxies ; @xcite ) .
the extinction corrections are key to the discussion on the luminosities of h@xmath3 emission line , and thus the sfrs and ssfrs ; larger extinction correction makes extinction - corrected sfr and ssfr larger . in this section
, we discuss the reliability of our extinction correction through comparisons of the h@xmath3 sfrs with other sfr indicators , and the reason why the higher ssfr galaxies are found by our observation .
the uv continuum is used as a sfr indicator , because the stellar continuum at 1500 - 2800 is sensitive to emission from massive stars .
we compare the sfrs with those estimated from the uv continuum .
we calculate a relation between uv continuum at @xmath143 , @xmath144 , and sfr as for equation [ eq : hasfr ] .
@xmath144 is measured from the sed templates of galaxev with the same parameters as in [ subsec : sedms ] with the @xmath28-band filter of acs / hst redshifted to @xmath145 .
the relation is given by , @xmath146 for our samples , @xmath144 is measured with @xmath14715 in @xmath28-band with an aperture correction obtained from the @xmath1-band total magnitude .
the dust extinction is corrected with that derived from the sed fit .
the comparison between sfr(h@xmath3 ) and sfr(uv ) is shown in figure [ fig : sfrcompuv ] .
for the equal - extinction case , the sfr(h@xmath3 ) is underestimated for the galaxies with higher sfr(uv ) . on the other hand , using the @xcite relation the sfr(h@xmath3 ) and sfr(uv ) are roughly consistent for a broad range of sfr , but are systematically larger by 0.3dex for both sbzk - mips and sbzk - non - mips galaxies .
we note that the luminosity ratio of the uv continuum to h@xmath3 varies with the population of a galaxy .
for example , because the uv continuum is contaminated by less massive stars than stars emitting ionizing photon , galaxies populated by younger massive stars are more luminous in h@xmath3 . for our template ,
the ratio is larger by factor @xmath148 for age@xmath149myr and becomes constant for age@xmath150myr .
in fact , high - ssfr galaxies have a mass - weighted age@xmath151myr .
if a top - heavy imf is assumed , moreover , the ratio gets larger , so that more ionizing photons are emitted than uv continuum . in order to examine the reliability of the extinction corrections further , we compare the sfr(h@xmath3 ) estimates with those derived from the thermal infrared luminosities ( sfr(ir ) ) , because the infrared luminosity is an sfr indicator almost unaffected by dust extinction .
we use the rest-8@xmath2 m luminosity ( @xmath152 ) derived from the mips 24@xmath2 m flux as representative of the thermal infrared luminosity . independently from the public catalog , the 24@xmath2 m fluxes of the present galaxies are measured using the @xmath153 mips image ( dickinson et al . in preparation ) with the iraf / daophot package @xcite for both sbzk - mips and sbzk - non - mips samples .
some galaxies have 24@xmath2 m fluxes severely contaminated by nearby bright 24@xmath2 m sources because of the large psf of mips ( @xmath154 ) . in order to reduce the effect of the contamination ,
the fluxes are measured by fitting crowded objects simultaneously with the psf measured from isolated point sources .
the objects are centered using the @xmath1-band light distribution .
the photometric error is estimated on a blank - sky region on the image . in order to take into account the residuals of the fitting ,
the image from which the fitted models are subtracted is measured and the residual is added to the error .
then , @xmath155 is estimated from the spectroscopic redshift of h@xmath3 emission line and the observed 24@xmath2 m flux densities with @xmath156-corrections derived from the luminosity dependent sed models of @xcite .
the measured mips-24@xmath2 m flux and estimated rest-8@xmath2 m luminosity are also listed in table [ tab : sfrcat ] .
the derived @xmath157 and h@xmath3 luminosities are plotted in figure [ fig : sfrcomp ] . the h@xmath3 luminosities are corrected using both the @xcite relation ( _ left _ ) and the equal extinction case ( _ right _ ) for comparison . the expected relation between @xmath157 and h@xmath3 luminosity
is also shown using a solid line in the figure .
the relation is derived based on that expected between @xmath152 and sfr .
the @xmath157 is converted to total infrared luminosity ( @xmath158 ) in the wavelength range between 8 and 1000@xmath2 m with the luminosity dependent sed models @xcite , and the relation between the total infrared luminosity and sfr , @xmath159 is used .
the relation is derived by dividing that in @xcite by 1.5 to convert the @xcite imf to the @xcite imf .
as we see in figure [ fig : sfrcompuv ] , it is suggested that the extinction correction between the stellar continuum and ionized gas depends on the sfr of a galaxy . for higher ( lower )
sfr , the h@xmath3 luminosity corrected with the @xcite relation ( equal - extinction case ) is consistent with @xmath157 .
we divide the present sample into two groups , lirgs and ulirgs , at @xmath160 , which corresponds to @xmath161 . for ulirgs , the systematic offset between sfr(ir ) and sfr(h@xmath3 ) ( sfr(ir)/sfr(h@xmath3 ) )
is 0.31 ( 0.72 ) dex , with a scatter of 0.47 ( 0.29 ) dex for the @xcite relation ( equal - extinction case ) ; although the @xcite relation results better consistency , sfr(h@xmath3 ) is still underestimated . for lirgs , the systematic offset is -0.42 ( 0.02 ) dex with a scatter of 0.29 ( 0.43 ) dex for the @xcite relation ( equal - extinction case ) ; using the equal - extinction case results in a good agreement between the two sfr indicators .
we need to note that sfr estimated from the rest-8@xmath2 m luminosity has a large uncertainty .
rest - frame luminosity at @xmath162 m contains a very small fraction of thermal infrared emission , so that the @xmath158 is estimated by extrapolating the sed to longer wavelengths .
the rest-8@xmath2 m luminosity could also be affected by polycyclic aromatic hydrocarbons ( pahs ) emission lines , whose energy source is dominated by late short - lived b - type stars rather than o - type stars @xcite .
these factors can produce large random error in the estimation of sfr from the rest-8@xmath2 m luminosity . for local galaxies , @xcite demonstrated that the 1@xmath70 error for the @xmath158 estimated from rest-6.7@xmath2 m was a factor of two .
the uncertainty is consistent with the observed scatter around the solid line in figure [ fig : sfrcomp ] .
a correlation between fir luminosity and radio continuum is known for local galaxies @xcite .
non - thermal radio continuum is mainly emitted by synchrotron emission from the remnants of supernovae , whose frequency is proportional to the number of massive stars , and thus the sfr .
the relation between these luminosities for local galaxies is given by @xmath163 with about 40% of uncertainty @xcite .
we use a publicly available 1.4ghz radio catalog and images @xcite , whose 1@xmath70 rms noise in the goods -
n region is 5.8@xmath2jy@xmath164 ( the beam size @xmath165 ) .
the catalog contains sources with a @xmath166 detection .
although our h@xmath3-emission - line galaxies are not matched with the sources in the catalog , some galaxies are marginally detected ( @xmath167 ) , according to our eye - ball check on the 1.4ghz radio images .
therefore , their 1.4ghz fluxes are lower than 29@xmath2jy ( 5@xmath70 ) , and some galaxies possibly have @xmath13 a few 10@xmath2jy .
if we assume the local radio - ir correlation , we can calculate an expected @xmath168 from sfr using equation [ eq : irsfr ] and [ eq : ir_radio ] .
we assume a spectral index of ( @xmath3 ; @xmath169 ) of the non - thermal radio continuum @xmath170 when we calculate the expected 1.4ghz flux density .
the expected 1.4ghz flux estimated from the mips-24@xmath2 m flux exceeds the detection limit only for one emission line galaxy ( mods12 - 0125 ) , even if the uncertainty in the estimation of @xmath158 from rest-8@xmath2 m luminosity is considered .
this object is a sbzk - mips galaxy and has a low ssfr . for this object the sfr(h@xmath3 )
is also smaller than sfr(ir ) , so that only the rest-8@xmath2 m luminosity exceeds that at other wavelengths .
this infrared excess when compare to the radio luminosity can be explained by a warm dust component , which is thought to be caused by a compact nuclear starburst or dust - enshrouded agn @xcite . on the other hand , when we calculate the radio flux from the extinction - corrected h@xmath3 luminosity using the @xcite relation , the radio emission of 5 emission line galaxies ( mods22 - 1411 , mods21 - 2612 , mods22 - 5133 ,
mods11 - 0390 , and mods22 - 2658 ) exceed the detection limit .
the first four galaxies are high - ssfr galaxies .
although most of these galaxies have marginal detection of 1.4ghz flux , the expected radio fluxes ( @xmath171 ) are higher than the observed radio flux , even if the uncertainty in the radio - ir correlation is taken into account . on the other hand , for the h@xmath3 luminosity calculated assuming the equal - extinction case , none of the galaxies exceed the detection limit .
this would suggest that the amounts of extinctions in the ionized gases are overestimated by using the @xcite relation , though the uncertainties in their extinctions are large .
we also note that , equal extinction in turn underestimates the extinction for ulirgs ( see figure [ fig : sfrcomp ] ) .
these h@xmath3-excess galaxies consist of three ulirgs ( mods11 - 0390 , mods22 - 1411 , and mods22 - 2658 ) .
thermal emission from dust heated by massive stars ( @xmath172k ) is observed at sub - millimeter wavelengths .
the goods - n region is also observed with scuba / jcmt ( 850@xmath2 m ; * ? ? ?
* ) , aztec / jcmt ( 1.1 mm ; * ? ? ?
* ) , and mambo / iram ( 1.2 mm ; * ? ? ?
* ) down to @xmath173mjy ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , @xmath16mjy @xcite , and @xmath174mjy @xcite , respectively .
one of our sample , mods31 - 0033 , is identified with an aztec source , azgn27 , whose 1.4ghz and 24@xmath2 m counterparts are also identified by @xcite , although this object is not detected at both 850@xmath2 m and 1.2 mm wavelengths .
interestingly , the 1.1 mm flux ( 2.31@xmath175mjy ) of mods31 - 0033 is consistent with that expected from the 24@xmath2 m flux ( 1.6mjy ) , when the @xcite sed templates are assumed , as though the sfr(h@xmath3 ) is lower than the sfr(ir ) .
none of our samples including mods31 - 0033 are detected at 850@xmath2 m , in spite that some of them have sfr(h@xmath3 ) and/or sfr(ir ) exceeding several hundred @xmath128 .
it is known that 850@xmath2 m selected galaxies seem to have cooler thermal emission than local ulirgs , which results from the fact that the 850@xmath2 m flux is lower than that expected from total ir luminosity @xcite .
therefore , it is possible that 850@xmath2 m emission is not detected from all ultra - luminous starburst galaxies at @xmath7 . the linear relation between x - ray luminosity and sfr is calibrated with local galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
x - ray emission of star - forming galaxies is dominated by high mass x - ray binaries ( hmxb ) , whose progenitor masses are @xmath176 and thus they are short - lived .
the correlation of x - ray luminosity with uv @xcite and h@xmath3 @xcite luminosity have also been investigated for high - z galaxies . due to the sample selection , the present samples are not detected in the _ chandra _ 2ms catalog , whose limiting flux is @xmath177 in soft x - ray band ( 0.52.0kev ) , which detects the redshifted hard x - ray fluxes of objects at @xmath7 .
according to the calibration by @xcite , the flux corresponds to sfr(x - ray)@xmath178 for our galaxies ( the imf is converted to the @xcite imf ) . even if the errors of the sfrs are taken into account , five of our samples ( mods11 - 0390 , mods21 - 2612 , mods22 - 1411 , mods42 - 0112 ,
mods22 - 2658 ) have the extinction corrected sfr(h@xmath3)s with the @xcite relation larger than the upper limit of the sfr(x - ray ) .
the first four galaxies are high - ssfr galaxies , and the galaxies except mods42 - 0112 are also h@xmath3 excess galaxies against radio flux .
this would also suggest that the amounts of extinctions are overestimated by using the @xcite relation .
we note that @xcite reported the higher h@xmath3 luminosities of their bm / bx galaxies than the x - ray luminosities .
the evolution of the first massive stars to hmxb takes @xmath179myr , which is comparable to the age of our high - ssfr galaxies .
extinction and age could be degenerate in sed fitting , because both parameters make the sed redder .
for example , degeneracy in extinction and age is found in the probability distribution for some galaxies ; see the bottom panels of figure [ fig : sed ] .
however , if we limit the age to be larger than 50myr in the sed fit , for example , then the estimated value of extinction is almost the same as that of the original fit , showing that the degeneracy has only a marginal effect on the estimate of extinction .
the choice of star - formation history applied to the sed model also affects the estimate .
we use @xmath81-models in the fitting , but the constant star - formation ( csf ) model tends to result in larger attenuation . in fact , if we apply the csf model to our galaxies , the @xmath180s for about half of the galaxies with @xmath181gyr become larger by @xmath182 than that with the @xmath81-model , while the @xmath180s are unchanged for younger galaxies . in short
, the csf model gives similar or larger best - fit extinction than the @xmath81-model in general .
although a detailed analysis of the rest - optical continuum would reduce further the possible range of solutions for best - fit parameters @xcite , it is difficult to measure the continuum from the present spectrum data .
our result shows no correlation between stellar mass and sfr .
however , @xcite demonstrated that bzk - selected star - forming galaxies have a tight correlation within 0.2dex using sfr derived from uv and mid - ir .
we compare our result for sfr and stellar mass with the @xcite relation ( figure [ fig : sed_ms_comp ] ) . for the @xcite relation , significant fractions of sbzk galaxies ( 4/18 and 4/19 for sbzk - mips and sbzk - non - mips ,
respectively ) deviate from the correlation .
massive galaxies ( @xmath183 ) in our sample are distributed around the relation given by @xcite , while some of the less massive galaxies ( @xmath184 ) have much higher sfr than that expected from the correlation . even for the equal - extinction case , which makes the sfrs of the higher sfr galaxies lower ,
there is still no correlation between stellar mass and sfr like that of @xcite .
one of the reasons for the discrepancy could be the different extinction estimates used in @xcite , who assumed a linear correlation between @xmath185 color and @xmath180 derived from their sed fitting result @xcite to estimate extinction - corrected uv luminosities .
in fact , if we apply the @xmath180 values estimated from the @xmath185 color , the extinction corrected sfr(h@xmath3 ) becomes smaller than the @xmath180 values estimated from our stellar population analysis ( figure [ fig : sed_ms_bz ] ) .
especially when the equal - extinction case is applied , the extinction corrected sfr(h@xmath3)s are similar to the @xcite relation . although our @xmath180 values correlate with @xmath185 color , the correlation has larger scatter than that of @xcite ( figure [ fig : ebv_bz ] ) .
the high - ssfr galaxies have larger @xmath180 values than the galaxies with similar @xmath185 color .
the @xcite relation is also supported by @xcite , who shows the averaged properties of extinction from the radio - to - uv luminosity ratio using a large numbers ( @xmath186 ) of sbzk galaxies .
they also find a strong correlation between the stellar mass and the amount of extinction , and thus the correlation accentuates the @xcite relation . in figure
[ fig : ms_ebv ] , the amount of extinction of our galaxies is shown as a function of the stellar mass .
our results show a larger scatter than the errors of the correlation by @xcite . although the extinction values of the low - ssfr galaxies correlate with the stellar masses , consistent with that of @xcite , the high - ssfr galaxies have higher extinctions than the low - ssfr galaxies with similar stellar masses .
because our emission line galaxies would be biased towards h@xmath3 bright systems , galaxies that deviate from the average relation shown by @xcite can be found , which probably results in the larger scatter .
@xcite also shows the same correlation using a sfr indicator unrelated to extinction .
they estimated the sfr from the sum of the extinction uncorrected uv and mid - ir luminosities .
on the other hand , @xcite reported that there exist significant fraction of lower mass ( @xmath187 ) galaxies with high ssfrs ( @xmath188 ) at @xmath7 using sfr inferred from the 24@xmath2 m flux .
@xcite argues that the correlation is not accurately recovered with sfrs inferred only from mid - ir , due to a substantial presence of mips - excess galaxies , which are excluded from their sample .
a mips - excess galaxy could harbor a heavily obscured agn @xcite .
however , our results on the sfr inferred from h@xmath3 do not recover the correlation even if mips - excess galaxies are excluded from the sample .
we calculate the mips excesses of our sample following @xcite , and found that only a fraction of the high - ssfr galaxies show mips - excess . in figure
[ fig : sed_ms_comp ] , mips - excess galaxies are denoted by white dots , which found to be more typically low - ssfr galaxies rather than high - ssfr galaxies .
previous studies on the relation between sfr and stellar mass using the sfr derived from the h@xmath3 luminosity show a similar trend to ours .
the sfrs and stellar masses of bm / bx galaxies @xcite and sbzk galaxies @xcite are also plotted on figure [ fig : sed_ms_comp ] .
we convert the sfr and stellar mass of these studies to those calibrated with a @xcite imf . in the literature
, they applied different extinction corrections to estimate the h@xmath3 luminosities from that used to estimate the stellar extinction .
the equal - extinction case is applied to the bm / bx galaxies , while the @xcite relation to the sbzk galaxies . for comparison
, we applied the same extinction correction to the samples in the literature as that for the present sample in each panel .
there are no correlations between sfr and stellar mass of the samples in the both panels . however , some high - ssfr galaxies with ssfr@xmath188 appear when the @xcite relation is applied . in the studies for bm / bx and sbzk galaxies ,
validity of the extinction correction is confirmed with the consistency between sfr(h@xmath3 ) and sfr(uv ) @xcite .
we also confirmed the consistency between them in the previous section . the extinction corrected sfr(h@xmath3 )
values assuming the equal - extinction case are consistent only for lower sfr galaxies ( sfr@xmath189 ) .
sfrs calculated assuming the @xcite relation are broadly consistent , but the sfr(h@xmath3)s are systematically larger than the sfr(uv)s by 0.3dex .
these are almost consistent with the previous sfr(h@xmath3 ) works .
although @xcite applied only the equal - extinction case , galaxies with sfr(uv ) @xmath190 are rare in their sample . on the other hand , @xcite , who used the @xcite relation , also showed that their sfr(h@xmath3 ) values are systematically larger than sfr(uv ) by a factor of 3 .
recently , the relation between balmer decrement and stellar continuum attenuation has been investigated for galaxies at low - to - medium redshift .
@xcite studied spectroscopically confirmed @xmath14-selected sample at @xmath191 .
their extinctions from the balmer decrement and the continuum extinctions are equal on average . on the other hand , @xcite studied nearby galaxies , including starburst galaxies such as infrared - luminous galaxies .
comparison between their extinctions shows large scatter , but the average stellar - to - nebular extinction ratio is consistent with the @xcite relation .
the dust properties of the samples in these two studies are quite different .
for example , the median reddening of @xcite sample is @xmath192 , while about half of the galaxies of @xcite have @xmath193 .
the values of reddening of the high - z samples discussed above and our samples are similar to that of @xcite , but vary with the sample selection .
the median values of reddening are @xmath194 for bm / bx galaxies by @xcite and sbzk galaxies by @xcite , respectively .
this can imply that the @xcite relation gives better estimation of the value of reddening for ionized gas in the active star - forming galaxies at high redshift .
the gas - phase metallicity reflects the past star - formation history of a galaxy since the heavy elements are produced by nuclear fusion in stars and fed back into gas .
therefore , it is expected that the gas - phase metallicity of galaxies in the high - z universe is lower than that in the local universe .
the metallicities were found to be lower in high - z galaxies than those in local galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we measured the gas - phase metallicity of our samples to investigate their star - formation histories .
the gas - phase metallicity is estimated from the [ nii]@xmath626583-h@xmath3 line ratio ( n2 index ) , defined by @xmath195\lambda 6583/\mathrm{h\alpha}\}},\ ] ] which correlates with oxygen abundance @xcite .
the relation between the n2 index and gas - phase oxygen abundance is given by @xmath196 @xcite .
[ nii ] and h@xmath3 are so close to each other in wavelength that the ratio is not affected by reddening .
therefore , the n2 index has been used to estimate the oxygen abundances of distant galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we note that the n2 index saturates at @xmath197 ( @xmath198 * ? ? ?
in order to gain a sufficiently high s / n to reliably measure the line flux ratio , we stacked the [ nii]-h@xmath3 emission lines .
we divided the present sample into three groups based on stellar masses at @xmath199 and @xmath200 .
in addition , we also divide the sample into two groups based on a ssfr of @xmath201 .
the stacking , measurement of flux , and estimates of error are performed in the same manner as described in [ subsec : stack ] .
the results are plotted in figure [ fig : mzr ] as a function of the median stellar mass of each group . for comparison , @xmath202 local objects from sdss dr4 , local lyman
break analog ( lba ) galaxies @xcite , bm / bx galaxies @xcite , and sbzk galaxies @xcite are also plotted with the same calibration .
the mass - metallicity properties of our high - ssfr galaxies are possibly different from that of rest - uv selected galaxies .
the mass - metallicity relation of the two largest mass bins is consistent with that of bm / bx galaxies described by @xcite , while the smallest mass bin has higher metallicity than the similar mass bins of @xcite ; our sample is closer to the local galaxies .
the discrepancy in the smallest mass bin could be due to the galaxies with high ssfr .
the high - ssfr group has higher metallicity , close to that of the local galaxies with similar stellar mass .
as shown in figure [ fig : sed_ms ] , high - ssfr galaxies dominate our sample in the smallest mass bin , while most of the galaxies in @xcite have ssfr smaller than @xmath201 .
the @xmath14-band selected samples would provide many galaxies with high ssfr and metallicity .
the existence of higher metallicity galaxies in @xmath14-selected sample is also supported by the observations of sbzk galaxies by @xcite , as shown in figure [ fig : mzr ] . on the other hand , the local lbas , which are local analogs of high - z uv - selected galaxies , tend to have a lower mass - metallicity relation .
this would suggest that the sample by @xcite is biased toward lower metallicities due to the rest - uv selection .
the results indicate that in order to reveal the unbiased cosmic evolution of mass - metallicity relation , we need to evaluate metallicities of a larger sample of @xmath14-band selected ( i.e. , mass - selected ) galaxies .
it is surprising that the high - ssfr galaxies have already acquired metallicities as high as local galaxies at @xmath7 .
the age of the stellar population of the high - ssfr galaxies is young , so that the metal enrichment must have occurred in a short timescale .
the n2 index is known to be sensitive to a hard ionizing radiation field from an agn or strong shock excitation and ionization parameter @xcite .
in fact , our stacked [ oiii]@xmath625007/h@xmath50 and [ nii]@xmath626583/h@xmath3 line ratios show evidence for a harder ionizing spectrum .
if the metallicity is really high , the result is in conflict with the observations of local galaxies .
these galaxies can not keep their stellar masses and metallicities after the observed epoch , because the studies for local galaxies demonstrate that galaxies in a similar mass range formed a significant fraction of their stars in the low redshift universe @xcite , suggesting that these local galaxies must have increased their metallicity recently .
it is possible that such high - ssfr galaxies evolved into larger stellar mass galaxies with low star - forming efficiency in the local universe , depleting the gas and/or merging with higher mass galaxies .
we estimate the contribution of the observed galaxies to the cosmic sfr density at @xmath7 with the sfr(h@xmath3)s , dividing the sample into two groups with ssfr @xmath188 and @xmath203 for the extinction corrected sfrs with the @xcite relation .
since we observed only a limited number of sbzk galaxies , we need to assume that the sfr and redshift distributions of the other galaxies are the same as the observed galaxies ; this is not unreasonable since we randomly selected the present sample from the sbzk - mips and sbzk - non - mips population .
first , we further divide the sample based on the mips detection in the public catalog .
second , we calculated the total sfr in a certain group ; we multiply the total sfr in the group by the ratio of the number of the entire galaxies in the group to that of the number of galaxies observed . in the calculation of total sfr
, we assumed that the sfr of the galaxies without h@xmath3 detection is negligible ; we set those sfrs to zero .
then the total sfrs of galaxies with and without mips detection are summed .
third , we divide the total sfr by the expected detection rate of h@xmath3 emission line ( 87.7% ; see [ subsec : haemission ] ) to recover the h@xmath3 emission lines not detected due to strong sky emission lines or atmospheric absorption bands .
finally , we divide the total sfrs by comoving volume of the survey region for @xmath204 .
the resultant total cosmic sfr densities of the sbzk galaxies are @xmath205 , @xmath206 @xmath207 for extinction correction calculated using the @xcite relation and the equal - extinction case , respectively . as we see in the previous discussions , the relation between extinction estimated from the stellar continuum and the emission lines probably depend on the intrinsic luminosities , and thus the sfrs . to examine this effect ,
when we apply extinction correction with the @xcite relation for the sbzk - mips galaxies and equal extinction for the sbzk - non - mips galaxies ( differential correction ) , the result is @xmath208 . in any case of the extinction correction ,
high - ssfr galaxies have a large contribution to the sfr density .
although these galaxies account for only @xmath209 of all of the galaxies , they contribute to @xmath210 of the sfr density for the extinction correction with the @xcite relation and the differential extinction . even for the equal - extinction case , they contribute to no less than @xmath211 of the sfr density .
we compare our sfr density at @xmath7 with previous works ( e.g. , @xcite and the references therein ; @xcite ) .
if we use the same calibration of imf as that of these studies , the sfr densities are 0.203 , 0.044 , 0.134@xmath212 for the @xcite relation , the equal - extinction case , and the differential correction , respectively . the values compiled by @xcite and @xcite
are 0.05 - 0.38@xmath207 , which are consistent with both of those with the @xcite relation and differential extinction .
recent studies on the sfr density at @xmath7 using ir luminosities ( 850@xmath2 m ; @xcite ; 24@xmath2 m ; @xcite ) show slightly lower sfr densities ( 0.08 - 0.09@xmath212 ) , which are still consistent with the differential extinction correction .
this would also suggest that the relation between the dust properties of the stellar continuum and the emission lines are different depending on the intrinsic sfr .
the sfr density using the equal - extinction case is lower than that of any study .
the lower sfr density implies that applying the equal - extinction case to all galaxies underestimates sfr(h@xmath3 ) .
we performed near - infrared spectroscopic observations for 37 @xmath0 color selected star - forming galaxies with moircs on the subaru telescope .
the sample galaxies are picked from the deep @xmath1-selected catalog of mods .
we select 18 of our sample from the publicly available 24@xmath2 m source catalog of mips on board the @xmath153 space telescope . in order to avoid contamination by agns
, we exclude the bzks detected in the _ chandra _ deep field - north 2ms point - source catalogs .
all of the sample galaxies have strong 1.6@xmath2 m bumps in their seds , so that their optical emission is thought to be dominated by the stellar continuum .
the h@xmath3 emission lines are detected from 23 galaxies , of which the median redshift is @xmath213 . in the diagnostic diagram with [ oiii]@xmath625007/h@xmath50 and [ nii]@xmath626583/h@xmath3 ,
the line ratios of the stacked spectra are above that found for local star - forming galaxies but closer to the expectations from star - formation activity rather than agn activity .
we derived the sfrs from the extinction - corrected h@xmath3 luminosities .
the extinction correction for the h@xmath3 luminosity is estimated from the sed fitting of multi - band photometric data with stellar population models of exponentially declining sfr and the @xcite extinction curve . to estimate the extinction correction for the ionized gas
, we assumed the @xcite relation ( @xmath214 ) as well as the equal - extinction case @xmath215 , and compared the extinction corrected sfrs .
although the balmer decrement of the stacked emission lines is consistent with the @xcite relation , a fraction of the h@xmath50 detected galaxies are consistent with the equal - extinction case .
we further investigated these galaxies by comparing the extinction corrected h@xmath3 luminosities with other sfr indicators .
the comparison with the uv and mips 24@xmath2 m flux shows that the sfrs of lower sfr galaxies ( @xmath216 ) are in good agreement for the equal - extinction case , while those of the higher sfr galaxies agree for the @xcite relation .
this implies that the relation between the dust properties of the stellar continuum and emission lines are different depending on the intrinsic sfr ( differential extinction correction ) .
if the calzetti relation is assumed , a fraction of the high - sfr galaxies should be detected in radio , ( sub-)mm , and x - ray catalogs , but none of them except one aztec source are detected .
we could not explain the reason for this discrepancy consistently .
however , most of the discrepancy could be explained by the short timescale of star - formation activity , because the younger ages of the stellar population are derived from our sed analysis .
we note that other mechanisms ( e.g. , top - heavy imf , lower dust temperature , etc . )
could also explain a part of the discrepancy .
we compared the extinction - corrected h@xmath3 luminosities with the stellar masses obtained from the sed fitting .
when we apply the @xcite relation , the comparison shows no correlation between sfr and stellar mass . a fraction of the galaxies with stellar mass smaller than @xmath4 show sfrs higher than @xmath5 .
the ssfrs of these galaxies are higher than @xmath6 .
even if equal extinction is applied , their ssfrs are still high ( @xmath134 ) .
this is inconsistent with the correlation demonstrated by @xcite .
our value of extinction correction shows large scatter against @xmath185 color and stellar mass , with which @xcite and @xcite show strong correlations . from the best - fit parameters of the sed fitting for these high - ssfr galaxies , we found that the average age of the stellar population is younger than 100myr , which are consistent with sfr age ( @xmath217 ) with the @xcite relation .
we also calculate the contribution of the present h@xmath3 emission to the cosmic sfr density at @xmath7 .
our results are consistent with previous works , if we apply the @xcite relation or differential extinction correction .
the high - ssfr galaxies significantly contribute ( @xmath218 ) to the cosmic sfr density .
the metallicity of the high - ssfr galaxies , which is estimated from the n2 index , is larger than that expected from the mass - metallicity relation of uv - selected galaxies at @xmath7 by @xcite .
although metallicities estimated from the n2 index is known to have significant uncertainties due to the harder ionizing radiation field and ionization parameter , the high metallicity comparable to galaxies in the same mass range is inconsistent with the known star - formation history of such galaxies in the local universe , which formed a significant fraction of their stars in the low redshift universe .
the high - ssfr galaxies , therefore , possibly evolve into larger stellar mass galaxies with low star - forming efficiency , depleting the gas .
we thank the anonymous referee for valuable comments .
we thank the staff of the subaru telescope for their assistance with the development and observation of moircs .
ma is supported by the ministry of education , culture , science and technology ( mext ) grant - in - aid for young scientist ( b ) , 18740118 , 2006 - 2008 .
dma is supported by the royal society and leverhulme trust for financial support .
some of the data presented in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
support for mast for non - hst data is provided by the nasa office of space science via grant nag5 - 7584 and by other grants and contracts .
this work is based in part on archival data obtained with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
this work makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
this work makes use of archival data provided by the sloan digital sky survey ( sdss ) .
the sdss web site is http://www.sdss.org/. osterbrock , d. e. 1989 , research supported by the university of california , john simon guggenheim memorial foundation , university of minnesota , et al .
mill valley , ca , university science books , 1989 , 422 p. ,
|
we present the results of near - infrared multi - object spectroscopic observations for 37 @xmath0-color - selected star - forming galaxies conducted with moircs on the subaru telescope .
the sample is drawn from the @xmath1-band selected catalog of the moircs deep survey ( mods ) in the goods - n region .
about half of our samples are selected from the publicly available 24@xmath2m - source catalog of the multiband imaging photometer for _ spitzer _ on board the _ spitzer space telescope_. h@xmath3 emission lines are detected from 23 galaxies , of which the median redshift is 2.12 .
we derived the star formation rates ( sfrs ) from extinction - corrected h@xmath3 luminosities .
the extinction correction is estimated from the sed fitting of multi - band photometric data covering uv to near - infrared wavelengths .
the balmer decrement of the stacked emission lines shows that the amount of extinction for the ionized gas is larger than that for the stellar continuum . from a comparison of the extinction
corrected h@xmath3 luminosity and other sfr indicators we found that the relation between the dust properties of stellar continuum and ionized gas is different depending on the intrinsic sfr ( differential extinction ) .
we compared sfrs estimated from extinction corrected h@xmath3 luminosities with stellar masses estimated from sed fitting .
the comparison shows no correlation between sfr and stellar mass .
some galaxies with stellar mass smaller than @xmath4 show sfrs higher than @xmath5 .
the specific sfrs ( ssfrs ) of these galaxies are remarkably high ; galaxies which have ssfr higher than @xmath6 are found in 8 of the present sample . from the best - fit parameters of sed fitting for these high ssfr galaxies , we find that the average age of the stellar population is younger than 100myr , which is consistent with the implied high ssfr .
the large sfr implies the possibility that the high ssfr galaxies significantly contribute to the cosmic sfr density of the universe at @xmath7 .
when we apply the larger extinction correction for the ionized gas or the differential extinction correction , the total sfr density estimated from the h@xmath3 emission line galaxies is @xmath8 , which is consistent with the total sfr densities in the literature .
the metallicity of the high - ssfr galaxies , which is estimated from the n2 index , is larger than that expected from the mass - metallicity relation of uv - selected galaxies at @xmath7 by @xcite .
|
it has long been recognized @xcite , that standard candles constructed from long duration gamma - ray bursts ( grbs ) would have several potential advantages over type ia supernovae ( sne ia ) , the most important being high redshift detection . whereas detected sne ia are currently spectrally classifiable out to a maximum of @xmath8 with _ hst _
@xcite ( and in the future with _ snap _
@xcite ) , @xmath9 of grbs with known @xmath3 ( @xmath10 of @xmath11 ) already have measured redshifts @xmath12 .
although there are diminishing returns for observations at higher redshifts which primarily probe the matter - dominated regime such measurements may be of great interest if the dark energy shows exotic time variation . in practice ,
@xmath13 ( @xmath14 of @xmath15 ) grbs in the current sample of bursts with measured redshifts ( @xmath3 ) , jet break times ( @xmath16 ) , and peak energies ( @xmath1 ) , are in the redshift range @xmath17 , which is _ already _ comparable to the number of high - z sne ia discovered with _ hst _ @xcite .
this regime is clearly important for constraining @xmath5 , @xmath6 , its possible time variation , and the transition redshift to the epoch of deceleration @xcite .
in addition to high-@xmath3 detection , @xmath18rays penetrate dust , grb spectra @xcite are simpler than sne ia spectra , yielding potentially cleaner @xmath19-corrections @xcite , and with massive star progenitors @xcite , any long grb evolution would likely be orthogonal to that of sne ia , ensuring different systematics .
grb standard candles could thus provide a useful independent check to cosmography with sne ia .
previous attempts to constrain cosmological parameters with grb energetics @xcite were thwarted by what are now known as wide distributions in the isotropic equivalent energies ( @xmath20 ) and beaming - corrected energies ( @xmath0 ) , which span more than @xmath21 and @xmath22 orders of magnitude , respectively @xcite . in particular , the once - promising @xmath0 distribution @xcite , widened with the discovery of new low energy bursts ( e.g. 030329 ) . expanding upon the well known @xmath1@xmath20 `` amati '' relation @xcite , ghirlanda recognized that many under - energetic bursts appeared softer in the rest - frame prompt @xmath18ray spectrum than those with higher @xmath0 , discovering the remarkable @xmath1@xmath0 `` ghirlanda '' relation @xcite , which can be cast as a power - law @xmath23 . in @xcite , we confirm this correlation ( fig .
1a herein ) , and demonstrate that although the goodness of fit to a power - law is strongly sensitive to input assumptions ( particularly the circumburst density ) , the relation itself is still highly significant over a range of plausible cosmologies .
as recently suggested @xcite , the relation could be used to create a standardized grb candle with an empirical correction to the energetics similar to the light - curve shape corrections used to standardize the peak magnitudes of sne ia @xcite . however , without knowing the slope of the power law _ a priori _ from physics or from a low-@xmath3 training set , in the cosmographic context , it is imperative to re - fit for the slope of the power law from the data for each cosmology @xcite ( fig .
1a inset ) , lest circularity problems arise ( e.g. @xcite ) .
even so , greater obstacles to cosmography with current data involve small number statistics and the sensitivity of the goodness of fit of the relation to input assumptions @xcite .
@xmath0 relation , plotted from tables 1,2 of @xcite , with only 19 grbs used in the fit ( filled circles ) .
upper / lower limits ( open circles ) are denoted with arrows .
inset shows weak cosmology dependence of slope @xmath24 . *
( b ) * sensitivity of @xmath25 to assumptions for density @xmath26 ( error @xmath27 ) .
data , are from @xcite : left , @xcite : right .
curves are for @xmath28 $ ] from top down .
plot symbols show assumptions from previous work .
independent of data set , @xmath19-cor bandpass , @xmath29@xmath30 cm @xmath31 minimizes @xmath25 .
derived @xmath25 are left : @xmath32 ( 17 dof ) , big triangle ; right : @xmath33 ( 13 dof ) , small diamond . in decreasing importance , discrepancy comes from data references , density assumptions , sample size , @xmath19-cor bandpass . *
( a ) * and * ( b ) * assume @xmath34 , ( @xmath4,@xmath5,@xmath35)=(@xmath36,@xmath37,@xmath36).,title="fig:",width=192 ] @xmath0 relation , plotted from tables 1,2 of @xcite , with only 19 grbs used in the fit ( filled circles ) . upper / lower limits ( open circles ) are denoted with arrows .
inset shows weak cosmology dependence of slope @xmath24 . *
( b ) * sensitivity of @xmath25 to assumptions for density @xmath26 ( error @xmath27 ) .
data , are from @xcite : left , @xcite : right .
curves are for @xmath28 $ ] from top down .
plot symbols show assumptions from previous work .
independent of data set , @xmath19-cor bandpass , @xmath29@xmath30 cm @xmath31 minimizes @xmath25 .
derived @xmath25 are left : @xmath32 ( 17 dof ) , big triangle ; right : @xmath33 ( 13 dof ) , small diamond . in decreasing importance , discrepancy comes from data references , density assumptions , sample size , @xmath19-cor bandpass . *
( a ) * and * ( b ) * assume @xmath34 , ( @xmath4,@xmath5,@xmath35)=(@xmath36,@xmath37,@xmath36).,title="fig:",width=307 ] unlike peak magnitudes of sne ia light curves , computing @xmath0 ( see eq s 1 , 2 of @xcite ) depends on * ( a ) * the cosmology , * ( b ) * a model assumption for the energy structure of the jet , * ( c ) * the effective rest - frame `` bolometric '' bandpass for the grb @xmath19-correction @xcite , and * ( d ) * parameters which are often unknown for most bursts ; namely the ambient ism density @xmath26 ( possibly a stellar wind profile @xcite ) , and the @xmath18ray creation efficiency ( @xmath38 ) , where the same values of @xmath26 , @xmath38 are assumed for all grbs when unknown . assuming @xmath39 cm@xmath31 , @xmath34 ( 20% ) , for our sample of 19 grbs , we find a goodness of fit of @xmath41/dof @xmath42 ( 17 dof ) for the @xmath43 $ ] kev bandpass @xcite . at this conference ,
ghirlanda et al . first reported @xmath44 ( 13 dof ) for their sample of 15 grbs @xcite . in @xcite
, we demonstrate that the discrepancy in @xmath25 arises mainly from different references for individual bursts ( small number statistics ) , and from slightly different input assumptions for @xmath26 , @xmath27 , illustrated here in fig .
poor fits to the relation result in unacceptable fits in the grb hubble diagram , rendering the @xmath45 contours over ( @xmath4,@xmath5 ) meaningless ; however , this sensitivity also allows most outlier bursts in fig .
1a to be made consistent with the relation by changing the density ( or its error ) to otherwise reasonable values @xcite . using the relation for cosmography , in @xcite ,
although the authors claim tight constraints on the matter density ( assuming flatness ) , they fail to self - consistently re - fit the slope of the relation for each cosmology and exclude several outlier bursts ( e.g. 990510 , 030226 ) on grounds not adequately justified .
recently , those authors have improved upon these points in follow - up work @xcite . in @xcite ,
the authors re - fit the relation self - consistently , but did not stress the cosmographic power of grbs alone , instead performing a joint fit with sne ia , claiming that the joint fit is more consistent with flatness than sne ia alone .
however , the analysis in @xcite understresses the fact that grbs alone appear to favor a loitering cosmology @xcite ( although see @xcite ) .
despite improvements , the most recent follow up work @xcite does not address the sensitivity to input assumptions @xcite .
however , by performing simulations of future data @xcite , and developing new statistical techniques @xcite , the recent work had indicated some potentially exciting new directions for grb cosmology .
although a major step forward , a grb standard candle constructed from the @xmath1@xmath0 relation can not yield meaningful constraints on the cosmological parameters for the current data , mainly due to the small sample ( with very few low-@xmath3 events ) and to the strong sensitivity of the goodness of fit of the relation to density assumptions .
selection effects @xcite and other potential systematics @xcite must also be addressed before using grbs for precision cosmography . however , _ swift _ should detect @xmath46 grbs over the next @xmath22 years @xcite . of these
, redshift constraints are expected for a majority of the bursts , either from the on - board broad - band spectroscopy or ground - based follow - up spectra . with early - time light curves from the _
swift _ uvot instrument and a fleet of dozens of ground - based follow - up programs , @xmath16 could be measured for a substantial fraction of these bursts .
unfortunately , future @xmath1 measurements may be hindered by the relatively narrow spectral range of _ swift _ ( [ 15,150 ] kev ) , further strengthening the science case for the ongoing symbiosis with _ hete ii _
, due to its larger [ 30,400 ] kev bandpass . even for a sample including an order of magnitude more grbs with measured @xmath3 , @xmath16 , and @xmath1 , we believe that density constraints will remain the limiting factor for cosmography , since each requires detailed broadband afterglow modeling ( e.g. @xcite )
this is independent of a low-@xmath3 training set or a theoretical prediction that constrains the slope of the relation _ a priori_. future work
will then require using the known information about the _ distribution _ of densities to marginalize over or sample statistically from such a distribution with monte carlo simulations @xcite .
while the current data do not uniquely support a good fit for the relation to a power law , they certainly do not rule out one .
as such , it is still possible that grb standard candles from this relation might place meaningful constraints on the cosmological parameters , most notably the time variation of the dark energy .
however , even if the relation is never able to seriously constrain cosmology , with the beaming - corrected energy @xmath0 , it is more physically motivated than the amati relation @xcite ( also see @xcite ) , it may lend insight into grb radiation physics @xcite , and could help us to identify new classes of grbs ( different progenitors ) @xcite .
independent of its ultimate fate as a potential grb standard candle , the discovery of the @xmath1@xmath0 ghirlanda relation @xcite clearly represents an exciting new direction for the grb field .
|
following our previous work , we conclude that a grb standard candle constructed from the ghirlanda et al .
power - law relation between the geometry - corrected energy ( @xmath0 ) and the peak of the rest - frame prompt burst spectrum ( @xmath1 ) is not yet cosmographically useful , despite holding some potential advantages over sne ia .
this is due largely to the small sample of @xmath2 grbs with the required measured redshifts , jet - breaks , and peak energies , and to the strong sensitivity of the goodness - of - fit of the power - law to input assumptions .
the most important such finding concerns the sensitivity to the generally unknown density ( and density profile ) , of the circumburst medium . although the @xmath1@xmath0 relation is a highly significant correlation over many cosmologies , until the sample expands to include many low-@xmath3 events , it will be most sensitive to @xmath4 but essentially insensitive to @xmath5 and @xmath6 , with some hope of constraining @xmath7 with high-@xmath3 grb data alone .
the relation clearly represents a significant improvement in the search for an empirical grb standard candle , but is further hindered by an unknown physical basis for the relation , the lack of a low-@xmath3 training set to calibrate the relation in a cosmology - independent way , and several major potential systematic uncertainties and selection effects .
until these concerns are addressed , a larger sample is acquired , and attempts are made to marginalize or perform monte carlo simulations over the unknown density distribution , we urge caution concerning claims of the utility of grbs for cosmography and especially the attempts to combine grbs with sne ia .
[ 1999/12/01 v1.4c il nuovo cimento ]
|
in the last few years the increased accuracy of the experimental techniques has renewed interest in nuclear reactions at fermi energies .
exclusive measurements , event - by - event analysis , and a @xmath1 coverage allow a deeper investigation of the evolution of the reaction mechanisms with beam energy and centrality .
new insights into the understanding of the nuclear matter equation - of - state ( @xmath2 ) were gained @xcite .
in particular , recent experimental and theoretical analyses were devoted to the study of the properties and effects of the symmetry term of the @xmath2 ( asy-@xmath2 ) away from saturation conditions @xcite .
indeed , the two - component character of nuclear matter adds some special interest to the dynamics of heavy ion collisions at intermediate energies , between @xmath3 and @xmath4 . in _
central _ collisions isospin distillation is an important effect in multifragmentation of charge asymmetric systems . here
phase separation is driven by isoscalar - like unstable fluctuations , i.e. local in phase variations of proton and neutron densities @xcite .
this leads to a more symmetric `` liquid '' phase of fragments surrounded by a more neutron rich `` gas '' relative to the original asymmetry of the system .
isoscaling phenomena , observed experimentally , provide indications for such a scenario @xcite . in _ semi - peripheral
_ collisions between nuclei with different @xmath0 ratio , isospin dynamics will drive the system toward a uniform asymmetry distribution .
the degree of equilibration , correlated to the interaction time , should provide some insights into transport properties of fermionic systems @xcite , in particular give information on transport coefficients of asymmetric nuclear matter @xcite .
the aim of this work is to investigate the isospin transfer through the neck region in semi - peripheral collisions of asymmetric nuclei at fermi energies .
of particular interest is the role of the density dependence of the symmetry energy in this process .
the isospin transfer was measured for collisions of different @xmath5 isotopes at msu @xcite and interpreted theoretically with the result that the asy-@xmath2 should be rather stiff . in these
works the effect of pre - equilibrium emission , which changes the isospin content of the interacting system , was not analyzed explicitly . here
we will discuss this questions in detail , as well as the different transport processes affecting the final isospin content .
quantitatively , dynamical isospin effects can be properly understood only from microscopic calculations based on transport models .
we will base our study on a stochastic bnv transport model ( see refs .
@xcite for more details on the main ingredients of this approach ) .
we are focusing on the charge asymmetric collision @xmath6 , at @xmath7 bombarding energy , to which we refer as the mixed system , @xmath8 . to investigate the density ( @xmath9 ) dependence we consider here two representative parameterizations of the symmetry energy , @xmath10 : one showing a rapidly increasing behaviour with density , roughly proportional to @xmath11 ( asysuperstiff ) and one where a kind of saturation is observed above normal density ( soft , @xmath12 ) ( see ref.@xcite for more detail ) .
the bnv simulations have been performed for semi - peripheral collisions at impact parameters @xmath13 . in the last two cases the reaction has dominantly a binary character and the charge asymmetry of primary projectile ( target)- like fragments
, @xmath14 ( @xmath15 ) , should provide the essential information about the isospin equilibration rate . at @xmath16 already about @xmath17 of the events
are ternary .
an imf can be formed in the mid - velocity region by neck fragmentation @xcite . for more central events
this mechanism becomes dominant : at @xmath18 , one or two imf s are found in more than @xmath19 of events @xcite .
the fragment formation in the neck region will influence the final isospin distribution of the plf / tlf , and will render considerably more difficult the interpretation of the results .
thus in the following we select only binary events in our analysis .
we define the average interaction time , @xmath20 , as the time elapsed between the initial touching and the moment when @xmath14 and @xmath15 reseparate . from our simulations
we obtain @xmath21 for the impact parameters @xmath22 , respectively .
four hundred events were calculated for each initial condition and for each asy-@xmath2 .
typical density contour plots , at @xmath23 and @xmath24 , are shown in figure [ difusden810 ] .
we note the dynamical evolution of the overlap region : driven by the fast leading motion of @xmath25 and @xmath26 prefragments , the formation of a lower density interface can be clearly observed after around @xmath27 .
an isospin migration , or transport , takes place during this transient configuration of two residues with densities close to the normal one , separated by a dilute neck region .
in contrast , in deep - inelastic collisions at lower energies the isospin equilibration is driven by the n / z difference between the interacting nuclei having a quite uniform density profile without a low density interface until separation @xcite .
@xmath16 ( upper row ) and @xmath28 ( lower row ) collision : density contour plots .
the side of each box is 40 fm . _ _ ] we quantify the degree of equilibration by the isospin transport ( imbalance ) ratio @xcite , defined as : @xmath29 here @xmath30 stands for the projectile - like ( target - like ) fragment .
the quantities @xmath31 refer to the isospin , @xmath32 , or in general to any isospin dependent quantity , characterizing the fragments at separation time , for the mixed reaction @xmath33 , the reactions between neutron rich ( @xmath34 ) , and between neutron poor nuclei ( @xmath35 ) , respectively .
a value of @xmath36 approaching zero is an indication of a larger degree of equilibration .
the extreme cases @xmath37 and @xmath38 correspond to the absence of any isospin transfer . the results of the calculations are shown in figure [ imbalance3 ] for the dependence of @xmath39 on the interaction time @xmath20 for the asysoft ( squares ) and asysuperstiff ( circles ) @xmath2 s .
the figure also shows the experimental values extracted in ref .
@xcite at semi - peripheral collisions at about @xmath23 .
we conclude , as in ref .
@xcite , that an asystiff - like @xmath2 provides a better agreement with the experimental observations .
a significant difference between the two equations of state is evident for larger interaction times , i.e. smaller values of the impact parameter , @xmath40 .
the smaller values of isospin transport ratios for the asysoft @xmath2 point toward a faster equilibration rate . in refs .
@xcite an explanation was based on the observation that below normal density the asysoft @xmath2 has a larger value of the symmetry energy .
therefore an enhanced isospin equilibration will occur if the diffusion takes place at uniform lower density .
we intend to show , that in fact the mechanism of charge equilibration is more complicated due to dynamical evolution of the reaction at these energies .
fast particle emission and density gradients will also play a role . in the next section
we investigate more in detail the various influences on the isospin transfer process . , eq .
( 1 ) , for the asysoft ( full squares ) and asysuperstiff ( circles ) @xmath2 s as a function of the interaction time @xmath20 , corresponding to different impact parameter @xmath41 .
the band between the two solid lines corresponds to the experimental data of @xcite . _ ]
the isospin content of the two residues in a mixed collision system at separation time is determined by the interplay between the particle emission to the gas from each nucleus during the overlap and the transfer of nucleons through the neck .
we thus write simple balance equations : @xmath42 here @xmath43 ( @xmath44 ) and @xmath45 ( @xmath46 ) are the @xmath14 ( @xmath15 ) asymmetry and mass at separation , @xmath47 ( @xmath48 ) and @xmath49 ( @xmath50 ) the initial projectile ( target ) asymmetry and mass
. then @xmath51 ( @xmath52 ) , @xmath53 ( @xmath54 ) are the asymmetries and masses of the projectile / target `` gas '' , i.e. of the pre - equilibrium particles emitted by the projectile ( target ) during the interaction time .
finally @xmath55 ( @xmath56 ) , and @xmath57 ( @xmath58 ) are the asymmetry and mass of all nucleons transferred from projectile ( target ) to target ( projectile ) . in figure [ agasiso ]
we plot the time evolution of the quantities @xmath51 , @xmath52 and @xmath53 , @xmath54 , as well as the values of @xmath55 and @xmath56 for the asysoft and the asysuperstiff @xmath2 for two impact parameters .
we note that @xmath51 is much larger than @xmath47 .
the same is true for the target but the difference is smaller .
thus the pre - equilibrium emission reduces the @xmath0 difference between the two nuclei , competing with the transfer process .
this is also clearly seen in figure [ isoevol ] where the time evolution of the projectile , target , and composite system apparent asymmetry , @xmath59 , @xmath60 and @xmath61 is shown .
the apparent asymmetry is calculated taking into account only the preequilibrium emission but not the isospin transfer inside the matter .
a stronger variation is observed for the neutron rich projectile .
the composite system asymmetry , @xmath62 , is the limiting value toward which the two participants would evolve in complete isospin equilibrium .
in contrast , we show in the figure the final asymmetries @xmath43 and @xmath44 including the transfer processes . comparing the results for the two @xmath2 s in figs . 3 and 4 we see that a more neutron rich composition of pre - equilibrium emission is generated in the asysoft case because below normal density , from where most of the emitted nucleons originate , the neutrons ( protons ) are less ( more ) bound than for the asysuperstiff @xmath2 .
the differences between the two asy-@xmath2 s are strongly reduced at larger impact parameters , as seen in the results for @xmath28 in figures [ agasiso ] and [ isoevol ] , since the interaction times are much shorter .
collision at @xmath16 ( left ) and @xmath28 ( right ) : time evolution of isospin content and mass of preequilibrium particles @xmath51 , @xmath53 ( @xmath52 , @xmath54 ) emitted from the projectile ( top panels ) ( resp .
target , lower panels ) for asysoft ( solid lines ) and asysuperstiff ( dashed lines ) @xmath2 .
the squares ( circles ) indicate the mass and isospin of the nucleons that have migrated from projectile to target @xmath55 ( and vice versa , @xmath56 ) for asysoft ( full symbols ) and asysuperstiff ( empty symbols ) at separation time .
the horizontal lines in the left panels give the initial isospin of the projectile , @xmath63 , and target,@xmath64 , respectively . _
_ ] at @xmath16 ( left windows ) and @xmath28 ( right windows ) : influence of pre - equilibrium emission on isospin content of projectile ( top ) and target ( bottom ) systems .
the curves show the `` apparent '' isospin without transfer , while the symbols denote the real isospin at separation time including also tranfer . dashed lines and open symbols correspond to asysuperstiff @xmath2 and the solid lines and full symbols to the asysoft case . _ ] we next show in figure [ difusionpap3 ] the dependence of the quantities introduced above on interaction time , i.e. impact parameter . here
the projectile ( target ) and gas final asymmetries @xmath43 , @xmath51 ( @xmath44 , @xmath52 ) are confronted with the transferred asymmetries @xmath55 and @xmath56 . for the projectile , that is neutron rich ,
both pre - equilibrium emission and nucleon transfer drive the system toward a more symmetric configuration ( left ) .
the two processes have opposite effects and thus tend to compensate for the target ( right ) .
therefore the projectile asymmetry has a more pronounced deviation from the corresponding initial value in comparison to the target .
collision with asysoft ( top ) and asysuperstiff ( bottom ) @xmath2 as a function of collision time or impact parameter .
left panels : @xmath43 ( black circles ) , @xmath51 ( crosses ) , @xmath55 ( rhombs ) and @xmath56 ( squares ) .
right windows : @xmath44 ( black circles ) , @xmath52 ( crosses ) , @xmath55 ( rhombs ) and @xmath56 ( squares ) .
the horizontal solid lines refer to the initial asymmetry of the projectile ( target ) . _
_ ] we find a a clear dependence on the asy-@xmath2 of the isospin transferred between the two nuclei . in particular , for @xmath65
we observe that @xmath66 we will show in the next chapter that the origin of these inequalities lies in the existence of the low density interface and the density dependence of symmetry energy .
the asysuperstiff @xmath2 favors the neutron migration toward the neck region from both participants .
this explains why simultaneously @xmath67 and @xmath68 .
we will see that for the asysoft @xmath2 this effect is weakened .
the above arguments can be made more explicit by considering that the proton and neutron migration is dictated by the spatial gradients of the corresponding chemical potentials @xmath69 @xcite .
the currents of the two species can be expressed as follows @xmath70 \nonumber \\ j_{p } = - ct \nabla \mu_{p}(\rho_p,\rho_n , t ) = - ct [ \left({\partial \mu_p \over \partial \rho_n}\right)_{\rho_p , t } \nabla \rho_n + \nonumber \\
\left({\partial \mu_p \over \partial \rho_p}\right)_{\rho_n , t } \nabla \rho_p ] \nonumber,\end{aligned}\ ] ] where @xmath71 is a constant . rewriting these expressions in terms of landau parameters , @xmath72 ( @xcite and references therein ) , using :
@xmath73 ( where @xmath74 is the level density ) , we obtain : @xmath75 \nabla \rho + % \frac{\rho}{2 } [ \frac{1+f_0^{nn}-f_0^{np}}{n_n } ] \nabla i ) \nonumber \\
% & = & -d_{n}^{\rho } \nabla \rho - d_{n}^{i } \nabla i \\
% j_{p } & = & - ct ( [ \frac{1}{2}~\frac{1+f_0^{pp}+f_0^{pn}}{n_p } - % \frac{i}{2}~ \frac{1+f_0^{pp}-f_0^{pn}}{n_p } ] \nabla \rho - % \frac{\rho}{2 } [ \frac{1+f_0^{pp}-f_0^{pn}}{n_p } ] \nabla i ) \nonumber \\ % & = & - d_{p}^{\rho } \nabla \rho - d_{p}^{i }
\nabla i j_{n } & = & -\frac{ct}{2 } ( [ n_n^{-1}(1+i ) + f^{\rho}_n + i f^{i}_n ) ] \nabla \rho + \rho [ n_n^{-1 } + f^{i}_n ] \nabla i ) \nonumber \\ & = & -d_{n}^{\rho } \nabla \rho - d_{n}^{i } \nabla i \\ j_{p } & = & -\frac{ct}{2 } ( [ n_p^{-1}(1-i ) + f^{\rho}_p - i f^{i}_p ) ] \nabla \rho - \rho [ n_p^{-1 } + f^{i}_p ] \nabla i ) \nonumber \\ & = & -d_{p}^{\rho } \nabla \rho - d_{p}^{i } \nabla i , \end{aligned}\ ] ] where @xmath76 @xmath77 is the neutron or proton mean - field potential . the second lines in eqs . ( 7,8 ) define @xmath78 and @xmath79 drift and diffusion coefficients due to density and isospin gradients , @xmath80 and @xmath81 , respectively .
the terms @xmath82 and @xmath83 , that appear in the density drift coefficients @xmath80 , can be expressed as @xmath84 @xmath85 where the function @xmath86 depends only on the isoscalar part of the interaction .
we have denoted by @xmath87 and @xmath88 the kinetic and the potential part of the symmetry energy coefficient @xmath89 , respectively . combining eqs .
( [ drho ] ) , one can see that the isovector part of the nuclear interaction enters the coefficients @xmath80 through the derivative of the total symmetry energy @xmath89 . on the other hand the isospin diffusion coefficients @xmath90 are proportional to the quantity @xmath91,\ ] ] and depend , in leading order , on the value of the symmetry energy coeffcient @xmath89 .
various particular situations can be derived from these relations . in symmetric nuclear matter @xmath92 and @xmath93 . in the absence of density gradients
the proton current will flow oppositely and equal in magnitude to the neutron current .
on the other hand , for density gradients only , in asymmetric nuclear matter the proton and neutron currents may have the same direction but assume different values , inducing isospin gradients @xcite .
such a situation can be encountered in semi - pheripheral collisions between identical , charge asymmetric nuclei with the formation of a dilute intermediate region .
for the two asy-@xmath2 s we calculate the coefficients @xmath94 , @xmath95 and @xmath96 .
we plot the ratios @xmath97 in figure [ difus1 ] as a function of the density for a fixed asymmetry @xmath98 .
these values of the asymmetry and density are close to the physical conditions expected for the projectile or target region .
the only negative coefficient is @xmath99 .
the isospin gradients , directed from the projectile to the neck and from the neck to the target , induce neutron and proton flows in opposite directions .
however the ratios of the corresponding coefficients are quite close to unity for the two asy-@xmath2 s and therefore the effects are similar . , @xmath100 and @xmath96 , as a function of the density for fixed asymmetry @xmath98 . _
_ ] since the density gradient is oriented from projectile and target residues to the neck , neutrons and protons migrate from higher toward lower density regions . around and below saturation density @xmath101 and @xmath102 .
these inequalities suggest that more neutrons and less protons migrate from projectile toward neck in the case of asysuperstiff @xmath2 resulting in the formation of a more neutron rich intermediate region .
this is due to the larger value of the derivative of the symmetry energy around normal density and is in agreement with the behavior observed in the numerical simulations , see eqs .
( 4,[inequal ] ) .
we note that for asysuperstiff @xmath2 , in spite of an enhanced isospin migration toward the neck at separation time , the projectile residue is more asymmetric in comparison to the asysoft case .
one reason is that for the asysuperstiff @xmath2 during the pre - equilibrium emission not as many neutrons are removed as for the asysoft @xmath2 . also , as it was shown , for the asysuperstiff @xmath2 more asymmetric matter is also transferred from the target .
we now estimate the effect of isospin transport and pre - equilibrium emission on the transport ratios . within the following approximations @xmath103 where the indices @xmath104 refer again to the mixed ( 124 + 112 ) , the neutron rich @xmath105 and the neutron deficient @xmath106 systems , we arrive at a simplified expression for the isospin transport ratio for the projectile which shows explicitely the dependence on the isospin transport @xmath107 and on the pre - equilibrium emission @xmath108 : @xmath109 with similar approximations the target isospin transport ratio can be expressed as : @xmath110
it is observed that the transport ratios depend on the difference @xmath111 as expected .
however , in contrast to what was assumed in refs .
@xcite , it is seen that they also depend on the pre - equilibrium emission which reduces the absolute value of the transport ratios .
both effects are smaller in the case of the asysuperstiff @xmath2 , for which indeed a larger @xmath112 ratio is obtained .
in this work we have studied processes related to isospin equilibration in semipheripheral collisions at fermi energies and their dependence on the symmetry term of the @xmath2 .
a special feature of these reactions is the development of a low density interface between the two residues .
the neck region is controlling the proton and neutron currents and their direction .
the presence of density gradients also affects the isospin exchange between projectile and target and we have shown that this is sensitive to the density dependence of the symmetry energy .
the neutron to proton ratio emitted during the interaction stage is also influenced by the asy-@xmath2 .
the interplay between the two processes leads to a stronger equilibration for asy - soft @xmath2 , as it is evidenced by the isospin transport ( imbalance ) ratio . actually , in the asy - stiff case
, a larger isospin transfer is observed , due to the presence of density gradients , directed from @xmath14 and @xmath15 towards the neck region .
however , since we are studying binary processes , finally we observe a kind of compensation between the asymmetry of the matter transferred from projectile to target ( @xmath55 ) and from target to projectile(@xmath56 ) .
> from this point of view , to put in better evidence effects due to the presence of density gradients , it would be more appropriate to study events where fragments originating from the neck region are also detected , possibly with their isospin content . in the present study a rapidly increasing symmetry energy at subnormal densities ( asysuperstiff ) appears to be in better agreement with the existing data @xcite .
more recent work also considered momentum dependent interactions , showing that the more repulsive character of the overall dynamics may reduce the symmetry energy stiffness required to reproduce the data @xcite .
the study of the interplay between the effects due to the isoscalar and isovector part of the @xmath2 on isospin transport observables deserves further attention . in conclusion , charge equilibration measurements in semi - peripheral heavy ion collisions at fermi energies provide new independent observables to study the poorly known density dependence of the symmetry term of the nuclear eos .
this is of interest for other properties of asymmetric matter , like neutron skin and isovector collective response in finite nuclei , and may also be important for neutron star crust structures @xcite .
p. danielewicz , r. lacey , and w.c .
lynch , science * 298 * , 1592 ( 2002 ) .
, edited by bao - an li and w. udo schroeder , nova science publishers , inc . , new york,2001 .
v. baran , m. colonna , v. greco , m. ditoro , phys.rep . *
410 * , 235 ( 2005 ) .
v. baran et al .
phys.rev.lett . * 86 * , 5023 ( 2001 ) .
m. colonna , p.chomaz , s. ayik , phys.rev.lett .
* 88 * , 122701 ( 2002 ) . v. baran et al .
, nucl . phys . *
a703 * , 603 ( 2002 ) .
j.margueron and ph.chomaz , phys . rev . * c67 * , 041602 ( 2003 ) .
h. xu et al . , phys.rev.lett .
* 85 * , 716 ( 2000 ) .
e.geraci et al . , nucl .
* a732 * , 173 ( 2004 ) .
uehling and g.e .
uhlenbeck , phys.rev . * 43 * , 552 ( 1933 ) .
hellund and e.a .
uehling , phys.rev . *
56 * , 818 ( 1939 ) .
anderson , c.j .
pethick , k.f .
quader , phys.rev . *
b35 * , 1620 ( 1987 ) .
l. shi and p. danielewicz , phys.rev .
* c68 * , 064604 ( 2003 ) .
m.b.tsang , nucl .
a734 * , 605 ( 2004 ) .
m.b.tsang et al . , phys.rev.lett . *
92 * , 062701 ( 2004 ) .
we note that the experimental impact parameter selection corresponds to rather peripheral collisions ( estimated impact parameters b@xmath1138 fm @xcite ) . in this sense
the appropriate comparison should be with @xmath20 = 110 - 120
fm / c interaction time simulations .
et al . , nucl.phys .
* a642 * , 449 ( 1998 ) .
ph.chomaz , m.colonna and j.randrup , phys . rep . * 389 * , 263 ( 2004 ) .
v. baran , m. colonna , m. di toro , nucl .
a730 * , 329 ( 2004 ) ; e.de filippo et al . ,
phys . rev . * c71 * , 044602 ( 2005 ) .
m.farine , t.sami , b.remaud , f.sebille , z.phys .
* a339 * , 363 ( 1998 ) .
f.rami et al .
, phys.rev.lett . * 84 * , 1120 ( 2000 ) .
l.w.chen , c.m .
ko , bao - an li , phys .
. lett . * 94 * , 032701 ( 2005 ) ; a.w .
steiner , bao - an li , arxiv : nucl - th/0505051 .
r. balian _ from microphysics to macrophysics _ ,
vol ii , springer verlag , berlin 1992 .
we note that this is also the mechanism behind the isospin distillation ( n - enrichment of the gas phase ) in the liquid - gas phase transition for asymmetric matter @xcite .
|
in this paper we investigate isospin transport mechanisms in semi - peripheral collisions at fermi energies .
the effects of the formation of a low density region ( neck ) between the two reaction partners and of pre - equilibrium emission on the dynamics of isospin equilibration are carefully analyzed .
we clearly identify two main contributions to the isospin transport : isospin diffusion due to the @xmath0 ratio and isospin drift due to the density gradients .
both effects are sensitive to the symmetry part of the nuclear equation of state ( eos ) , in particular to the value and slope around saturation density .
keywords : isospin transport ; binary collisions ; neck dynamics ; symmetry energy .
+ pacs numbers : 21.30.fe , 25.70.-z , 25.70.lm , 25.70.pq .
|
mt , light potential depth @xmath0 and @xmath1 .
the blue bar represents @xmath2 .
see text for explanation.,title="fig:",scaledwidth=100.0% ] [ fig : fig0 ] a basic concept for a system driven in a non - equilibrium process by the change of some external parameters is that of work . for mesoscopic systems ,
e.g. colloidal particles or biomolecules , the work spent in such a process becomes a fluctuating quantity which follows a probability distribution of finite width @xcite . during the last two decades , various non - equilibrium work relations , above all the jarzynski @xcite and the crooks relations @xcite ,
have been established and shown to restrict the shape of this probability distribution depending on the underlying specific features of both the system and the non - equilibrium process . from a more fundamental point of view
, these relations refine the second law of thermodynamics at the mesoscopic scale .
further exact statements involving the applied work are rare , even though it has been demonstrated in the framework of stochastic thermodynamics @xcite that the aforementioned relations as well as different ones for other thermodynamic quantities @xcite , namely entropy production and dissipated heat , can be derived from a broader perspective @xcite . experimental tests of non - equilibrium work relations have been carried out for a variety of different systems : for single colloidal particles in time - dependent harmonic @xcite and non - harmonic potentials @xcite , biomolecules in folding - unfolding assays @xcite , mechanical torsion pendulums coupled to a heat bath @xcite , and charge transitions in electronic devices @xcite .
common to all experimental systems studied so far is that they only consist of a small number of degrees of freedom , whose internal interactions are irrelevant . from an experimental point of view
the reason for this is straightforward : controlling all external forces acting on a system of many coupled degrees of freedom during a non - equilibrium process in a well - defined way , i. e. , measuring the applied work , is a huge challenge @xcite . on the other hand , a non - equilibrium phenomenon where the concept of work plays a prominent role
is friction . for atomically flat surfaces sliding against each other , friction results from the interplay between
externally applied forces and the nonlinear interaction of a large number of degrees of freedom making up extended contacts at the interface .
inspired by simplistic models @xcite , a 2d system which has attracted much attention in recent years in the field of tribology in order to investigate in a controlled manner the spatiotemporal dynamics of crystalline surfaces under friction consists of a monolayer of interacting particles suspended in a fluid and mechanically driven through a periodic potential . since experimental realizations @xcite and numerical simulations @xcite of this system have successfully shed light on the mechanisms behind friction , it represents also an appropriate model to investigate the statistical properties of the non - equilibrium work done by a well - controlled external force on a system composed of many interacting degrees of freedom .
so far , this kind of analysis has only been numerically carried out to characterize plastic depinning of interacting particles within a stochastic thermodynamic context @xcite . here
, we experimentally study the fluctuations of the work done by an external force on a mesoscopic system with many coupled degrees of freedom : a crystalline monolayer of magnetically interacting colloidal particles moving on a periodic light field under commensurate conditions
. the dynamics of this system mimics the transition from static to sliding friction , where a solid surface ( the colloidal monolayer ) is driven across a corrugated substrate ( the periodic light potential ) by an external force .
we investigate the role of several parameters which can influence the shape of the work distribution , e.g. the number of particles used to locally probe the properties of the system and the integration time to measure the work .
we find that , when tuning the control parameters to induce particle depinning from the substrate , there is an abrupt change of the shape of the work distribution . while in the completely static and sliding friction regime ,
the work distribution is gaussian , non - gaussian tails show up due to the spatiotemporal heterogeneity of the particle dynamics during the transition between these two regimes .
finally , we discuss the asymmetry of the work distribution within the context of the non - equilibrium fluctuation theorems .
and different applied forces @xmath3 . from left to right : @xmath4 fn , 18.6 fn and 24.8 fn . ( b ) trajectories of 50 particles measured over 200 s at constant applied force @xmath5 fn and different potential depths @xmath6 . from left to right : @xmath7 , @xmath8 and @xmath9.,title="fig:",scaledwidth=100.0% ] [ fig : fig1 ] parts of the experimental setup have been described in detail elsewhere @xcite and will be discussed only briefly . a schematic illustration of the experimental setup is shown in fig .
[ fig : fig0](a ) .
the system consists of a monolayer of superparamagnetic colloidal particles with @xmath10 in diameter ( dyna m-450 epoxy , life technologies ) , suspended in a @xmath11 sodium dodecyl sulfate aqueous solution and situated in a sample cell of @xmath12 height .
the total number of particles forming the monolayer is @xmath13 .
using video microscopy , we simultaneously track the center of mass of approximately 500 particles in the full field of view at 3.3 frames per second with a spatial accuracy of @xmath14 @xcite .
the viscous drag coefficient of the particles in the solvent under this confinement , measured from their equilibrium mean - square displacement , is @xmath15 .
a coupling between the particles is obtained by a static homogeneous magnetic field @xmath16 applied perpendicular to the sample plane .
this field induces a repulsive dipole - dipole interaction @xmath17 with @xmath18 the magnetic constant , @xmath19 the particle separation distance and @xmath20 a constant , which allows to control the stiffness of the colloidal monolayer . by interference of three laser beams ( @xmath21 ) a light field with hexagonal symmetry inside the sample cell
is generated , corresponding to a 2d periodic potential landscape @xmath22 , whose maximum depth is @xmath23 .
the potential profile is described by the function @xmath24.\ ] ] the lattice constant @xmath25 and the depth @xmath6 can be tuned by the intensities and the angles of incidence of the laser beams @xcite . before a measurement
is performed , the colloidal monolayer is allowed to equilibrate at room temperature @xmath26 k for at least one hour in presence of a magnetic field of @xmath27 .
the resulting homogeneous crystalline state with hexagonal symmetry exhibits a lattice constant of approximately @xmath28 with an interaction potential of @xmath29 , as shown in fig .
[ fig : fig0](b ) .
the lattice constant @xmath25 of the light potential @xmath22 is adjusted to the same value , i.e. we focus on commensurate conditions , in order to resolve the transition from static to sliding friction of the colloidal monolayer @xcite .
of particle trajectories counted in the monolayer for two different frictional regimes ( a ) @xmath30 fn and @xmath31 ( stick - slip motion ) and ( b ) @xmath5 fn and @xmath32 ( complete sliding ) , computed over different time intervals : @xmath33 s ( @xmath34 ) , 1.2 s ( @xmath35 ) , 2.4 s ( @xmath36 ) , 4.8 s ( @xmath37 ) , 9.6 s ( @xmath38 ) , 19.2 s ( @xmath39 ) , and 38.4 s ( @xmath40 ) .
the dotted lines represent eq .
( [ eq : meanwork ] ) , with no fitting parameters.,title="fig:",scaledwidth=100.0% ] [ fig : fign ] the whole sample cell is displaced in the x - direction with velocity @xmath41 by use of a piezo table in order to move the particles across @xmath22 , as sketched in fig .
[ fig : fig0](a ) .
the movement allows to create a controlled homogeneous flow @xmath42 and to drive the particles into non - equilibrium steady states ( ness ) . by tuning the value of @xmath41 we can mimic the response of the colloidal crystal to the effective external force @xmath43
the reason of this choice is that many tribological processes take place under these conditions @xcite , where an external force is applied to a crystalline surface with the purpose of moving it against a substrate , and where the surrounding fluid , in our case the solvent , only plays the passive role of a thermal bath to keep the system at fixed temperature . in order to induce a transition of the monolayer in response to @xmath44 from static to sliding friction
, we perform two different experimental protocols . in the first
, we keep the depth of the light potential at its maximum value @xmath31 and then we displace the cell at different velocities to tune the magnitude of the applied force @xmath45 . in fig .
[ fig : fig1](a ) we plot some trajectories of 50 neighboring particles moving according to this protocol . at small velocities , corresponding to values of @xmath3 much smaller than the maximum restoring force exerted by the light potential ( [ eq : potentialprofile ] ) , @xmath46 ,
the particles remained pinned by the potential wells , as shown in the left panel of fig [ fig : fig1](a ) for @xmath47 ( @xmath48 ) .
note that , even when immobile in average , the position of each individual particle fluctuates due to the collision of the solvent molecules .
as @xmath3 approaches values comparable to or larger than @xmath49 , i.e. when overcoming static friction , the particles are able to move across the potential barriers , thus resulting in collective motion . for example , at @xmath50 ( @xmath51 ) , the monolayer is partially depinned and some of the particles start to move in the direction of @xmath52 , as shown in the central panel of fig .
[ fig : fig1](a ) . for @xmath53 ( @xmath54 , right panel of fig .
[ fig : fig1](a ) ) , all the particles in the monolayer are able to move in the direction of @xmath55 . under these conditions ,
the particles undergo stick - slip motion because their mobility is hindered by the presence of the potential barriers @xcite , resulting in a mean particle velocity @xmath56 . in the second protocol
, we fix the velocity of the sample at @xmath48 , i.e. at constant @xmath57 , and we vary the depth @xmath6 of the light potential . the response of the particles to this protocol with decreasing values of @xmath6 is qualitatively similar to that observed when increasing @xmath3 at constant @xmath6 and also has a transition from a static to a sliding friction regime .
this effect can be observed in fig .
[ fig : fig1](b ) for potentials depths @xmath58 , @xmath59 and @xmath60 , at which the maximum restoring force @xmath61 has the values 34 fn , 23 fn and 17 fn , respectively .
we point out that , although qualitatively similar , the particle dynamics resulting from these two protocols are not completely equivalent . indeed ,
close inspection of the trajectories in fig .
[ fig : fig1](a ) and [ fig : fig1](b ) shows that in the second protocol , brownian motion is more significant and the particle mobility is higher because the local confinement created by the substrate is reduced when decreasing @xmath6 .
therefore , the second protocol reproduces the effect of changing the roughness of the substrate , which in turn results in higher particle velocities compared to the first protocol for the same values of the parameter @xmath62 .
for instance , at @xmath63 , the mean particle velocity obtained by means of the first protocol is only 26% of the velocity @xmath41 of the sample cell ( right panel of fig .
[ fig : fig1](a ) ) , whereas in the second protocol , it almost reaches free sliding at 93% of @xmath41 ( right panel of fig .
[ fig : fig1](b ) ) .
we focus on the work done on the colloidal monolayer of @xmath64 interacting particles driven across the corrugation potential @xmath22 by an external force @xmath44 , which is the common situation encountered in many tribological problems @xcite .
we fist present the equations of motion for our specific experimental protocol under flow @xmath65 , which allows to mimic in a controlled manner the dynamics under applied force @xmath44 .
then , we derive the corresponding stochastic - thermodynamic quantities of the latter tribological process . when the @xmath66-th particle ( @xmath67 ) moves at instantaneous position @xmath68 and velocity @xmath69 in presence of flow @xmath65 , the viscous drag force relative to the flow is @xmath70 . in our system , there is no actual external force but only conservative forces derived from the magnetically - induced repulsive interactions and the periodic light field . in addition , each particle is subject to the random thermal collisions of the solvent molecules .
therefore , the dynamics of the @xmath66-th particle is described by the langevin equation @xmath71 where @xmath72 is the total potential energy of the system , which includes the light potential @xmath22 , the pair - interaction potential @xmath73 of all the particles and the confining potential exerted by the sample cell , @xmath74 , which maintains the monolayer in a packed configuration and prevents the particles at the boundaries from escaping from the monolayer due to the repulsive interactions @xmath75 whereas the fast interactions with the surrounding solvent molecules are modeled by a gaussian white noise @xmath76 of zero mean and correlations @xmath77 . because of the structure of eq .
( [ eq : langevin ] ) , the dynamics of every particle in response to the flow and _ in absence of an external force _
@xcite is equivalent to the dynamics in response to an external uniform force , @xmath78 , and _ _ without external flow__. in our experiments this assumption is fully justified because the reynolds number is @xmath79 . ]
@xmath80 therefore , hereafter we will focus only on eq .
( [ eq : langevinforce ] ) in order to study the stochastic thermodynamics of a the monolayer under external constant force and without flow . in the context of stochastic thermodynamics , the first law for the potential energy variation along a single stochastic realization of the dynamics of the system can be written as @xcite @xmath81 where @xmath82 and @xmath83 are the heat dissipated into the solvent and the work done by @xmath52 , respectively , and are given by @xmath84 \cdot \mathrm{d } { \bf{r}}_i , \nonumber\\ \mathrm{d}w & = & \sum_{i=1}^n { \bf{f } } \cdot \mathrm{d } { \bf{r}}_i.\end{aligned}\ ] ] then , from eq .
( [ eq : workheat ] ) the work done by the uniform force @xmath55 on the colloidal monolayer , normalized by @xmath85 , over the time interval @xmath86 $ ] reads @xmath87.\end{aligned}\ ] ] it should be noted that the expression of the work in eq .
( [ eq : workforce ] ) only involves the value of the force @xmath3 , which can be tuned experimentally by means of @xmath41 , and the instantaneous values of the x - coordinates of each particle , which are determined by videomicroscopy .
consequently , the work can be directly determined from the particles trajectories without the need to measure the pair interactions . from eq .
( [ eq : workforce ] ) , we can conclude that , regardless of the nature of the pair interactions , the mean value of the ness work done by @xmath44 over a time interval of duration @xmath88 can be expressed as @xmath89 where the brackets stand for an ensemble average over @xmath64 particle trajectories and @xmath90 is the drift particle velocity in response to @xmath44 .
we check that eq .
( [ eq : meanwork ] ) is valid in all the frictional regimes investigated in our experiments is not necessarily fulfilled for systems with many coupled degrees of freedom driven by time dependent forces , see for example @xcite . ] .
for instance , it is trivially satisfied for static friction , where @xmath91 yields @xmath92 because in average no mechanical work is done by @xmath44 on the monolayer . on the other hand , for @xmath93 the linearity of @xmath94 with respect to @xmath64 and @xmath88 predicted by eq .
( [ eq : meanwork ] ) is also verified .
for example , in figs .
[ fig : fign](a ) and [ fig : fign](b ) we plot for different integration times @xmath88 the value of the mean work @xmath94 for stick - slip motion and free sliding , respectively , as a function of the number @xmath64 of particle trajectories used in the computation of @xmath95 . in this case
, @xmath94 is determined by taking the average over all possible values of @xmath95 at fixed @xmath64 and @xmath88 .
we also plot as dotted lines the values of @xmath94 computed by means of eq .
( [ eq : meanwork ] ) , where the ness drift velocity @xmath96 is independently determined from the particle dynamics .
we observe that the agreement between both kinds of calculations is excellent .
therefore , from the validity of eq .
( [ eq : meanwork ] ) we conclude the mean work mirrors the bulk frictional properties of the monolayer , namely a smooth transition from @xmath92 ( static friction with zero mobility at small @xmath3 ) to @xmath97 ( sliding friction with constant finite mobility at sufficiently large @xmath3 ) @xcite .
fn over @xmath98 s on subsystems composed of different number @xmath99 of particles across two potentials of depths ( a ) @xmath100 and ( b ) @xmath101 . from top to bottom in [ fig : fig2](a ) and from left to right in [ fig : fig2](b ) : @xmath102 ( dark blue ) , 5 ( light blue ) , 10 ( cyan ) , 25 ( dark green ) , 50 ( light green ) , 100 ( orange ) , and 250 ( red ) . insets : work distribution rescaled according to eq .
( [ eq : rescaling ] ) .
the symbols correspond to @xmath102 ( @xmath40 ) , 5 ( @xmath103 ) , 10 ( @xmath34 ) , 25 ( @xmath39 ) , 50 ( @xmath35 ) , 100 ( @xmath104 ) , and 250 ( @xmath37 ) .
the solid lines represent the rescaled gaussian distributions for non - interacting particles given by eqs .
( [ eq : case1 ] ) and ( [ eq : case2 ] ) , respectively.,title="fig : " ] [ fig : fig2 ] in principle , the fluctuations of @xmath95 depend on the strength of the repulsive interactions , the number of particles @xmath64 , the integration time @xmath88 , the force ratio @xmath105 and the depth @xmath6 of the substrate potential .
nevertheless , using the langevin model of eq .
( [ eq : langevinforce ] ) , we can gain some insight into the statistical properties of the work by analyzing two limit ideal cases which bear resemblance to static and sliding friction , respectively .
the first case corresponds to @xmath64 non - interacting particles moving under the influence a very weak force @xmath106 in presence of a very high potential barrier @xmath107 , such that the inverse kramers rate of each particle becomes much larger than the other characteristic time - scales of the system . in this case , which resembles static friction conditions
, the system is in a quasi - equilibrium state , where the particles are pinned by the potential wells at an average distance @xmath108 from the minima in order to balance the external force @xmath3 .
the probability density function of the work @xmath95 is gaussian , i.e. @xmath109 $ ] , with mean @xmath94 and variance @xmath110 given by @xmath111,\end{aligned}\ ] ] where @xmath112 ^ 2v_{max}$ ] is the effective stiffness of the restoring force exerted by a periodic light potential with hexagonal symmetry ( [ eq : potentialprofile ] ) .
note that , while @xmath92 because no mechanical work is done in average by @xmath52 , @xmath110 is non - zero .
this is due to the thermal fluctuations of the solvent molecules , which can promote either positive or negative work fluctuations by randomly moving the particles with or against the applied force .
the second ideal case is when @xmath64 non - interacting particles are driven by a sufficiently large force @xmath113 , such that they move at the highest possible average velocity @xmath114 , where the influence of the periodic potential is negligible , similar to free sliding friction .
in such a case , the probability density function of @xmath95 is also gaussian , with mean and variance @xmath115 respectively .
we point out that only in this particular case , the non - equilibrium work trivially satisfies the detailed steady - state fluctuation theorem @xcite @xmath116 because @xmath95 is actually equal to the total entropy production of the system , normalized by @xmath117 . in the following ,
we discuss how the previous ideal expressions for @xmath118 compare to the experimental work distributions for interacting particles in the static and sliding friction regimes .
furthermore , we also investigate the work distribution in the intermediate regime when tuning the control parameters to induce a transition from static to sliding friction of the colloidal monolayer .
fn on @xmath119 particles against a periodic light potential of depth @xmath100 over different time intervals : @xmath33 s ( @xmath40 ) , 1.2 s ( @xmath120 ) , 2.4 s ( @xmath121 ) , 4.8 s ( @xmath39 ) , 9.6 s ( @xmath35 ) , 19.2 s ( @xmath104 ) , and 38.4 s ( @xmath37 ) .
the solid line represents the case of non - interacting particles under the same @xmath122 and @xmath3 .
inset : standard deviation of @xmath95 as a function of @xmath88 for non - interacting ( solid line ) and magnetically coupled ( dashed line ) particles ( b ) probability density function of the work done by a constant force @xmath5 fn on @xmath119 particles against a light potential of depth @xmath123 over different time intervals . from left to right : @xmath33 s , 1.2 s , 2.4 s , 4.8 s , 9.6 s , 19.2 s , and 38.4 s. inset : work distribution measured over different @xmath88 and rescaled according to eq .
( [ eq : rescaling_time ] ) .
same symbols as in fig .
[ fig : fig3](a ) .
the solid line represents the case of non - interacting particles.,title="fig : " ] [ fig : fig3 ] since only a portion of the complete monolayer of @xmath13 particles is accessible for data analysis , a frequent problem encountered in spatially extended systems @xcite , we first investigate the effect of measuring the work done on a smaller subsystem composed of @xmath124 particles , thus ignoring its coupling with the @xmath125 degrees of freedom of the rest of the system . a possible way to probe the role of such a coupling is by means of the differences between the statistical properties of the work applied on a subsystem of @xmath99 interacting particles with those observed in a subsystem of the same size @xmath99 of non - interacting ones , where there is no coupling
. note that in absence of interactions , the work is a gaussian variable with mean and variance proportional to the number of components @xmath99 of the subsystem for the two limit cases described by eqs .
( [ eq : case1 ] ) and ( [ eq : case2 ] ) , i.e. the width of the distribution scales in both cases as @xmath126 . therefore , upon translating the work distribution to the origin by an amount @xmath94 and then squeezing it by its width @xmath127 any subsystem composed of @xmath99 non - interacting particles exhibits a @xmath99-independent profile @xmath128 .
this means that the statistical properties of the work done on the whole system can be probed by measuring the work done on any subsystem of arbitrary size .
this situation can change drastically in presence of particle interactions , though . as discussed in @xcite , because of the spatio - temporal correlations created by the interactions between the subsystem and the surroundings , the undercount of slow degrees of freedom can give rise to strong modifications of the statistical properties of the thermodynamic quantities of the subsystem with respect to those of the complete system .
then , it is not expected that the variance scales as @xmath129 for sufficiently small @xmath99 in presence of interactions .
the effect of the coupling with the surroundings only vanishes when the size of the sampling subsystem spans a length - scale larger than the typical correlation length induced by the interactions , thus recovering the actual statistical properties of the complete system @xcite . indeed , in presence of repulsive interactions we observe this kind of non - trivial dependence of the work distribution on the number @xmath99 of ness trajectories used to compute @xmath95 from eq .
( [ eq : workforce ] ) in both static and free sliding frictional regimes .
in fig [ fig : fig2](a ) we plot the probability density function @xmath118 of the work computed over @xmath130 s for subsystems composed of different number @xmath99 of particles ( @xmath131 ) at @xmath47 and @xmath132 , for which all the particles in the monolayer are pinned by the potential wells over the observation times accessible in the experiment , see left panel of fig .
[ fig : fig1](a ) .
each subsystem is chosen in such a way that @xmath99 neighboring particles cover an approximately square area @xmath133 .
we find that for all the values of @xmath99 , @xmath118 is symmetric and peaked around @xmath134 because no work is done in average in this quasi - equilibrium state , whereas its width increases with increasing @xmath99 . in the inset of fig .
[ fig : fig2](a ) we plot the work distribution rescaled according to eq .
( [ eq : rescaling ] ) .
we observe that @xmath135 has a gaussian profile , and unlike the case of non - interacting pinned particles , its width increases with increasing @xmath99 .
this implies that the variance @xmath110 grows faster than @xmath99 in presence of repulsive interactions for sufficiently small @xmath99 , an indication that the correlations between the subsystem and the @xmath136 surrounding particles are significant . nevertheless , for sufficiently large values of @xmath99 , we find that @xmath135 seems to converge to a size - independent profile , as shown in the inset of fig .
[ fig : fig2](a ) for @xmath137 .
the convergence demonstrates that finite - size effects due to the spatial correlations between the sampling subsystem and rest of the particles in the monolayer become negligible compared to the global behavior of @xmath95 for sufficiently large @xmath99 .
however , the effect of the particle interactions on the fluctuations of @xmath95 persists even for sufficiently large @xmath99 . as a matter of fact , when comparing the experimental @xmath135 for @xmath138 and @xmath139 with that computed from eq .
( [ eq : case1 ] ) with @xmath140 for non - interacting particles ( solid line in the inset of fig .
[ fig : fig2](a ) ) , we find that the former are much wider than the latter .
this suggest that , with increasing numbers of particles @xmath99 , the randomness created by the strongly non - linear coupling accumulate , giving rise to fluctuations of @xmath95 larger than those that would be otherwise observed in absence of interactions . a different behavior is observed for free sliding , where all the particles are able to move across the potential landscape at a mean velocity close to @xmath141 , as those shown in the right panel of fig .
[ fig : fig1](b ) .
an example of such a behavior is shown in fig .
[ fig : fig2](b ) where we plot the probability density function @xmath118 of the work done by a force @xmath5 fn on subsystems formed by different number of particles , @xmath142 , across a potential of depth @xmath143 . in this case
, the work distribution is gaussian , whose maximum is located at positive values of @xmath95 , because the applied force is able to perform mechanical work by moving the monolayer . the mean work , which coincides with the location of the maximum of @xmath118 , increases linearly with increasing @xmath99 , in quantitative agreement with eq .
( [ eq : meanwork ] ) , as shown in fig .
[ fig : fign](b ) . on the other hand
, the presence of interactions affects the behavior of the fluctuations of @xmath95 compared to the ideal sliding case described by eq .
( [ eq : case2 ] ) . in order to highlight these differences , in the inset of fig .
[ fig : fig2](b ) we plot the rescaled work distribution @xmath144 defined in eq .
( [ eq : rescaling ] ) .
once more , the effect of the correlation between the subsystem and the rest of the monolayer can be observed for small values of @xmath99 , for which the width of the rescaled distribution increases with @xmath99 .
however , for @xmath145 , @xmath144 converges to a @xmath99-independent profile , thus probing the actual statistical properties of @xmath95 for the complete system .
this convergence implies that the variance of the work scales as @xmath129 for sufficiently large @xmath99 . in the inset of fig .
[ fig : fig2](b ) we also plot as a solid line the rescaled work distribution of non - interacting sliding particles , described by eq .
( [ eq : case2 ] ) .
interestingly , we find that the width of the rescaled work distribution in the presence of interactions is smaller than that of the non - interacting case
. we can interpret this narrowing as a reduction of the work fluctuations due to the repulsive interactions , which give rise to an effective stiffening of the monolayer , thus preventing large random excursion of the particles induced by the thermal fluctuations around the drift imposed by @xmath44 .
this is consistent with the fact that for a perfectly stiff colloidal crystal , which can be realized in the limit of infinitely strong repulsive interactions , thermal fluctuations are suppressed @xcite , which gives rise to a complete sharpening of the work distribution around the mean value of eq .
( [ eq : meanwork ] ) .
we now focus on the dependence of the probability density function of the work on the integration time @xmath88 .
we point out that for values of @xmath88 smaller than the relaxation time - scales of the system , time - correlations can affect also the statistical properties of the work , because the expression of @xmath95 in eq .
( [ eq : workforce ] ) involves differences at distinct times of the particle positions .
nonetheless , for sufficiently large values of @xmath88 , such time - correlations vanish and therefore the shape of the work distribution must converge to a single profile upon time rescaling .
we first show in fig .
[ fig : fig3](a ) the results for the case of a pinned colloidal monolayer , where in average no mechanical work is done .
here we plot the probability density function of the work @xmath95 done on @xmath119 particles by a force @xmath5 fn against a light potential of depth @xmath146 over different integration times , @xmath147 .
we observe that , for all the values of @xmath88 .
@xmath118 is gaussian and centered around @xmath148 , whose width increases with increasing @xmath88 .
however , for @xmath149 s , the width of the distribution levels off and all curves collapse onto a master curve regarless of @xmath88 .
this is further verified in the inset of fig .
[ fig : fig3](a ) , where we plot as a dashed line the dependence of the standard deviation @xmath150 of the work on @xmath88 , observing a saturation to a constant value at sufficiently large @xmath88 .
the dependence of @xmath150 on @xmath88 is qualitatively similar to that for non - interacting particles , shown as a solid line in the inset of fig .
[ fig : fig3](a ) . in this case , according to eq .
( [ eq : case1 ] ) , the variance of the gaussian work distribution levels off exponentially for integration times larger than the viscous relaxation time of the particles in the potential wells , @xmath151 s. this behavior of @xmath95 can be actually understood at the single - particle level . for @xmath152 ,
the particle motion is strongly auto - correlated in time due to the energy stored by the confining light potential , which translates into a very narrow distribution @xmath118 .
the motion becomes less and less correlated when @xmath88 approaches @xmath153 , and therefore each particle is able to perform larger brownian displacements within the potential wells both with and against the applied force , thus resulting in a broadening of @xmath118
. nevertheless , the fluctuations of @xmath95 can not grow indefinitely with increasing @xmath88 because the single - particle motion is always bounded to the potential wells , giving rise to a saturation of @xmath150 for @xmath154 .
although qualitatively similar as a function of @xmath88 , we observe a quantitative difference at @xmath155 between the standard deviation of @xmath95 for interacting particles with respect to that in the non - interacting case , as shown in the inset fig .
[ fig : fig3](a ) .
this difference is due to the strong coupling between the particles , which gives rise to a complex non - linear particle dynamics within the potential wells . in fig .
[ fig : fig3](b ) we illustrate the dependence of the work distribution @xmath156 on the integration time @xmath88 for the sliding friction regime of @xmath119 particles driven at @xmath4 fn and @xmath157 .
we find that @xmath156 is gaussian and the location of the maximum increases linearly with increasing @xmath88 in accordance with eq .
( [ eq : meanwork ] ) .
once more , inspired by the comparison with non - interacting particles , where the width of the distribution scales as @xmath158 ( see eq .
( [ eq : case2 ] ) ) , we can test a scaling with respect to @xmath88 similar to eq .
( [ eq : rescaling ] ) @xmath159 interestingly , in the inset of fig .
[ fig : fig3](b ) we show that the work distributions , rescaled according to eq .
( [ eq : rescaling_time ] ) , collapse onto a master curve for all @xmath88 .
this essentially means that in this frictional regime the variance of the work scales as @xmath160 even in presence of particle interactions .
note that in this case , the particles are not confined to move in the potential wells , and consequently there is no intrinsic relaxation time in the dynamics , which explains the very fast convergence of @xmath144 to the master curve .
quantitative differences are observed between the experimental @xmath144 and the case without interactions ( solid line in the inset of fig .
[ fig : fig3](b ) ) , though .
this occurs due to the narrowing of the work distribution due to the effective stiffening of the monolayer .
we now show how the shape of the work distribution changes between the two very distinct cases previously studied , i.e. when changing the experimental parameters to induce a transition from the regime where all the particles are pinned on the substrate , to the depinning of the colloidal monolayer and subsequent free sliding .
we recall that in the two extreme regimes of static and sliding friction , the work distribution is gaussian , even though the mean and variance behave differently as a function of @xmath99 and @xmath88 . while in static friction these quantities also depend strongly on both the elastic stiffness @xmath161 exerted by the substrate potential and the repulsive pair - interactions , they are only affected by the strength of the interactions for sliding friction . in fig .
[ fig : fig4](a ) we illustrate the effect on the shape of the work distribution for @xmath119 particles , computed over @xmath162 s , when increasing the value of the applied force @xmath3 at constant potential depth @xmath31 in order to induce particle depinning .
interestingly , we observe that the work distribution becomes asymmetric with respect to the maximum with increasing @xmath3 , as can be observed for @xmath163 fn and @xmath164 fn . in particular ,
non - gaussian tails appear at positive values of @xmath95 , as highlighted in the semilog plot of the inset of fig .
[ fig : fig4](a ) . for these values of @xmath3 ,
the spatio - temporal dynamics of the monolayer becomes heterogeneous , as can be observed from the particle trajectories of fig .
[ fig : fig1](a ) . for example
, for @xmath163 fn there are regions where the particles are still confined by the potential wells , because the external force is still smaller than the maximum conservative force exerted by the light field : @xmath165 .
however , the combination of thermal fluctuations and non - linear repulsive interactions can promote hops to the neighboring potential wells , thus creating collective motion of clusters of particles .
the collective motion is in turn facilitated by the symmetry beaking induced by the external force .
while the stagnant particles do not contribute to the mean value of the work but only to the fluctuations around @xmath166 , the non - gaussian tails originate from the work done on the sliding particles .
this regime persists even when the complete monolayer can slide across the periodic substrate , as observed at @xmath164 fn and @xmath31 , where the particles undergo stick - slip motion . in this case , the spatial heterogeneity is induced by the corrugation potential , creating zones around the potential minima where the particles slow down , whereas they move faster when overcoming the potential barriers , as illustrated by the trajectories in the right panel of fig .
[ fig : fig1](a ) .
note that this heterogeneity results also in a mean particle velocity @xmath167 much smaller than the maximum velocity that could be achieved in presence of a completely flat substrate ( @xmath54 ) .
gaussianity of @xmath95 is recovered at sufficiently large @xmath105 , though , with a respective narrowing of @xmath118 .
this can be observed in fig .
[ fig : fig4](b ) , where we plot @xmath118 for the second experimental protocol with which we can reach more easily the free sliding regime .
a reduction of only @xmath168 in the potential depth @xmath6 , from @xmath169 to @xmath9 , is enough to observe a prominant change of the shape of the work distribution , as shown in the inset of fig .
[ fig : fig4](b ) . for this values of @xmath6 , the mean particle velocity changes from @xmath170 to @xmath171 , whereas the maximum velocity that could be achieved for this value of @xmath3 on a completely flat surface is @xmath172 .
this dramatic change in the shape of @xmath118 reveals that not only the average tribological properties of the monolayer @xcite but also the properties of the fluctuations of the work done on it become very sensitive when tuning the experimental parameters close to the depinning transition .
on @xmath119 particles against a potential of depth @xmath173 , computed over @xmath162 s. from left to right : @xmath4 fn , 18.6 fn , and 24.8 fn .
( b ) probability density function of the work done by an external force @xmath5 fn on @xmath119 particles against potentials of different depth @xmath6 , computed over @xmath162 s. from left to right : @xmath174 , and @xmath9 .
the insets are semi - logarithmic representations of the same plots in the main figures.,title="fig : " ] [ fig : fig4 ] finally , we investigate the work distribution @xmath118 within the context of non - equilibrium work relations .
more specifically , we focus on the evaluation of quantity @xmath175 , which quantifies the asymmetry of the probability of observing positive work fluctuations , where the monolayer moves in the direction of the applied force , with respect to the probability of observing rare negative fluctuations , where the monolayer moves against the force .
we point out that , although empirically satisified in many steady - state complex systems @xcite , a simple linear relation such as eq .
( [ eq : ft ] ) for the asymmetry function , @xmath176 , is not expected to hold generally for the ness system we study .
indeed , for this kind of frictional processes , such a linear asymmetry relation is strictly valid only for the work done by an external force on a collection of non - interacting particles freely sliding on a perfectly flat substrate .
in such a case , eq . ( [ eq : ft ] ) is a direct consequence of the detailed fluctuation theorem , which only applies to the total entropy production of the system , and which in that specific case equals the work done by an external force . on the other hand , since the work defined by eq .
( [ eq : workforce ] ) has a definite parity under time - reversal , it must satisfy a generalized fluctuation theorem in presence of the corrugation potential , the pair - interactions and the global confining potential of the sample cell @xcite @xmath177 in eq .
( [ eq : detailedft ] ) , @xmath178/(k_b t)$ ] is the variation of the total potential energy of the system during a time interval @xmath88 , given by eq .
( [ eq : potential ] ) , @xmath179 is the stochastic entropy change over @xmath88 @xcite , and the brackets denote a conditional average over the stochastic realizations for which @xmath95 equals the value @xmath180 . the last term on the right - hand side of eq .
( [ eq : detailedft ] ) is in general non - zero in presence of particle interactions , and therefore @xmath118 does not necessarily satisfy the exact linear relation of eq .
( [ eq : ft ] ) .
note that if @xmath118 is gaussian , the asymmetry function can still be proportional to @xmath180 , i.e. @xmath181 where the parameter @xmath182 , i.e. the slope of the linear relation , is given by @xmath183 however , unlike the ideal case of non - interacting particle described by eq .
( [ eq : ft ] ) , the parameter @xmath182 is in general different from 1 because the second term on the right - hand side of eq .
( [ eq : detailedft ] ) , which involves the particle interactions and the substrate potential , is non - zero . for instance , we observe that eq .
( [ eq : asymmetrygaussian ] ) holds in the static friction regime ( completely pinned monolayer ) and in the sliding regime , as shown in fig .
[ fig : fig5](a ) . in the static friction regime ,
the parameter @xmath182 is equal to @xmath184 for all @xmath99 and @xmath88 , because the system is in a quasi - equilibrium state , with equal probabilities @xmath185 and @xmath186 . on the other hand , we find @xmath187 for all @xmath88 in the free sliding regime , as illustrated by the symbols around the dashed line in fig . [
fig : fig5](a ) .
this implies that in this case the second term on the right hand side of eq .
( [ eq : detailedft ] ) is non - zero : @xmath188 .
it should be noted that the direct computation of the asymmetry function from @xmath118 is restricted to rather small values of @xmath99 and @xmath88 , because negative work fluctuations are difficult to sample with increasing values of such parameters .
however , taking into account that @xmath118 is gaussian , @xmath182 can be estimated from the mean and the variance by means of eq .
( [ eq : slope ] ) . surprisingly , in the inset of fig .
[ fig : fig5](a ) we show that the value @xmath187 holds even for @xmath99 and @xmath88 as large as @xmath189 and @xmath190 s , respectively , thus demonstrating that the particle interactions give rise to a robust behavior of the term @xmath188 in eq .
( [ eq : detailedft ] ) .
this unconventional behavior of the asymmetry function can be traced back to the strong coupling between the particles forming the crystalline monolayer . indeed , with increasing strength of repulsive interactions , which is fixed in our experiment by the magnetic field @xmath191 , the width of the work distribution decreases because of the increasing stiffness of the colloidal crystal .
note that in the limit of a perfectly stiff colloidal crystal , i.e. created by infinitely large repulsive interactions , the fluctuations of @xmath95 are completely suppressed . in this case
the work distribution becomes a delta function , @xmath192 , which gives rise to @xmath193 .
hence , for sliding friction @xmath182 must be an increasing function of the pair interaction strength , bounded by the values @xmath194 ( no interactions ) and @xmath195 ( infinitely large repulsions ) .
the value @xmath187 , specific to our experimental conditions , clearly illustrates that repulsive pair interactions reduce the fluctuations of @xmath95 compared to the value @xmath196 of eq .
( [ eq : ft ] ) in absence of interactions .
the intermediate frictional regime , where @xmath118 exhibits non - gaussian tails due to the heterogeneous spatio - temporal dynamics of the monolayer , is particularly interesting . in this case
, we observe that the asymmetry function is not even linear in @xmath180 , as shown in fig .
[ fig : fig5](b ) for @xmath4 fn and @xmath197 , at which only partial depinning from the substrate is achieved .
the asymmetry function is approximately linear for small values of @xmath180 , with slope @xmath198 due to large negative work fluctuations on the stagnant particles .
nevertheless , significant deviations from this linear behavior show up at larger work fluctuations , @xmath199 , when probing values of @xmath95 on the non - gaussian tails plotted in fig .
[ fig : fig4](b ) .
once more , the behavior of the asymmetry functions seems to be robust , as shown in fig .
[ fig : fig5](b ) , where all the data points collapse to a master curve for different values of @xmath88 and @xmath99 .
we point out that in general , such a non - linear behavior of the asymmetry function is not easily observed in systems described by a small number degrees of freedom , because in such a case large negative fluctuations are difficult to sample @xcite . in our experimental system
we are able to achieve this because of the existence of strong negative work fluctuations , which originate from the heterogenous dynamics of the coupled degrees of freedom of the system during the depinning transition of the colloidal monolayer .
we have investigated the statistical properties of the work done by an external force on a monolayer of magnetically interacting particles driven across a periodic potential , which mimics friction between crystalline atomic surfaces .
we have studied the influence of the number of particles used to probe these properties , the integration time , and the control parameters that are tuned to induce a transition from a pinned state ( static friction ) to complete depinning from the substrate potential ( sliding friction ) .
we have shown that , in the static and free sliding regimes , the work distribution converges to a gaussian master curve for sufficiently large number of particles and integration times upon rescaling of these parameters .
we have found that the mean and variance of such work distributions depend on the strength of the repulsive interactions , which in particular give rise to a stiffening of the monolayer for free sliding .
interestingly , we have also found that in the intermediate friction regime , where the monolayer undergoes a depinning transition , the work distribution becomes non - gaussian because of the heterogeneity of the particle dynamics , e.g. due to partial depinning and stick - slip motion .
we have shown that in general , the work distribution exhibits unconventional asymmetry properties within the context of non - equilibrium fluctuations relations .
we have demonstrated that such a behavior originates from the presence of repulsive particle interactions .
thus , we provide the first experimental measurements of a stochastic thermodynamic quantity with non - trivial properties for a mesoscopic system with many coupled degrees of freedom . for : static ( @xmath40 ) , and sliding friction regime measured for @xmath200 particles at @xmath5 fn , @xmath201 , over @xmath202 s ( @xmath34 ) , @xmath203 s ( @xmath35 ) , @xmath204 s ( @xmath36 ) .
the solid circles are measurements under the same conditions for @xmath205 particles during @xmath203 s. the dashed line is a guide to the eye with slope @xmath206 .
inset : dependence of the parameter @xmath182 on @xmath88 computed by means of eq .
( [ eq : slope ] ) for @xmath138 ( @xmath38 ) and @xmath189 ( solid line ) particles .
( b ) asymmetry function of the work distribution @xmath118 measured for @xmath205 particles moving with a heterogeneous dynamics under @xmath4 fn and @xmath197 over @xmath207 s ( @xmath40 ) , @xmath208 s ( @xmath120 ) , @xmath209 s ( @xmath34 ) , @xmath210 s ( @xmath39 ) , @xmath162 s ( @xmath35 ) , @xmath211 s ( @xmath38 ) , and @xmath212 s ( @xmath36 ) .
the solid circles are measurements under the same conditions for @xmath119 particles over @xmath213 s. the dashed line is a guide to the eye with slope @xmath214.,title="fig:",scaledwidth=100.0% ] [ fig : fig5 ]
we thank udo seifert for helpful discussions . we acknowledge financial support of the deutsche forschungsgemeinschaft ,
be 1788/10 - 1 .
|
we experimentally investigate the non - equilibrium steady - state distribution of the work done by an external force on a mesoscopic system with many coupled degrees of freedom : a colloidal crystal mechanically driven across a commensurate periodic light field . since this system mimics the spatiotemporal dynamics of a crystalline surface moving on a corrugated substrate , our results show general properties of the work distribution for atomically flat surfaces undergoing friction .
we address the role of several parameters which can influence the shape of the work distribution , e.g. the number of particles used to locally probe the properties of the system and the time interval to measure the work .
we find that , when tuning the control parameters to induce particle depinning from the substrate , there is an abrupt change of the shape of the work distribution . while in the completely static and sliding friction regimes
the work distribution is gaussian , non - gaussian tails show up due to the spatiotemporal heterogeneity of the particle dynamics during the transition between these two regimes . _ keywords _ : non - equilibrium work fluctuations , stochastic thermodynamics of interacting particles , friction , colloidal crystals
|
the field of evolutionary algorithms ( eas ) has been successfully applied to the areas of modern art @xcite .
evolutionary computation is an interesting approach to the generation of novel images and is beginning to have a broader impact on artistic fields .
more generally , evolutionary methods applied to problem solving in creative fields is an exciting and fast developing topic in computer science . in earlier years research
has been conducted using evolutionary algorithms for interactive generation of art .
dawkins @xcite and smith @xcite demonstrated the potential of darwinian variation to evolve biomorphs graphic objects . following this steps sims @xcite , latham and todd @xcite
have combined evolutionary techniques and computer graphics to create artistic images and to reproduce computer creatures of great complexity .
more recently unemi @xcite , hart @xcite created a collection of images using refined means of combining individual images and their genotypes .
these works have not only influenced the field of evolutionary computation , but were also displayed in galleries and museums around the world .
furthermore , research in computational art has employed deep neural networks to create artistic images @xcite . in our work
we introduce a novel approach to define evolutionary processes for evolutionary image transition .
the main idea comprises of using well - known evolutionary processes and adapting these in an artistic way to create innovative images .
( kinkaku - ji temple kyoto ) and target image @xmath0 ( daigo - ji temple kyoto),title="fig : " ] ( kinkaku - ji temple kyoto ) and target image @xmath0 ( daigo - ji temple kyoto),title="fig : " ] the evolutionary transition process starts from a given starting image @xmath1 and , through a stochastic process , evolves it towards a target image @xmath0 .
our goal is to maximise the fitness function where we count the number of the pixels matching those of the target image @xmath0 .
this problem mirrors the classical onemax problem , which has been widely studied in academic literature @xcite .
we start by introducing a variant of the asymmetric mutation operator studied in the context of minimum spanning trees @xcite and pseudo - boolean optimization@xcite .
this operator is based on flipping pixels with a given probability in each step and avoids the coupon collectors effect @xcite by using standard bit mutation applied to onemax .
afterwards , we introduce specific mutation operators for evolutionary image transition , namely strip and box mutations , which pick parts of the image consisting of strips and boxes and set them to the target image .
our experimental results show the artistic effects which can be produced by applying evolutionary process to an image via the operators described below .
the rest of the paper is structured as follows . in section [ sec2 ]
, we introduce evolutionary transition process and examine the behaviour of ( 1 + 1 ) eafor image transition in section [ sec3 ] .
section [ sec4 ] gives results of how _ strip mutation , combined strip mutation _ and _ box mutation _ can be used for evolutionary image transition process . in section [ sec5 ]
we investigate the use of _ asymmetric mutation _ as part of mutation operators and study their combinations with pixel - based mutations during the evolutionary process .
finally , we conclude with our overview about the insights from our investigations .
artificial intelligence has the potential to change our perception of the process of creating art .
using evolutionary algorithms can lead to significant novelty in the interaction between the computer and the artist in order to produce an artwork .
the application of evolutionary algorithms in an innovative way can change the apprehension of human - machine creativity .
artificial intelligence as a medium for creating art has a long and well - established tradition .
the growing influence and demand of artificial intelligence in our global industrial life has been discussed throughout a wide spectrum of our society .
so comes a meaningful question : how can computational creativity perform as a valuable addition to digital art to produce unique generative art ?
* let @xmath1 be the starting image and @xmath0 be the target image .
* set x:=s . *
evaluate @xmath2 . *
while ( not termination condition ) * * obtain image @xmath3 from @xmath4 by mutation . *
* evaluate @xmath5 * * if @xmath6 , set @xmath7 .
we consider an evolutionary transition process that transforms a given image @xmath4 of size @xmath8 into a given target image @xmath0 of size @xmath8 .
our goal is to study different ways of carrying out this evolutionary transformation based on random processes from an artistic perspective .
we start our process with a starting image @xmath9 .
our algorithms evolve @xmath1 towards @xmath0 and have at each point in time an image @xmath4 where @xmath10 .
we say that pixel @xmath11 is in state @xmath12 if @xmath13 , and @xmath11 is in state @xmath14 if @xmath15 . for our process
we assume that @xmath16 as pixels with @xmath17 can not change values and therefore do not have to be considered in the evolutionary process . to illustrate the effect of the different methods presented in this paper , we consider two images of famous kyoto temples ( see figure [ fig : opera ] ) .
the fitness function of an evolutionary algorithm guides its search process and determines how to move between images .
therefore , the fitness function itself has a strong influence on the artistic behaviour of the evolutionary image transition process .
an important property for evolutionary image transition , in this work , is that images close to the target image get a higher fitness score .
we measure the fitness of an image @xmath4 as the number of pixels where @xmath4 and @xmath0 agree .
this fitness function is isomorphic to that of the onemax problem when interpreting the pixels of @xmath1 as @xmath18 s and the pixels of @xmath0 as @xmath19 s .
formally , we define the fitness of @xmath4 with respect to @xmath0 as @xmath20 we consider simple variants of the classical ( 1 + 1 ) eain the context of image transition .
the algorithm is using mutation only and accepts an offspring if it is at least as good as its parent according to the fitness function .
the approach is given in algorithm [ alg : ea ] .
using this algorithm has the advantage that parents and offspring do not differ too much in terms of pixel which ensures a smooth process for transitioning the starting image towards the target .
all experimental results in this paper are shown for the process of moving from the starting image to the target image given in figure [ fig : opera ] where the images are of size @xmath21 pixels .
the algorithms have been implemented in matlab . in order to visualize the process ,
we show the images obtained when the evolutionary process reaches 12.5% , 37.5% , 62.5% and 87.5% of pixels evolved towards the target image .
we consider a simple evolutionary algorithm that has been well studied in the area of runtime analysis , namely variants of the classical ( 1 + 1 ) ea @xcite .
as already mentioned , our setting for the image transition process is equivalent to the optimization process for the classical benchmark function onemax .
the standard variant of the ( 1 + 1 ) eaflips each pixel with probability @xmath22 where @xmath23 is the total number of pixels in the given image .
using this mutation operator , the algorithm encounters the well - known coupon collector s effect which means that the additive drift towards the target image when having @xmath24 missing target pixels is @xmath25 @xcite . -6pt
+ * obtain @xmath3 from @xmath4 by flipping each pixel @xmath11 of @xmath4 independently of the others with probability @xmath26 if @xmath27 , and flip @xmath11 with probability @xmath28 if @xmath15 , where @xmath29 and @xmath30 are constants . in order to avoid the coupon collector s effect but obtaining images where all pixel are treated symmetrically , we use the _ asymmetric mutation _ operator introduced in @xcite .
jansen and sudholt @xcite have shown that the ( 1 + 1 ) ea using _ asymmetric mutation _ optimizes onemax in time @xmath31 which improves upon the usual bound of @xmath32 when using standard bit mutations .
we denote by @xmath33 the number of pixels where @xmath4 and @xmath0 agree .
similarly , we denote by @xmath34 the number of pixels where @xmath4 and @xmath1 agree .
each pixel is starting state @xmath12 is flipped with probability @xmath26 and each pixel in target state @xmath14 is flipped with probability @xmath28 . the mutation operator is shown in algorithm [ alg : asym ] .
the mutation operator differs from the one given in @xcite by the two constants @xmath35 and @xmath36 which allow one to determine the expected number of new pixels from the starting image and the target image , respectively .
the choice of @xmath35 and @xmath36 determines the expected number of pixel in the starting state and target state to be flipped . to be precise , the expected number of pixels currently in the starting state @xmath12 to be flipped is @xmath37 and the number of pixels in the target state @xmath14 to be flipped is @xmath38 as long as the current solution @xmath4 contains at least that many pixel of the corresponding type . in @xcite
the case @xmath39 has been investigated which ensures that at each point in time has an additive drift of @xmath40 and therefore avoiding the coupon collector effect at the end of the process .
using different values for @xmath35 and @xmath36 allows us to change the speed of transition as well as the relation of the number of pixels switching from the starting image to the target and vice versa while still ensuring that there is constant progress towards the target . for all our experiments we use @xmath41 and @xmath42
-6pt + we now introduce specific mutation operators for image transition . the design of our three mutation operators , namely _ strip mutation _ , _ combined strip mutation _ and _ block mutation _ , is strongly oriented towards production of interesting artistic images .
all of these mutation operators transition a region of the current image @xmath4 to an area of the target image @xmath0 , starting at randomly chosen pixel position @xmath11 .
_ strip mutation _ mutates a vertical strip of pixels from @xmath11 for a length of @xmath43 pixels or to the boundary of the image whichever comes first . _
the strip mutation _ imitates well - know technique in generative art called a glitch .
this effect intentionally corrupts data of an image by encoding the jpeg process .
all experiments shows that in each new generation the initial image and the target image overlay through a series of randomly placed strips . in figure [ fig:2 ]
we show the experimental results of _ strip mutation_. the effect of _ strip mutation _
creates generative art with highly interesting components .
furthermore , we have conceptualised a _ combined strip mutation _
operator where both horizontal and vertical strips gradually overlay the original image in random locations .
the parameter settings for _ the combined strip mutation _ are @xmath44 pixels in width and @xmath45 pixels in height for the horizontal strip and one pixel in width and @xmath44 pixels in height for the vertical orientation .
figure [ fig:3 ] shows a very interesting transition from the point of view of an image processing which also has artistic value .
our third mutation operator is _
block mutation_. this operator randomly selects a position in the space of the matrix @xmath46 and flips a block of @xmath47 pixels . because the position chosen for mutation can be any point @xmath11 we can observe that mutated blocks can overlap .
figure [ fig:4 ] shows the experimental results of _ block mutation_. we have executed a smooth transition so as to interest the viewer .
additionally , we have made the changes in the image clearly visible . -6pt
_ the asymmetric , strip , combined strip _ and _ box mutation
_ operators have quite distinct behaviours when applied to image transition .
we now study the effect of combining the approaches for evolutionary image transition in order to obtain a more artistic evolutionary process .
we explore the combination of _ the asymmetric mutation _ operator and _ strip , combined strip _ and _ box mutation_. here we run _ the asymmetric mutation _ operator as described in algorithm [ alg : asym ] .
the process alternates between _ asymmetric mutation _ and one of either _ strip _ , _ combined strip _ or _
box mutation_. figure [ fig:5 ] shows the results of the experiments for _ combined asymmetric _ and _ strip mutation_. the transitional images are significantly different to the images produced in _ strip mutation _ or _ box mutation_.
-6pt + -6pt + -6pt + interesting blurred elements randomly emerge over the image during the transition stage .
similar elements are shown as a characteristically steadily changing , whereas the elements of the target image continuously appear .
on the last image we can almost completely see the details of the target image .
the blurred elements seem to be created through the bright autumn colour in the target image in this stage of the sequence .
figure [ fig:7 ] shows the results of the experiments for _ combined asymmetric _ and _ combined strip mutation_. the lower sequence of the picture is similar to the top in terms of the behaviour of the patches .
more differences can be seen as the transition progresses , but the picture continues to be partially unrecognisable during the transition .
new elements can be discovered on the affected image which slowly appear and create an element of surprise .
we see less of the dotted image pattern , but now observe combined blurred elements randomly occurring in the upper sequence of the images .
the image provides an interesting comparison to the old linen pictures of the rembrandt epoch . in the advanced stages , the patches are well integrated into the images using _
asymmetric mutation_. finally , the last image we see is well integrated and shows more detail in all areas of the image .
figure [ fig:8 ] shows the results of the experiments for _ combined asymmetric _ and _ box mutation_. in the advanced stages , the patches are well integrated into the images using _ asymmetric mutation _ to produce a smooth transition process . in summary ,
the image transition of _ combined asymmetric _ and _ box mutation
_ differs significantly from _ the box mutation _ transition introduced in section [ sec4 ] .
in this paper we have investigated how evolutionary algorithms and related random processes can be used for image transition .
we have shown how _ the asymmetric mutation _ operator , introduced originally for the optimization of onemax , can be applied to our problem .
we examined how the varying mutation settings influences our results through _ box mutation , strip mutation , combined strip mutation _ and combined approaches with _
asymmetric mutation_. by investigating combinations of the different approaches , we have shown a variety of interesting evolutionary image transition processes .
all our investigations are based on a fitness function that is equivalent to the well - known onemax problem .
it would be interesting to study more complex fitness functions and their impact on the artistic behaviour of evolutionary image transition in future research .
hart , d. : toward greater artistic control for interactive evolution of images and animation . in finnegan , j.w .
, pfister , h. , eds . : 33 . international conference on computer graphics and interactive techniques , siggraph 2006 , acm ( 2006 ) 2
|
evolutionary algorithms have been used in many ways to generate digital art .
we study how evolutionary processes are used for evolutionary art and present a new approach to the transition of images .
our main idea is to define evolutionary processes for digital image transition , combining different variants of mutation and evolutionary mechanisms .
we introduce box and strip mutation operators which are specifically designed for image transition .
our experimental results show that the process of an evolutionary algorithm in combination with these mutation operators can be used as a valuable way to produce unique generative art .
|
recent experiments on ultracold gases in optical potentials @xcite and experiments on photons in microwave cavities @xcite have demonstrated that it is possible to prepare a fock state as a pure state in a finite - dimensional system . after the preparation of the fock state ,
the parameters of the system can suddenly be changed ( performing a `` quench '' ) such that the fock state is not an eigenstate of the new system hamiltonian @xmath0
. then the evolution of the many - body state due to the evolution operator @xmath1 will lead to a random walk inside the available hilbert space .
the visited states include other fock states as well as superpositions of fock states .
typical questions in this context are : what is the probability for visiting different states and how is this affected by the interaction of the particles ?
a natural quantity for measuring this probability is the spectral density function of the hamiltonian @xmath0 with respect to the initial fock state @xcite .
a classical candidate for modeling the evolution of a fock state is the hubbard model @xcite .
the corresponding discrete hamiltonian describes the tunneling of a particle between neighboring potential wells and a local particle - particle interaction . the hubbard model for bosons ( bose - hubbard model )
was realized as an atomic system in an optical lattice @xcite .
a possible realization of the bose - hubbard model by photons in coupled microwave cavities was proposed recently by hartmann et al .
an anharmonicity of the microwave cavities plays the role of the photon - photon interaction @xcite .
the simplest system for discussing the evolution of a fock state within the hubbard model is a double well , where particles can tunnel between the two wells . for @xmath2 bosons the underlying hilbert space is spanned by the @xmath3dimensional fock base @xmath4 , where @xmath5 bosons are in one well and @xmath6 in the other well @xcite . the initial state is prepared as a fock state , where all the bosons are in one of the two wells ( i.e. @xmath7 or @xmath8 ) , while the tunneling between the wells is turned off . to start the evolution , a `` quench ''
is provided by switching on the tunneling between the two wells .
this is realized by a sudden reduction the potential barrier between the wells in an atomic system @xcite or by connecting the two microwave cavities with an optical fiber @xcite .
a similar experiment was performed with two atomic clouds , subject to weak interaction and separated by an adjustable potential barrier @xcite . on the theoretical side , mean - field descriptions of the bose - hubbard model , such as a hartree approximation or the gross - pitaevskii equation , may work well for clouds with many bosons and weak boson - boson interaction @xcite .
they provide a rather poor approximation for the dynamics of small many - body systems ( cf . ref .
this was also observed in a recent study by streltsov et al . who compared the results of a simple hartree ( gross - pitaevskii ) approximation with a sophisticated multi - orbital hartree approximation @xcite .
the latter reveals that the bosonic clouds are related to superpositions of fock states in the form of n00n states |n00n= .
[ noon ] in the following we will study the hubbard dynamics of bosons in a double well in more detail .
in particular , we are interested in the connection of spectral properties and the formation of n00n states , based on a fock state with all the particles in one well as the initial state . to avoid problems with uncontrolled approximations
, we will rely on a full quantum calculation .
an exact solution is available in a fock - state base , as described previously in refs .
@xcite .
the paper is organized as follows : in sect .
[ sect : model ] the model , based on the bose - hubbard hamiltonian , is defined and in sect . [
sect : dynamics ] the dynamics of an isolated quantum system is explained .
then we discuss the dynamics of a noninteracting bose gas in sect .
[ sect : double_well_non ] and the dynamics of an interacting bose gas in sect .
[ sect : double_well_int ] .
the latter is divided into a study of a symmetric double well ( sect .
[ sect : double_well_symm ] ) and of an asymmetric double well ( sect .
[ sect : double_well_asymm ] ) . finally , we summarize the results of our calculation in sect . [
sect : results ] and discuss them in sect .
[ sect : discussion ] .
the many - body hamiltonian @xmath9 of @xmath2 bosons with mass @xmath10 reads = _ j=1^n + _ j , k=1^nu(*r*_j,*r*_k ) , [ mb_ham0 ] where @xmath11 is the momentum of a boson , @xmath12 is the one - body potential of the double well and @xmath13 is the two - body interaction potential . for the latter
we assume that it decays very quickly with the distance @xmath14 of the particles .
this implies that particles located in different wells do not interact with each other .
then the many - body hamiltonian is expressed in fock - state representation as @xmath15 = n - k , k|h|n - k,k .
[ bhh0 ] for the new hamiltonian @xmath0 , which acts in the hilbert space spanned by the fock base , we can use the bose - hubbard hamiltonian with local interaction in each well as a reasonable approximation h = j(a_1^a_2+a_2^a_1)+u_1(a_1^a_1)^2+u_2(a_2^a_2)^2 , [ bhh ] where @xmath16 ( @xmath17 ) are creation ( annihilation ) operators for bosons in the fock states .
@xmath0 , which describes tunneling between the two wells and the local interaction inside the well with interaction strength @xmath18 , gives us a complete quantum description of the different fock states and their superpositions .
in particular , we can use it to study the evolution of a fock state to a n00n state of eq .
( [ noon ] ) .
we consider a system which is isolated from the environment .
furthermore , we assume that the system lives in an ( @xmath19)dimensional hilbert space . with the initial state
@xmath20 we can get for the time evolution of the state |_t = e^-iht|_0[evol0 ] or the evolution of the return probability @xmath21 with the amplitude _ 0|_t=_0|e^-iht|_0 .
[ return0 ] in general , the amplitude @xmath22 can be expressed via an integral transformation of the resolvent as _ 1|_t=_1|e^-iht|_0=__1|(z - h)^-1|_0e^-iztdz , [ laplace ] where the contour @xmath23 encloses all the eigenvalues @xmath24 ( @xmath25 ) of @xmath0 . with the corresponding eigenstates @xmath26
the spectral representation of the resolvent is a rational function : _
1|(z - h)^-1|_0=_j=0^n = , [ resolvent2 ] where @xmath27 , @xmath28 are polynomials in @xmath29 of order @xmath2 , @xmath19 , respectively , with the common denominator @xmath30 these polynomials are readily evaluated by the recursive projection method ( rpm ) @xcite . the expression in eq .
( [ resolvent2 ] ) for @xmath31 can be interpreted as the bosonic spectral density @xmath32 with respect to the state @xmath20 : _
( e)=im _ 0|(e - i - h)^-1|_0= _
[ spectrald ] the amplitude of the return probability then reads as the fourier transform of the spectral density _ 0|_t=_0_(e)e^-ietde .
[ fourier ] analogously , the overlap @xmath22 reads in terms of the resolvent _
i m _ 1|(e - i - h)^-1|_0e^-ietde [ fourier2 ] with _ 0 i m _ 1|(e - i - h)^-1|_0=_j_1|e_je_j|_0(e - e_j ) , [ off_diagonal1 ] provided that the matrix elements are symmetric .
the latter is the case for the hubbard hamiltonian .
the purpose of the subsequent calculation is to determine the evolution of the fock state under the influence of the bose - hubbard hamiltonian of eq .
( [ bhh ] ) . in general ,
this is expressed in the fock base as @xmath33 .
for the n00n state we only need to focus on the coefficients @xmath34 and @xmath35 . comparing the result in eq .
( [ evolution1 ] ) with the expressions in eqs .
( [ fourier ] ) , ( [ fourier2 ] ) , ( [ off_diagonal1 ] ) , it turns out that the fourier transform of @xmath34 and @xmath35 are just the imaginary parts of the matrix elements of the resolvent _ 0(e)=_0im n,0|(e - i - h)^-1|n,0 = _ jn,0|e_je_j|n,0(e - e_j ) [ fourier_c1 ] and _
n(e)=_0im 0,n|(e - i - h)^-1| n,0 = _ j0,n|e_je_j|n,0(e - e_j ) .
[ fourier_c2 ] these two expressions will be called spectral coefficients , where @xmath36 measures the relative weight @xmath37 .
integration over the energy @xmath38 gives 1 for this coefficient .
the coefficient @xmath39 measures the correlation between @xmath8 and @xmath7 due to the product @xmath40 .
the latter is real for a symmetric hamiltonian .
integration over the energy @xmath38 gives 0 for this coefficient .
the bose - hubbard hamiltonian has two simple limits : the local limit @xmath41 and the noninteracting limit @xmath42 . in the local limit for a symmetric double well with @xmath43 pairs fock states @xmath44 , @xmath45 are doubly degenerate eigenstates with energy @xmath46 $ ]
. a perturbation by a small tunneling term will break the degeneracy .
this effect is stronger at lower energies because the parabolic spectrum is denser there .
this agrees with a numerical study @xcite .
the fact that the states @xmath8 and @xmath7 are very close in energy may support the formation of a n00n state . in the absence of particle - particle interaction the bose - hubbard hamiltonian @xmath47 ( i.e. the hamiltonian in eq .
( [ bhh ] ) with @xmath42 ) describes only tunneling .
a straightforward calculation shows that the eigenstate @xmath48 of @xmath47 with @xmath49 has an overlap with the fock states @xmath8 and @xmath7 as n,0|n - k;k= 2 ^ -n/2 , 0,n|n - k;k= ( -1)^k 2 ^ -n/2 .
[ binomi1 ] this implies that the spectral coefficients of eqs .
( [ fourier_c1 ] ) and ( [ fourier_c2 ] ) have a binomial form _
0(e)= 2 ^ -n_k=0^nn
k(e+j(2k - n ) ) [ free_bose3a ] _
n(e)= 2 ^ -n_k=0^nn k(-1)^k(e+j(2k - n ) ) .
[ free_bose3b ] a fourier transformation reveals a periodic behavior of the evolutionary coefficients as c_0(t ) = n,0|e^-iht|n,0=^n(jt ) , c_n(t ) = 0,n|e^-iht|n,0=(-i)^n^n(jt ) .
[ non_int_exp ] thus the evolution of the fock state leads to a n00n state with a probability that decays exponentially with @xmath2 .
this is a consequence of the fact that for an increasing @xmath2 the particles disappear in the @xmath3dimensional hilbert space because there is no constraint due to interaction .
the double well with the two fock states @xmath8 , @xmath7 as possible initial states can be treated within the rpm .
this method is based on a systematic expansion of the resolvent @xmath50 , starting from the initial base @xmath51 .
the method can also be understood as a directed random walk in hilbert space .
this means that in comparison with the conventional random walk the directed random walk of the rpm visits a subspace @xmath52 only once and never returns to it . in terms of @xmath2 bosons , distributed over the double well , the subspace @xmath52 is spanned by the base @xmath53
a step from @xmath52 to @xmath54 is given by the hamiltonian @xmath0 in such a way that @xmath54 is created by acting @xmath0 on @xmath52 ( cf .
this step is provided by the tunneling of a single boson .
thus , the directed random walk follows a path with increasing numbers @xmath55 .
the directed random walk is the main advantage of the rpm which allows us to calculate the matrix elements @xmath56 , @xmath50 of the resolvent on a @xmath3-dimensional hilbert space exactly .
now we choose @xmath57 for the bose - hubbard hamiltonian . assuming that @xmath2 is even , all projected spaces @xmath52 are two - dimensional and spanned by @xmath58 ( @xmath59 ) .
this leads to a recurrence relation in the base of the two fock states @xmath60 as initial states .
the recurrence relation reads ( app . a ) g_k+1= , g_0= ( k=0,1, ... ,n/2 - 1 ) [ 2d_recursion ] with coefficients a_k+1= [ coeff_a ] b_k+1= [ coeff_b ] and @xmath61 the iteration terminates after @xmath62 steps with g_n/2= , [ termination ] where a_n/2 = n,0|(z - h)^-1|n,0= 0,n|(z - h)^-1|0,n , [ fock1 ] and b_n/2 = 0,n|(z - h)^-1|n,0= n,0|(z - h)^-1|0,n .
[ fock2 ] there exists an invariance of the recurrence relation under the following simultaneous sign changes in eqs .
( [ coeff_a ] ) and ( [ coeff_b ] ) z - z , u - u , a_j - a_j , b_j - b_j .
[ transf ] this implies that a change from a repulsive to an attractive hubbard interaction results in a mirror image with respect to energy of the spectral coefficients c_0(e , u)=c_0(-e ,- u ) , c_n(e , u)=c_n(-e ,- u ) .
[ image ] in the case @xmath63 we have one more variable , namely @xmath64 , @xmath65 and @xmath66 with the following recurrence relations ( app . a ) g_k+1 = , g_0= ( k=0,1, ... ,n/2 - 1 ) [ 2d_recursion1 ] with matrix elements ( @xmath67 ) : a_k+1= [ acoeff_a ] b_k+1= [ acoeff_b ] c_k+1=- [ acoeff_c ] and with @xmath68[z+4un(k+1)-j^2b_k(n+k+1)(n - k)]\ ] ] @xmath69 the final result of the iteration is g_n/2= , [ termination2 ] with a_n/2=n,0|(z - h)^-1|n,0 , b_n/2=0,n|(z - h)^-1|0,n , [ fock3a ] c_n/2=n,0|(z - h)^-1|0,n=0,n|(z - h)^-1|n,0 . [ fock3b ]
the iteration of eqs . ( [ coeff_a ] ) , ( [ coeff_b ] ) for a symmetric double well and the iteration of eqs . ( [ acoeff_a])-([acoeff_c ] ) for an asymmetric double
well gives us , according to eqs .
( [ fock1 ] ) , ( [ fock2 ] ) and ( [ fock3a ] ) , ( [ fock3b ] ) , the following matrix elements of the resolvent @xmath70 these are rational functions of @xmath29 , as shown in eq .
( [ laplace ] ) . for n bosons these are lengthy expressions with @xmath19 poles .
therefore , it is convenient to present the results as plots with respect to energy .
examples of the spectral coefficients @xmath36 and @xmath39 are shown for a symmetric double well with 100 bosons in fig .
[ fig:1 ] and with 20 bosons in fig .
[ fig:3 ] , and for an asymmetric double well with 100 bosons in fig .
. a larger number of bosons shows a richer spectral structure .
the diagonal coefficient @xmath36 in the case of 100 bosons is remarkably different from the off - diagonal coefficient @xmath39 because the latter does not have spectral weight from eigenstates whose energy @xmath24 is larger than the energy of the initial fock state @xmath71 .
the reason for this feature is the double degeneracy of the eigenvalues mentioned in sect .
[ sect : double_well_non ] : the signs of the product @xmath40 for adjacent eigenvalues are opposite to each other .
since the eigenvalues get closer pairwise as we increase their energy , the contribution of the two levels cancel each other for each pair inside the sum of eq .
( [ fourier_c2 ] ) .
this interaction effect is also visible for 20 bosons ( fig .
[ fig:3 ] ) , although the cancellation is incomplete then due to a larger level distance .
this can be considered as an effect of spectral fragmentation , where the spectrum has a nondegenerate low - energy and a degenerate high - energy part , caused by the competition of tunneling and interaction .
the contribution of the two fock states @xmath7 , @xmath8 to the evolution in eq .
( [ evolution1 ] ) is given by the coefficients @xmath72 , @xmath73 . in fig .
[ fig:2 ] the real parts of these coefficients are plotted for 100 bosons .
their evolution indicates a collapse and revival behavior .
the plot of @xmath74 , @xmath75 as a two - dimensional vector for 20 bosons in fig .
[ fig:4 ] shows a complex dynamical behavior that can not be described by a simple equation of motion .
this observation suggests a statistical description with a probability @xmath76 which measures how often certain values of @xmath74 , @xmath75 are visited during the evolution in a period of time .
the result for 20 bosons is plotted in fig .
[ fig:5 ] for @xmath77 , @xmath78 .
it indicates that there is a strong correlation between the coefficients , where the most favored values are @xmath79 . for an asymmetric double well with interaction strength
@xmath80 the spectrum is different because of the absence of double degeneracy of the eigenvalues ( cf .
[ fig:6 ] ) .
there are two `` bands '' , one around @xmath81 , the other around @xmath82 , where the widths of the bands is characterized by the tunneling rate @xmath83
. moreover , the off - diagonal part @xmath39 appears closer to zero energy and its values are very small .
this indicates that the off - diagonal part has overlaps with energy levels which are different from those of the diagonal part @xmath36 . for the evolution
only the latter contribute substantially , preventing the system to create a n00n state .
in order to understand the evolution of an isolated many - body bosonic system , we start with noninteracting bosons ( i.e. @xmath42 ) of sect . [ sect : double_well_non ] .
the spectral properties are characterized by ( i ) equidistant energy levels with distance @xmath83 and ( ii ) a binomial weight distribution of the energy levels .
the evolution of a fock state is characterized by a periodic behavior with a single frequency @xmath84 as a direct consequence of the equidistant energy levels .
the amplitudes for visiting the initial fock state @xmath8 or the complimentary fock state @xmath7 vary with @xmath85 or @xmath86 , respectively .
this implies for a large number @xmath2 of bosons that ( i ) these states are visited only for a very short period of time and ( ii ) the two fock states are visited at different times .
thus the formation of a n00n state is very unlikely for noninteracting bosons .
a simple qualitative picture for the general evolution of the fock state is the random walk in hilbert space . in case of noninteracting bosons
the particles can walk independently of each other which enables them to explore the entire hilbert space spanned by the fock states without restriction .
a simultaneous overlap of @xmath87 with both fock states @xmath8 and @xmath7 is very unlikely then .
once we have turned on the boson - boson interaction the particles experience a mutual influence which restricts their individual random walks .
this is related to the fact that the systems stays much longer in the energetically ( almost ) degenerate fock states @xmath8 and @xmath7 than in the noninteracting case ( cf .
[ fig:4 ] ) and , what is even more important here , they can have a simultaneous overlap with both fock states , such that they create a n00n state . in terms of the spectral properties the interaction modifies ( i ) the energy levels , which are not equally spaced , and ( ii ) the weight distribution of the levels , which are not binomial any longer ( cf . fig .
[ fig:1]-[fig:6 ] ) .
this , of course , affects also the evolution of the fock state which is more complex now , since many different frequencies are involved .
a particular feature is the spectral fragmentation ( cf .
[ fig:1 ] ) , where only a part of the spectrum contributes to the off - diagonal coefficient @xmath39 .
this is a kind of hilbert - space localization , where transitions to the high - energy part of the hilbert space are completely suppressed , similar to the self - trapping found in the hartree approximation of the bose - hubbard model @xcite .
it should be noticed , however , that spectral fragmentation appears at a much weaker interaction strength than the self - trapping effect . for @xmath88 , which is the threshold for self - trapping @xcite ,
there is only one eigenvalue with significant weight @xmath37 @xcite .
thus it is not clear whether or not the two effects are directly are connected . for the asymmetric double
well the situation is different due to the existence of two `` bands '' and the absence of the double degeneracy .
the main consequence is the absence of a support for the formation of n00n states because the off - diagonal coefficient @xmath35 is strongly suppressed . from this observation
we can conclude that the evolutionary entanglement is much more favorable in the symmetric double well .
this is in agreement with the results of the multi - orbital hartree calculation of ref .
@xcite . in conclusion
, we have studied the evolution of a bosonic fock state @xmath8 in a double well and found that a local particle - particle interaction supports the formation of a n00n state , provided that the interaction is not too strong .
this is accompanied by a fragmentation of the spectrum .
the latter is characterized by the fact that only eigenstates with energies less than the energy of the initial fock state can be reached in the evolution .
this interaction effect causes a hilbert - space localization and prevents the evolution of the fock state to disappear in the depth of the hilbert space .
this is the main reason for a favorable creation of a n00n state .
the appearance of a n00n state is suppressed though for strong interaction because then the restriction of the hilbert space is too severe and does not allow to reach the complementary fock state @xmath7 .
i am grateful to a. streltsov for discussing his work on the n00n state .
99 s. trotzky et al . ,
science , bf 319 , 295 ( 2008 ) .
j. estve at al . ,
nature * 455 * , 1216 ( 2008 ) . c. gross et al .
, nature * 464 * , 1165 ( 2010 ) .
m. brune et al . , phys .
* 101 * , 240402 ( 2008 ) .
h. wang et al . , phys .
lett . * 101 * , 240401 ( 2008 ) .
k. ziegler , phys .
a * 81 * , 034701 ( 2010 ) .
e. s. annibale , o. fialko , k. ziegler , arxiv:1010.0527 ( in press ) k. ziegler , arxiv:1012.5848 .
j. hubbard , proc .
soc . a * 276 * , 238 ( 1963 ) .
m. lewenstein et al . adv . in phys .
* 56 * , 243 ( march 2007 ) .
m. greiner et al .
nature * 415 * , 39 ( 2002 ) .
hartmann , f.g.s.l .
brando and m.b .
plenio , nature physics 2 , 849 - 855 ( 2006 ) ; laser & photon .
rev . 2 , 527556 ( 2008 ) . g.j .
milburn , j. corney , e.m .
wright , d.f .
walls , phys . rev . a * 55 * , 4318 ( 1997 ) .
a. smerzi , s. giovanazzi and s.r .
shenoy , phys .
lett . * 79 * , 4950 ( 1997 ) .
mahmud , h. perry and w.p .
reinhardt , j. phys .
b : at . mol .
* 36 * , l265 ( 2003 ) .
salgueiro et al . , eur .
j. d * 44 * , 537 ( 2007 ) .
ji , q. sun , x.c .
xie , and w.m .
liu , phys .
lett . * 102 * , 023602 ( 2009 ) .
m. albiez et al . , phys .
rev . lett .
* 95 * , 010402 ( 2005 ) .
r. gati et al . , new j. phys .
bf 8 , 189 ( 2006 ) .
streltsov , o.e .
alon , l.s .
cederbaum , j. phys .
b * 42 * , 091004 ( 2009 ) .
given is a sequence of projectors @xmath89 ( @xmath90 ) , defined by the recurrence relation @xmath91 with initial conditions @xmath92 , @xmath93 and by the hamiltonian @xmath0 through the properties p_2khp_2k+1=p_2khp_2k+2 , p_2k+1hp_2k = p_2k+2hp_2k .
[ recurse ] the projection of the resolvent @xmath94 defines g_k = p_2(n - k ) ( z - h)_2(n - k)-1 ^ -1 p_2(n - k ) ( 0k
< n ) , [ proj_g ] where @xmath95 is the inverse on the @xmath96projected hilbert space .
then @xmath97 satisfies the recurrence relation g_k=(z - h_k)_2(n - k)^-1 [ projected4a ] with h_k= .
[ effham1 ] of interest is here only the case @xmath98 , where we have from eq .
( [ proj_g ] ) @xmath99 for the specific case of the double well we choose @xmath67 and the projectors @xmath100 @xmath101 with the hubbard hamiltonian of eq .
( [ bhh ] ) the diagonal terms of the effective hamiltonian in eq .
( [ effham1 ] ) read @xmath102 @xmath103|n/2+k , n/2-k\rangle\langle n/2+k , n/2-k|\ ] ] @xmath104|n/2-k , n/2+k\rangle\langle n/2-k , n/2+k| \ .\ ] ] the off - diagonal terms of the effective hamiltonian in eq .
( [ effham1 ] ) read @xmath105 @xmath106 @xmath107 this leads for @xmath43 to eqs .
( [ coeff_a ] ) , ( [ coeff_b ] ) and for @xmath108 to eqs .
( [ acoeff_a ] ) , ( [ acoeff_b ] ) and ( [ acoeff_c ] ) . ) and ( [ fourier_c2 ] ) for 100 bosons with @xmath109 and @xmath110 .
the energy of the initial fock state is @xmath111 .
the spectral fragmentation appears around @xmath112 , where the levels are nondegenerate at lower energies but almost degenerate for higher energies .
this is a consequence of the competition between tunneling and interaction , which the latter wins at higher energies .
, title="fig:",width=377,height=264 ] ) and ( [ fourier_c2 ] ) for 100 bosons with @xmath109 and @xmath110 .
the energy of the initial fock state is @xmath111 .
the spectral fragmentation appears around @xmath112 , where the levels are nondegenerate at lower energies but almost degenerate for higher energies .
this is a consequence of the competition between tunneling and interaction , which the latter wins at higher energies .
, title="fig:",width=377,height=264 ] and @xmath110 .
the energy of the initial fock state is @xmath113 .
the almost degenerate states appear above @xmath114 .
, title="fig:",width=377,height=264 ] and @xmath110 .
the energy of the initial fock state is @xmath113 .
the almost degenerate states appear above @xmath114 .
, title="fig:",width=377,height=264 ] and @xmath110 .
the energy of the fock states @xmath7 and @xmath8 is @xmath115 , respectively.,title="fig:",width=377,height=264 ] and @xmath110 .
the energy of the fock states @xmath7 and @xmath8 is @xmath115 , respectively.,title="fig:",width=377,height=264 ] ( upper panel ) , @xmath73 ( lower panel ) for 100 bosons with @xmath109 . the time scale is given in inverse units of the interaction strength @xmath116 and the energy of the initial state is @xmath111 .
, title="fig:",width=377,height=264 ] ( upper panel ) , @xmath73 ( lower panel ) for 100 bosons with @xmath109 .
the time scale is given in inverse units of the interaction strength @xmath116 and the energy of the initial state is @xmath111 .
, title="fig:",width=377,height=264 ] of @xmath74 , @xmath122 over a time period of @xmath119 for 20 bosons with @xmath123 .
the axes are scaled by a factor 100 and the vertical axis is in arbitrary units .
this plot indicates a strong correlation between the two spectral coefficients , supporting the formation of a n00n state .
, width=453,height=264 ]
|
we study the creation of a bosonic n00n state from the evolution of a fock state in a double well .
while noninteracting bosons disappear quickly in the hilbert space , the evolution under the influence of a bose - hubbard hamiltonian is much more restricted .
this restriction is caused by the fragmentation of the spectrum into a high - energy part with doubly degenerate levels and a nondegenerate low - energy part .
this degeneracy suppresses transitions to states of the high - energy part of the spectrum . at a moderate interaction strength
this effect supports strongly the dynamical formation of a n00n state .
the n00n state is suppressed in an asymmetric double well , where the double degeneracy is absent .
|
fedorov _ et al .
_ @xcite have recently measured the line - shape @xmath1 of photoemitted electrons in the ferromagnetic metal gd . for photoelectrons associated with a photohole in the @xmath2 surface band
, they find somewhat different life - time broadening depending on whether the emitted electron has up ( majority ) or down ( minority ) spin .
they interpret the source of lifetime broadening to be electron - phonon scattering for the majority spin component of the photohole , and electron - magnon scattering for the minority spin component of the photohole . although the arguments given by fedorov _ et al . _
seem perfectly sensible , nevertheless , this interesting diversity suggests a need for theoretical inquiry .
the magnitude of electron - phonon scattering in gd has previously @xcite been estimated , with results roughly agreeing with the assigned majority spin equilibration rate .
electron - magnon scattering has not previously been estimated . here
i argue that the `` s - f '' or `` ferromagnetic kondo lattice '' model allows reasonable estimates without free parameters .
i make an extreme model for the nature of surface electron and surface magnon states : surface electron states have amplitude 1 on the top layer and zero elsewhere , while magnon states at the surface @xcite are simply the bulk bloch states , ignoring surface boundary conditions . using this model , and the measured mass @xmath3 of the surface hole band , the zero temperature equilibration rate of minority spin holes more than 25mev from the top of the hole band
is predicted to be @xmath40.10 ev , agreeing with the experiment .
the generic hamiltonian for the `` s - f '' or `` ferromagnetic kondo lattice '' model couples electron bands @xmath5 with energy @xmath6 ( independent of spin , so far ) to localized spins @xmath7 located on atoms at lattice sites @xmath8 : @xmath9 the electron bands derive from the outer atomic orbitals @xmath10 with wavefunctions @xmath11 where @xmath12 is the spin part .
this model should describe the low - energy spin - related physics of a metal like gd , with three conduction electrons per atom in orbitals derived from atomic @xmath13 and @xmath14 states of gd .
these three orbitals have a magnetic interaction with the well - localized @xmath15 half - filled @xmath16 shell .
the exchange parameter @xmath17 comes from the atomic hund s rule trying to keep `` electron '' spins parallel to `` core '' spins .
the same hamiltonian , in the @xmath18 limit , is known as the `` double exchange hamiltonian '' @xcite and is very popular right now @xcite for discussions of hole - doped lamno@xmath19 . the zero temperature phase diagram of the s - f model ( for a single @xmath13 band ) as a function of filling and @xmath20band - width ratio has been computed approximately @xcite .
ferromagnetic order occurs over a wide range of parameters , with a curie temperature ( @xmath21=292k for gd ) proportional to @xmath17 .
electron bands acquire a spin splitting proportional to @xmath17 .
@xcite used the model to calculate ( in random - phase approximation ) the spin - wave dispersion , which was measured for gd by koehler _
@xcite , and which has been studied using spin - density functional theory by perlov _
_ @xcite . in lowest - order spin - wave approximation
, we replace the fourier - transformed spin operator @xmath22 in eq .
( [ eq : ham ] ) by spin - wave creation and destruction operators @xmath23 and @xmath24 using @xmath25 , @xmath26 , and @xmath27 .
the @xmath28 term in lowest order gives spin splitting @xmath29 , lowering the energy of bands with spin parallel to the localized spin @xmath30 and raising the other bands equally .
the @xmath31 terms give rise to spin - flip scattering events . in this note
i estimate the rate @xmath32 at which a single out - of - equilibrium hole in a surface band relaxes back toward equilibrium by spin - flip processes . the rate can be found from @xmath33 , where the leading feynman diagram for the self - energy @xmath34 is shown in fig .
[ fig : feynman ] . only the one - magnon process is considered .
this approximation can be questioned on the grounds of fallibility of the `` migdal approximation '' for electron - magnon processes @xcite .
however , it is a proper first estimate , the only one which can be reliably computed , and seems to me unlikely to make a large error when @xmath35 .
two magnon processes have been considered by lutovinov and reizer @xcite .
the actual states @xmath36 of interest are surface states of spin * primarily * but not exclusively up . the probability @xmath37 ( or @xmath38 for short ) that the spin state is up , is less than one .
the amplitude of the up - spin component of the wavefunction is @xmath39 and is close to 1 .
the corresponding probability @xmath40 ( or @xmath41 for short ) that the spin state is down , is non - zero for two reasons .
first , the spin - orbit interaction is not small and mixes spin states .
second , @xmath42-spins on gd atoms may deviate from perfect alignment by quantum and thermal fluctuations , and the conduction states are locally locked by hund s rule in the same spin orientation as the @xmath42-spins .
the second process is a renormalization of electron bands by magnon processes .
a combination of the two effects is seen experimentally @xcite as a small minority spin component in the photoemitted electron .
this component is sometimes referred to as a `` shadow band '' and has received recent theoretical treatments @xcite .
an elementary derivation of the `` migdal '' result follows from the standard `` golden rule '' rate equations of boltzmann theory .
suppose the down - spin component of state @xmath5 has population @xmath43 .
if this deviates from the equilibrium fermi - dirac population @xmath44 , then it will evolve back toward equilibrium according to @xmath45 @xmath46 \label{eq : emission}\end{aligned}\ ] ] @xmath47 \label{eq : absorption}\end{aligned}\ ] ] here @xmath48^{1/2}$ ] is the matrix element for the process @xmath49 scattering to @xmath50 by emission of the magnon @xmath51 or absorption of the magnon @xmath52 .
the factor @xmath53 , where @xmath54 is the number of unit cells and @xmath55 the number of atoms in the sample , comes from normalization of the spin - wave eigenvector .
there are two branches of spin waves , with amplitude @xmath56 on each atom .
this is discussed in the appendix .
a shorthand is used that @xmath57 and @xmath58 stand for the energy and occupancy of the quasiparticle state @xmath59 , @xmath60 and @xmath61 stand for the energy and occupancy of the quasiparticle state @xmath50 , @xmath62 and @xmath54 stand for the energy and occupancy of the magnon state @xmath52 , and @xmath63 and @xmath64 stand for the energy and occupancy of the magnon state @xmath51 . at temperature @xmath65 ,
depletion of an excess population toward equilibrium occurs both by emission and absorption of thermal magnons . each process ( emission or absorption )
has a time - reversed process which enhances the population , the `` scattering in '' terms with opposite sign . in thermal equilibrium , scattering out and in occur at equal rates .
this `` principle of detailed balance '' guarantees that the two parts of @xmath66(emission ) cancel each other when the distributions @xmath54 and @xmath67 become the equilibrium distributions @xmath68 and @xmath42 , and similarly for @xmath66(absorption ) .
now make the assumption that all quasiparticles are in equilibrium except for a particular state @xmath59 of interest , whose population ( @xmath67 ) deviates from equilibrium ( @xmath42 ) by @xmath69 .
then the rate equation [ eq : grule ] takes the form @xmath70
@xmath71 + \delta(\epsilon-\epsilon^{\prime}+\omega ) [ n+f^{\prime } ] \right\ } \label{eq : tau}\end{aligned}\ ] ] except for the factor @xmath72 , this magnon - limited scattering rate is a perfect analog of the usual phonon - limited quasiparticle relaxation rate from migdal theory . eq .
( [ eq : tau ] ) can equally well be derived by evaluation of the feynman diagram fig .
[ fig : feynman ] , analytic continuation of matsubara frequencies @xmath73 to complex frequency @xmath74 , and use of @xmath75 .
the energies @xmath57 and @xmath62 used to evaluate eq .
( [ eq : tau ] ) are the experimental quasiparticle and magnon energies , which is mandated by the occurrence of the full , not bare , green s functions in the self - energy graph of fig .
[ fig : feynman ] .
reasonable simplifying assumptions yield a `` zero parameter '' estimate of eq .
( [ eq : tau ] ) for @xmath32 .
first , we specialize to a state @xmath36 which is in an occupied surface band .
this state at low @xmath76 is dominantly spin up , but as seen experimentally , it has a spin down component , @xmath77 .
the green s function @xmath78 and self - energy are @xmath79 matrices in spin space .
eigenstates of the 2 @xmath80 2 matrix @xmath81 are the spin - split quasiparticles , of which only the lower ( primarily majority spin ) state is relevant .
the energy of this state will be denoted @xmath82 , with no band or spin index needed , and only a two - dimensional wavevector @xmath83 rather than a three - dimensional @xmath84 .
spin - resolved photoemission selectively depopulates a single spin component of this state .
when the photoemitted electron has spin down , there is a hole in the down - spin component of the of the majority spin hole band .
this component ( with distribution function @xmath85 ) decays to equilibrium with rate @xmath86 .
we assume that @xmath87 is independent of @xmath83 , and similarly @xmath88
. then we have @xmath89 where @xmath90 is the relevant density of decay channels . at @xmath91 , @xmath90 is @xmath92 we assume that the spin waves retain their bulk character .
therefore , although electron energies depend only on the two - component @xmath83 , the spin - wave energy @xmath93 depends on all three components of wavevector , the @xmath74 direction being normal to the surface . the first delta function in eq .
( [ eq : d1 ] ) is a process where an empty state @xmath94 lying above the fermi surface is filled from a state @xmath83 which scatters into it by spin - wave emission . in other words
, it refers to decay of an electron lying in a state @xmath83 which is above the fermi energy by at least a spin - wave energy .
the second delta function is a process where the filled state @xmath94 lying below the fermi surface scatters into an empty state @xmath83 by spin - wave emission ; in other words , it describes decay of a hole in the state @xmath83 ( below the fermi energy by at least a spin - wave energy . )
it is only this second process which is seen in the photoemission data .
let us also assume that the photohole state @xmath83 lies below the fermi energy by at least 25 mev , the maximum spin - wave energy , in which case @xmath95 is guaranteed to be 1 .
the @xmath96 integration then gives @xmath97 where @xmath98 is the normalized spin - wave slope @xmath99 and @xmath100 is the number of spin - wave states @xmath101 with fixed @xmath102 , but any value of @xmath96 , which conserve energy , _
i.e. _ which satisfy @xmath103 .
examining the measured dispersion curves @xcite , the normalized slope can lie between 0 and a maximum which is not very different from @xmath104=25 mev . if @xmath105 is not greater than @xmath106
, then there is a fairly good chance ( something like 50% probability ) that there are 2 values , @xmath107 , such that the spin - wave state @xmath101 obeys energy conservation . in other words
, @xmath100 can be expected to take the value 2 or 0 with about equal probability provided @xmath108 , and is definitely 0 elsewhere . the remaining integral is @xmath109 this integral should be evaluated using the correct dispersion relations for the surface electron state and for the spin - wave states , with allowance for boundary conditions and wave - function amplitudes altering the matrix elements . however
, this requires a large `` first - principles '' calculation of uncertain reliability .
therefore , a slightly cavalier estimate is in order .
assume that the surface state has parabolic dispersion @xmath110 .
experimentally this state is seen @xcite to disperse downwards in energy with effective mass @xmath111 , where @xmath112 is the electron mass .
the inequality @xmath108 is obeyed in an annular region shown in fig .
[ fig : circle ] .
the area of this region is @xmath113 and @xmath114 . inside this region
we assume @xmath115 has an average value of 1 and the normalized slope has an average value @xmath116
. then the density of decay channels is @xmath117 this is @xmath57-independent because of the 2-dimensionality of the surface band .
a more careful treatment , yielding exactly the same result , is in the appendix .
finally we evaluate the decay rate eq .
( [ eq : tau1 ] ) by choosing @xmath29 to be the spin splitting of the surface state , @xmath118 ev @xcite , @xmath119 to be 7/2 , and the fractional up - spin probability @xmath120 @xcite .
these assumptions yield @xmath121 ev for all photohole states @xmath83 which are not too much closer than 25 mev to the top of the surface state band . for states closer to the top
, the decay rate should diminish because of reduction of the number of decay channels @xmath90 .
it is also assumed that @xmath76 is fairly low . at higher @xmath76 ,
the rate is enhanced by thermal spin - wave population , but diminished by diminishing fractional up - spin probability @xmath122 .
the estimate @xmath121 ev agrees with experiment @xcite to greater precision than the uncertainties of the model .
the calculation supports the interpretation that the source of minority spin line - broadening is spin - flip decay . here
i try to illuminate some of the approximations in the model and also to demonstrate the surprising reliability of the estimate @xmath123 which comes from combining eqs .
( [ eq : tau1 ] ) and ( [ eq : d3 ] ) .
spin waves in gd have been fitted @xcite with the model @xmath124 where the spin wave eigenfrequencies are @xmath125 .
the elements of the matrix are defined by @xmath126 \nonumber\\ v_{12}(\vec{q})&= & -sj_{12}(\vec{q } ) \label{eq : vmatrix } \\
j_{11}(\vec{q})&= & { \rm re } \sum_{\vec{r } } j(\vec{r})e^{i\vec{q}\cdot\vec{r } } \nonumber\\ j_{12}(\vec{q})&= & { \rm re } \sum_{\vec{r } } j(\vec{r}+\vec{\tau } ) e^{i\vec{q}\cdot ( \vec{r}+\vec{\tau } ) } \label{eq : jelements}\end{aligned}\ ] ] where @xmath127 are fitted exchange coupling constants , @xmath128 runs over the translation vectors of the hexagonal lattice , and @xmath129 with @xmath74-component @xmath130 gives the position of the second atom in the two - atom basis of the hcp crystal structure .
the frequencies and density of states calculated from this model fit experiment @xcite well , and are shown in fig .
[ fig : disp ] .
notice that the two branches @xmath131 are degenerate at the top of the brillouin zone @xmath132 ( directions @xmath133 in fig [ fig : disp ] . )
these branches can be `` unfolded '' into a zone twice as large in the @xmath74-direction ; @xmath134 is the extension of @xmath135 under the mapping @xmath136 .
an accurate analytic approximation for eq .
( [ eq : d2 ] ) is possible provided lindgard s fitting parameters , eq .
( [ eq : jelements ] ) are slightly simplified by truncating off farther than first neighbor planes in the @xmath74 direction .
the alteration of @xmath137 caused by this approximation is shown in fig .
[ fig : disp ] .
the change is fairly small , and could be largely compensated by a further tuning of the nearer neighbor couplings . with this approximation , the one extended magnon branch @xmath138 has frequency given by @xmath139 then the normalized slope @xmath98 is given by @xmath140 the remaining integral eq .
( [ eq : d2 ] ) can be considered an integral over @xmath141 running over the annulus of fig .
[ fig : circle ] , @xmath142 where @xmath143 is the energy of the spin wave with wavevector @xmath144 which scatters the hole out of state @xmath83 into state @xmath145 . for a given two - vector @xmath146 ,
one searches over @xmath96 to find whether there is an energy conserving solution . either
there are no energy conserving spin - wave states , or else there are @xmath147 such solutions at @xmath148 .
consider a path in @xmath145-space at fixed azimuth @xmath149 shown as a dotted line in fig .
[ fig : circle ] .
the outer circle is states @xmath145 which are degenerate in energy with the starting state @xmath83 .
unless @xmath145 is the same as @xmath83 , there is no zero - energy spin wave which can couple these states . moving down the dotted line to higher energy electron states which can fall into the hole at @xmath83 by magnon emission , one finds the state labelled 1 which has just enough energy difference that an energy conserving magnon transition is found .
the magnon has @xmath150 because it is the least energy magnon allowed to couple on the dotted line .
moving farther down the dotted line , one comes eventually to the state labelled 2 which is the highest energy electron state which can fall into the hole in state @xmath83 by magnon emission .
the magnon has @xmath151 because it is the highest energy magnon .
for both these extreme states ,
@xmath152 , the slope @xmath13 is zero ( the magnon energy is quadratic in @xmath96 near @xmath150 and @xmath153 . ) the integrand of eq .
[ eq : d4 ] diverges at the two end points .
however , it is an integrable divergence , and in fact , almost exactly independent of the variables @xmath154 and @xmath155 .
we observe that in going from states 1 to 2 along the dotted line , the components @xmath156 and @xmath157 are not changing much , which allows us to set @xmath158 and @xmath159 to constants during the @xmath145-integration .
then we have @xmath160 this is exactly eq .
( [ eq : d3 ] ) .
finally it is appropriate to mention the hidden assumptions about surface and bulk states .
it is implicitly assumed that the electron surface state has a wavefunction of unit amplitude on the surface layer and zero amplitude elsewhere .
this kind of state will be absolutely insensitive to the @xmath74-component of the magnon wave - vector , and the electron - magnon matrix element will be @xmath161^{1/2}$ ] for all @xmath83 and @xmath145 .
suppose instead a surface electron wavefunction of amplitude @xmath162 on each of the top two layers .
this state will couple to the @xmath135 branch of magnons with the full matrix element @xmath163 , but will not see the @xmath134 branch .
conversely , a surface wavefunction which has amplitude @xmath56 on the top two layers , with a sign change , will couple only to the @xmath134 branch and not to the @xmath135 branch .
it is not reasonable that the net coupling to magnons should depend much on the depth or details of the surface electron wavefunction ; therefore we should ask where has the missing magnon coupling gone when the electron wavefunction extends two layers down instead of one .
the answer is in the orthogonal electron wavefunction on the top two layers , with opposite phase relation between layer 1 and 2 .
since by assumption there is only one surface state , the orthogonal state is not an eigenstate but a superposition of bulk states .
the missing magnon coupling is from the surface state into these bulk states . for the actual gd surface state ,
how much of the magnon - induced scattering is to bulk and how much is to surface states is an unknown element .
the extreme model used here hides this problem .
the justification is belief that the net scattering probability should have a tendency to be conserved , _
i. e. _ to be weakly dependent on depth .
similarly , we have not asked what is the nature of the magnon states near the surface , but instead assumed that we can use bulk magnon states . instead
, it might be that a surface band of magnon states grabs all the spectral weight
. then details would be quite different , but over - all coupling strength should be similar . c. zener , phys
. rev . * 82 * , 403 ( 1951 ) ; p. w. anderson and h. hasegawa , phys . rev . * 100 * , 675 ( 1955 ) ; p .- g . de gennes , phys . rev . * 118 * , 141 ( 1960 ) ; k. kubo and n. ohata , j. phys .
33 * , 21 ( 1972 ) .
j. a. hertz , k. levin , and m. t. beal - monod , solid state commun .
* 18 * , 803 ( 1976 ) ; i. grosu and m. crisan , phys .
b * 49 * , 1269 ( 1994 ) ; m. h. sharifzadeh amin and p. c. e. stamp , phys . rev .
letters * 77 * , 3017 ( 1996 ) .
|
the `` s - f model , '' also known as the ferromagnetic kondo lattice , contains a description of band electrons coupled to localized spins which is an appropriate description of the magnetic part of the low - energy physics of gd metal . here
the model is used to estimate the lifetime broadening of the minority spin component of the surface electron band in ferromagnetic gadolinium metal at temperatures below the curie temperature .
the low temperature result @xmath00.1 ev agrees nicely with a measurement by fedorov _
et al . _
|
stellar masses are now becoming a standard measure for galaxies , and are being used to trace the evolution of the galaxy population in terms of star formation rates and morphologies ( e.g. bundy et al . 2005 ; conselice et al .
2008 ; cowie & barger 2008 ) .
however , stellar mass only traces one aspect of galaxy mass , and ideally and ultimately , we aim to measure galaxy total masses , that include contributions from stars , gas , and dark matter .
galaxies are believed to be hosted by massive dark matter haloes that make up more than 85% of their total mass , and thus clearly tracing the co - evolution of galaxies and their haloes is a major and important goal .
one very powerful method for measuring the total masses of galaxies is to measure their clustering .
clustering measurements are independent of photometric properties , and as such they can be used to highlight fundamental properties of galaxy populations without assumptions concerning stellar populations or mass profiles .
halo clustering is a strong function of halo mass , with more massive haloes more strongly clustered , providing a method to study the relationship between galaxy properties and dark matter halo masses .
we present the first general study of the clustering properties for a stellar mass selected sample of galaxies up to @xmath2 .
we carry this out by measuring the correlation length and amplitude for galaxies selected with stellar masses @xmath3 within the palomar observatory wide - field infrared survey ( powir ) .
the palomar observatory wide - field survey ( bundy et al . 2006 ; conselice et al .
2007 ) was designed to obtain deep near - infrared photometry over a 1.5deg@xmath4 area , using the wide field infrared camera ( wirc ) on the palomar 5 m telescope .
it covers the extended groth strip ( davis et al .
2007 ) and three other fields covered by the deep2 survey ( davis et al .
2003 ) . optical imaging from the canada - france - hawaii telescope cover all fields , using the cfh12k camera in b- , r- and i - band ( coil et al .
deimos spectroscopy was acquired as part of the deep2 survey on the keck telescope .
around 20% of our k - band selected galaxies have a secure redshift ( up to z=1.4 ) . for the galaxies without spectroscopic redshift , we have determined their photometric redshifts based on our brijk photometry and our spectroscopically confirmed sample .
using our photometry we also derived stellar mass measures for our sample of galaxies ( bundy et al . 2006 ; conselice et al .
2007 ) . based on models from bruzual & charlot ( 2003 )
, we derived accurate stellar masses in the range @xmath5 over our range of redshifts with an error of 0.2 - 0.3 dex . in this work
, we study the average halo masses of samples of galaxies selected by their stellar - masses and redshifts .
in order to link these galaxies with their environment and their large scale structure , we quantify their distribution in the sky at different scales . we first measure the 2-point angular correlation function @xmath6 for our sample using the landy & szalay ( 1993 ) estimator , as shown in figure 1(a ) . in order to fit the clustering reliably ,
we avoid the small - scale excesses due to possible multiple galaxy occupation of a single dark matter halo . assuming a top - hat redshift distribution for galaxies in each of our narrow redshift bins , we also derive the correlation length @xmath7 from the amplitude of the angular correlation using the relativistic limber equation ( magliocchetti & maddox 1999 ) . as shown
if figure 1(b ) , we find that correlation lengths vary from @xmath8mpc to @xmath9mpc for galaxies selected by stellar masses with @xmath10 .
the correlation measurements of our different mass - selected samples indicate that the most massive galaxies are more clustered than less massive ones at all redshifts .
furthermore higher redshift samples are more clustered than lower redshift ones at a given mass range ( foucaud et al .
2010 ) .
the standard cold dark matter model predicts that at any redshift more massive dark matter haloes are on average more clustered than lower mass systems . to better understand where galaxies reside ,
their spatial distribution can be directly compared with the distribution of dark matter haloes as predicted by mo & white ( 2002 ) .
the very strong clustering shown by @xmath11 massive galaxies implies that they are hosted by very massive dark matter halos , consistent with progenitors of present - day elliptical galaxies ( foucaud et al .
furthermore we can directly extract dark matter halo masses from the models , and compare directly this mass with the stellar mass of the hosted galaxy , deriving the stellar - to - dark matter mass fraction for our different samples .
mpc of the range over which our data are fitted , in order to avoid any excess of pairs due to multiple halo occupation . *
( b ) * evolution of the spatial correlation length @xmath7 with redshift , for our different mass - selected samples ( foucaud et al .
the measurements are compared with prediction of the evolution for dark matter haloes ( mo & white 2002 ) and from the literature ( meneux et al .
, width=288 ]
in figure 2(a ) , we compared our stellar - to - dark matter mass fractions with measurements from different studies in the literature based on various methods to estimate the masses of dark matter haloes ( rotation curves , galaxy - galaxy lensing , groups ) .
as expected , the stellar - mass - to - total - mass ratio is far lower than the universal baryonic mass fraction , measured from wmap5 ( komatsu et al .
furthermore we confirm at high redshift the tight correlation found between the stellar - to - dark matter mass fraction and dark matter halo masses in the higher mass regime .
overall , we find that , in all our redshift ranges , more massive systems have a lower ratio of stellar - to - halo - mass .
this implies that at @xmath12 galaxies in the centres of the most massive dark matter halo stop increasing their stellar mass while their halos keep growing .
this also implies a limit to how much stellar mass a galaxy can have , with a cut off at a few times @xmath13 .
similarly , a decreasing stellar mass fraction with increasing halo mass has been observed in groups and clusters ( e.g. gonzalez et al .
2007 ) , with the majority of the baryonic mass taking the form of galaxy satellites , while the remaining being under the form of gas . as shown in figure 2(b ) , the stellar - to - dark matter mass fraction also increases with lower redshift at a given stellar mass .
this effect is likely linked with ` contamination ' of the massive sample over time by later forming galaxies hosted by less massive haloes . since the stellar masses of the most massive galaxies do not increase much with time , due to a quenching of their star formation and a weak merging rate ( drory & alvarez 2008 ) , the average mass of the haloes for these galaxies decreases , and the stellar mass ratio increases .
this is in agreement with a ` halo downsizing ' effect ( neistein et al .
2006 ) , i.e. more massive haloes are formed earlier than lower mass haloes .
we have exploited the near infrared data set from the palomar / deep2 survey to investigate the clustering evolution of the most massive galaxies observed from @xmath14 to @xmath15 . by deriving the mass of the host dark matter haloes from clustering analyses
, we show that the stellar fraction is lower in more massive dark matter haloes , and that more massive galaxies inhabit more massive dark matter haloes which form earlier , in agreement with downsizing effects .
the global picture resulting from this study is that the most massive galaxies are formed early in massive dark matter halos . at some stage
the evolution of these massive galaxies stops and henceforth evolves passively , while their halo keep growing in mass , with their baryonic mass in the form of satellites or gas .
bruzual & charlot , 2003 , mnras , 344 , 1000 bundy , ellis & conselice , 2005 , apj , 625 , 621 bundy , ellis , conselice , et al . , 2006 , apj , 651 , 120 coil , newman , kaiser , et al . , 2004 , apj , 617 , 765 cole , norberg , baugh , et al . , 2001 , mnras , 326 , 255 conselice , bundy , ellis , et al . , 2005 ,
apj , 628 , 160 conselice , bundy , trujillo , et al . , 2007 , mnras , 381 , 962 conselice , bundy , u , et al . , 2008 , mnras , 383 , 1366 cowie & barger , 2008 , apj , 686 , 72 davis , faber , newman , et al . , 2003 ,
spie , 4834 , 161 davis , guhathakurta , konidaris , et al .
, 2007 , apj , 660 , l1 drory & alvarez , 2008 , apj , 680 , 41 foucaud , conselice , hartley , et al . , 2010 , mnras , 406 , 147 gonzalez , zaritsky & zabludoff , 2007 , apj , 666 , 147 komatsu , dunkley , nolta , et al . , 2009 ,
apjs , 180 , 330 landy & szalay , 1993 , apj , 412 , 64 magliocchetti & maddox , 1999 , mnras , 306 , 988 mandelbaum , seljak , kauffmann , et al . , 2006 , mnras , 368 , 715 meneux , guzzo , garilli , et al . , 2008 , a&a , 478 , 299 mo & white , 2002 , mnras , 336 , 112 neistein , van den bosch , & dekel , 2006 , mnras , 372 , 933 van starkenburg , van der werf , franx , et al . , 2008 ,
a&a , 488 , 99 weinmann , van den bosch , yang & mo , 2006 , mnras , 366 , 2
|
we present a study on the clustering of a stellar mass selected sample of galaxies with stellar masses @xmath0 at redshifts @xmath1 , taken from the palomar observatory wide - field infrared survey .
we examine the clustering properties of these stellar mass selected samples as a function of redshift and stellar mass , and find that galaxies with high stellar masses have a progressively higher clustering strength than galaxies with lower stellar masses .
we also find that galaxies within a fixed stellar mass range have a higher clustering strength at higher redshifts .
we further estimate the average total masses of the dark matter haloes hosting these stellar - mass selected galaxies . for all galaxies in our sample
the stellar - mass - to - total - mass ratio is always lower than the universal baryonic mass fraction and the stellar - mass - to - total - mass ratio is strongly correlated with the halo masses for central galaxies , such that more massive haloes contain a lower fraction of their mass in the form of stars .
the remaining baryonic mass is included partially in stars within satellite galaxies in these haloes , and as diffuse hot and warm gas .
we also find that , at a fixed stellar mass , the stellar - to - total - mass ratio increases at lower redshifts .
this suggests that galaxies at a fixed stellar mass form later in lower mass dark matter haloes , and earlier in massive haloes .
we interpret this as a ` halo downsizing ' effect .
|
the photon - pion transition process @xmath2 provides a good example for tests of qcd .
this is because , at tree level in pqcd , there only involve electro - magnetic interactions .
the short distance interactions can be calculated in a definite way .
the amplitude for the process @xmath2 can be expressed as @xmath3 , where @xmath4 means the pion momentum , @xmath5 and @xmath6 denote the momentum and polarization of the real photon and @xmath7 is the virtuality of the virtual photon .
the qcd dynamics are contained in the form factor @xmath8 . in pqcd
, the leading form factor @xmath8 can be expressed as f_(q^2)=4c__0 ^ 1 dx with @xmath9 the charge factor and @xmath10 the pion wave function . in the high energy limit @xmath11 ,
the asymptotic of the form factor @xmath12 can be evaluated and have an extremely simple form @xcite @xmath13 that it is only in terms of the pion decay constant , @xmath14 mev and @xmath15 . however , the above asymptotic is about @xmath16 lower than the upper end of the experimental data @xcite .
there have been many explanations about this difference between the theory and the experiment . for expample , the inclusion of nlo correction in @xmath17 @xcite and the transverse structure of the pion wave function @xcite .
it has also been shown @xcite that the data can also be described by the brodsky - lepage interpolating formula @xcite for both @xmath11 and @xmath18 of @xmath8 @xmath19 where @xmath20 gev@xmath21 .
this implies that the nlo power correction , the @xmath0 correction , might be important . in this paper
, we shall present a perturbative approach to calculate the nlo power correction for form factor @xmath1 .
the calculations are related to the terminology of the collinear expansion @xcite .
the expansion has the following features:(1 ) it preserves individual gauge invariance of the soft function and the hard function;(2 ) it can systematically separate the leading twist ( lt ) contributions from the next - to - leading twist ( nlt ) contributions ; ( 3 ) it encounters the high twist contributions from non - collinear partons , wrong spin projection and higher fock states;(4 ) it is a twist - by - twist expansion and free from twist mixing problem;(5 ) it is a feynman diagram approach such that the parton picture for high twist contributions is preserved .
our main results are summarized as follows .
we shall employ collinear expansion to derive the lt and nlt contributions for tree diagrams of processes @xmath2 .
the nlt contributions to @xmath8 involve four nlt pion das .
two of them come from non - vanishing masses of the valence quarks . by the help of equations of motion ,
the four das are reduced to two das .
the asymptotic forms for the remaining das are employed .
the theory appears in a good agreement with the data .
the organization of the remaining text is as follows .
we sketch the collinear expansion for the process @xmath2 in sec .
ii . the explicit formula up to order @xmath22 of the process @xmath2 are given in sec .
iv is devoted to conclusions .
we sketch the procedures of performing collinear expansion for @xmath2 .
let @xmath23 represent the lowest order amplitude for @xmath24 as depicted in fig .
1(a ) and ( b ) .
the @xmath25 denotes the amplitude for partonic subprocess and the @xmath26 represents the meson da .
the @xmath27 means convolution integral over the loop momentum @xmath28 and the traces over the color indices and spin indices .
if we assign the momentum of the final state photon in the minus light - cone direction , @xmath29 , and the momentum of the light meson in the plus light - cone direction , @xmath30 , then the leading configuration for the process can be a function of the collinear momentum @xmath31 with @xmath32 the fraction of the momentum of the meson carried by the partons . @xmath33 and @xmath34 represent light - like vectors in @xmath35 and @xmath36 directions and satisfy @xmath37
the first step is to perform a taylor expansion for the parton amplitude [ st1 ] _
p(k)=_p(k = x p ) + ( _ p)_(x , x ) w^_^k^^+where we have assumed the low energy theorem _
p(k)|_k = x p=(_p)_(x , x ) . and
have employed @xmath38 and @xmath39 . the leading term @xmath40 contains leading , next - to - leading and higher twist contributions in accord with the spin structures of the leading parton amplitude @xmath41 .
that is those terms are proportional to @xmath42 or @xmath43 .
the terms proportional to @xmath42 would project collinear @xmath44 pair from the meson , while those terms proportional to @xmath43 would not diminish only when the @xmath44 pair carry non - collinear momenta .
the second step is to substitute the leading parton amplitude @xmath41 into the convolution integral with the meson wave function @xmath45 to extract the nlt contribution @xmath46 from the leading one @xmath47 [ st2 ] _
p=_p_0 + _ p_1 + , where @xmath48 and @xmath49 denote the leading and next - to - leading meson das , respectively .
the high twist da @xmath49 contains both short distance and long distance contributions .
the short distance contributions of @xmath49 arise from the non - collinear components of @xmath28 . by the equations of motion
, the non - collinear components of @xmath28 will induce one quark - gluon vertex @xmath50 and one special propagator @xmath51 @xcite .
because that the special propagator is not propagating , the quark - gluon vertex and the special propagator should be included into the hard function , @xmath41 . in this way
, we may factorize @xmath49 as @xmath52 and absorb the short distance piece @xmath53 into @xmath41 .
it leads to the third step [ st3 ] _
p_1&=&_p ( ( _ 1^h)_w^_^ ( _ 1^s)^^ ) & = & ( _ p_1^h)_w^_^ ( _ 1^s)^^ , where @xmath54 containing covariant derivative @xmath55 is implied .
notice that the light - cone gauge @xmath56 assures @xmath57 .
the second term of eq.([st1 ] ) also contribute to high twist corrections , as it convolutes with @xmath48 . the momentum factor @xmath58
will be absorbed by @xmath48 to become a coordinate derivative , denoted as @xmath59 .
consider another nlt contributions @xmath60 from fig .
1(c ) and ( d ) , where @xmath61 contains gauge fields .
we have employed the approximation that @xmath62 is the leading term of @xmath63 .
this comes to the fourth step : [ st4 ] ( _ p)_w^_^ _ 1,^^ + ( _ p)_w^_^_1,a^(_p)_1(_1^s ) , where we have employed @xmath64 .
however , it will be found that @xmath65 diminishes as convoluting with twist-4 da @xmath66 ( see below definition ) . up to nlo
, we may drop the @xmath65 term and arrive at the factorization for tree amplitudes _ 0 + _ 1_p_0 + ( _ p_1^h)_1 where @xmath49 means @xmath67 .
there involves only one nlt da @xmath49 for nlo power corrections .
to proceed , we need to consider the factorizations of the spin indices , the color indices and the momentum integrals over loop partons . for factorization of spin indices ,
we employ the expansion of the meson da into its spin components as _ 0,1 & = & _ _ 0,1^where @xmath68 means dirac matrix @xmath69 .
the factorization of the color indices take the convention that the color indices of the parton amplitudes are extract and attributed to the meson das .
the factorization of the momentum integral is performed by making use of the fact that the leading parton amplitudes depend only on the momentum fraction variables @xmath70 .
the identity can always be used ^1_0 dx_i(x_i - k_in)=1 .
the choice of the lowest twist components @xmath71 of @xmath72 is made by employing the power counting .
consider @xmath73 meson whose high twist da @xmath74 has the fermion index @xmath75 and the boson index @xmath76 .
the fermion index @xmath75 arise from the spin index factorization for @xmath77 fermion lines connecting da and parton amplitude and the boson index @xmath76 denotes the @xmath78 power of momenta in previous collinear expansion and the @xmath79 gluon lines as @xmath80 .
we may write [ pc1 ] ^_1_f;_1_b=_i ^_i-1e^_1_f;_1_b_i^i where @xmath81 denotes a small scale associated with da .
spin polarizers @xmath82 denote the combination of vectors @xmath83 , @xmath84 and @xmath85 .
variable @xmath86 represents the twist of da @xmath87 .
the restrictions over projector @xmath88 are [ pc2 ] n__je^_1_f;_1_j_b_i=0 which are due to the fact that polarizers @xmath82 are always projected by @xmath89 .
the dimension of @xmath74 is determined by dimensional analysis [ pc3 ] d()=3f+b-1 by equating the dimensions of both sides of eq.([pc1 ] ) , one can derive the minimum of @xmath86 [ pc4 ] _
i^ = 2f+b+[1-(-1)^b ] .
it is obvious from eq.([pc4 ] ) that there are only finite numbers of fermion lines , gluon lines and derivatives contributes to a given power of @xmath90 .
the factorization theorem for the nlo power correction should be proven in order to have a confident pqcd formalism . before this can be done
, we can at least show that the collinear expansion is compatible with the conventional approach for proving the factorization theorem .
the conventional approach is based on the factorization of the soft divergences from the ultraviolate divergences arising from the radiative corrections .
the soft divergences are shown to be cancelled or absorbed by the meson wave function such that the parton amplitude is free from the soft divergences .
the ultraviolate divergences can be absorbed by the parton amplitude .
we shall assume that the factorization of the soft and ultrviolate divergences can be done upto nlo power correction .
then , we can show that the the collinear expansion for tree diagrams can be straightforwardly extended to those diagrams containing arbitrary loop corrections .
the starting point is to notice that the collinear expansion for the one loop corrections in the collinear region of the radiative gluons can be written down as [ cla1 ] & & ( _ 0^(0 ) + _ 1^(0)+_0^(1 ) + _ 1^(1))| _ & & _ j=0,1 _ p^(0)_0^(j ) + _
p^(0)(_1^h)^(0)_1^(j)where superscript @xmath91 denote tree and one loop corrections , respectively .
this is because , as the collinear gluons with momentum @xmath92 go through the fermion lines , the valence fermion momenta behave similarly to those in the tree level expansion .
this leads to the fact that the collinear expansions for one loop amplitudes in collinear region can be performed just like for tree amplitudes .
the soft gluon corrections can not affect the collinear expansion .
the cancellations of double logarithms are assured in light - cone gauge by adding ladder and self energy diagrams .
the one loop corrected parton amplitudes are determined by subtracting the amplitudes in collinear and soft regions from the full one loop amplitudes .
following the standard considerations @xcite , the lt parton amplitude @xmath93 and nlt parton amplitude @xmath94 are infrared finite , and the soft divergences are absorbed by @xmath95 and @xmath96 .
the one loop factorization is derived up to nlt order [ cla2 ] & & ( _ 0^(0 ) + _ 1^(0)+_0^(1 ) + _ 1^(1 ) ) & & [ _ i=0 ^ 1 _ p^(i)]&&+[_i=0 ^ 1 _ j=0^i_p^(j)(_1^h)^(i - j ) ] .
the generalization to arbitrary loop orders can be obtained by iteration .
suppose that = ( _ p)_0_0+(_p)_1_1 where & = & _ i=0^n ^(i ) , ( _ p)_0=_i=0^n ( _ p)_0^(i ) , _ 0,1=_i=0^n _ 0,1^(i)(_p)_1&=&_i=0^n ( _ p)_1^(i ) , where ( _ p)_1^(i)=_j=0^n _ p^(j)(_1^h)^(i - j ) .
the above factorization still holds for @xmath97 order corrections , since the collinear gluons can not attach to the parton amplitudes .
the remaining proof of factorization requires the cancellations of double logarithms of soft divergences , the single soft logarithms absorbed by pion das and the infrared finiteness of the parton amplitudes .
this can be achieved by standard analysis ( see e.g. @xcite ) and it is left to other publish @xcite .
the lowest - order diagrams are displayed in fig . 1 .
by applying the previous collinear expansion to separate the lt and nlt contributions in a factorized form ,
we may write the result as @xmath98 where @xmath6 denotes the polarization vector of the final state photon . the leading order of @xmath8 is calculated from fig.1(a ) and ( b ) as @xmath99 where the charge factor @xmath100 . @xmath101 and @xmath102 mean the charges of @xmath103 and @xmath104 quark in unites of the elementary charge .
the nlo of @xmath8 is evaluated from fig.2 @xmath105}{q^4x(1-x)}\ .\end{aligned}\ ] ] the relevant das are expressed explicitly as follows @xmath106 the tensors @xmath107 and @xmath108 are defined as @xmath109 and @xmath110 .
the nonvanishing valence quark mass can also contribute to nlo correction .
we use the scheme that the partons involved in the hard function are massless .
this does not affect the final conclusion . by assigning new contributions from the quark mass operator @xmath111
, we get the result @xmath112}{q^4x(1-x)}\end{aligned}\ ] ] where two twist-4 das are introduced @xmath113 note that @xmath114 and @xmath115 are related to the conventional twist-3 pion da @xmath116 and @xmath117 @xcite by a factor @xmath118 , the average quark mass .
the das @xmath119 and @xmath120 are dependent and reducible under equations of motion to g^(x)&=&-_^_0 ^ 1 d x_1_0^e^i ( x_1-x ) e^ix&&0||q(0)_(n)_q(n)|(p_1 ) , and ^(x)&=&-d_^_0 ^ 1
d x_1_0^e^i ( x_1-x ) e^ix&&0||q(0)_5_(n)_q(n)|(p_1 ) . due to the factor @xmath121 for @xmath122
, @xmath123 becomes dominate .
the normalization of @xmath10 is fixed from process @xmath124 such that @xmath125 for asymptotic ( as ) model and @xmath126 for chernyak - zhitnitsky ( cz ) model @xcite .
similarly , the normalization of @xmath127 is determined from the axial anomaly @xmath128 @xcite to yield @xmath129 for as model and @xmath130 for cz model .
the comparison between the experiment and our result indicates that the data is in more favor of as model than of cz model ( see fig.3 ) .
this is consistent with that conclusion made by jakob _
we have shown that the collinear expansion for @xmath2 can be systematically performed .
the @xmath22 power correction for @xmath8 has been evaluated in terms of four twist-4 das .
the effects of nlo power correction are estimated to account for the data @xcite .
applications of the collinear expansion to other processes can be straightforwardly perfromed .
we have also limited ourselves to the tree level .
the factorization theorem for the nlo power correction should be proven if we require a confident pqcd formalism .
as we have shown , the collinear expansion is compatible with the conventional approach for proving the factorization theorem .
the collinear expansion can be performed order by order for radiative corrections .
s.j . brodsky and g.p .
lepage , phys .
lett.*b 87 * , 359 ( 1979 ) ; phys .
lett . * 43 * , 545 ( 1979);phys
. rev.*d 22 * , 2157 ( 1980 ) .
cleo collaboration ( j. gronberg et .
al . ) , phys . rev . *
d 57 * , 33(1998 ) .
p. kroll and m. raulfs , phys .
b * 387 * , 848 ( 1996 ) .
r. jakob _ et al .
_ , j. phys .
* g 22 * , 45(1996 ) .
et al . _ ,
phys . rev . *
d 53 * , 6582(1996 ) .
s.j . brodsky and g.p .
lepage , phys . rev .
* d 24 * , 1808 ( 1980 ) .
r.k.ellis , w.furmanski and r. petrozio , nucl . phys . * b 207*,1(1982);*b 212 * , 29(1983 ) ; j. qiu , phys . rev . *
d 42*,30(1990 ) .
yeh , contribution to the proceedings of the fourth international workshop on b physics and cp violation , ise - shima , japan , february 18 - 23 , 2001 .
yeh ( in preparation ) .
braun and i.e. filyanov , z. phys .
* c 48 * , 239(1990 ) .
v.l . chernyak and a.r .
zhitnitsky , phys .
rep . * 112 * , 173 ( 1984 ) .
fig.3 the pion - photon transition form factor @xmath131 calculated with different distribution amplitudes : the asymptotic ( solid line ) and cheryak - zhitnitsky ( dash line ) .
the experimental data are taken from @xcite .
|
we propose a perturbative evaluation for the next - to - leading - order ( nlo ) @xmath0 power correction to the photon pion transition form factor @xmath1 .
the effects of the nlo power correction are analyzed .
# 1/ # 1eq .
( [ # 1 ] ) # 1#2eqs .
( [ # 1]),([#2 ] ) # 1#2
|
design of chaotic electronic circuits offers a great challenge to the research community for the last three decades @xcite-@xcite .
the motivation for designing a chaotic electronic circuit comes mainly from two facts : first , one can ` observe ' chaos , and can also control the dynamics of the circuit by simply changing the physically accessible parameters of the circuit , e.g. , resistor , capacitor , voltage levels , etc . ;
second , there are multitude of applications of chaotic electronic oscillators starting from chaotic electronic communication to cryptography @xcite-@xcite .
a chua s circuit is the first autonomous electronic circuit where a chaotic waveform was observed experimentally , established numerically , and proven theoretically @xcite-@xcite .
moreover , it established that chaos is not a mathematical abstraction or numerical artifact , but is a very much realizable phenomenon .
after the advent of chaotic chua s circuit , a large number of works have been reported on different methods of realization of this circuit .
all of these realizations are mainly centered around following goals : inductor - free realization of chua s circuit , and realization of chua s diode .
the reason behind the inductor - free realization lies in the fact that the presence of an inductor makes the circuit bulky , unsuitable for ic design , less robust , etc . in the inductor - free realization ,
the inductor in chua s circuit is replaced by a general impedance converter ( gic ) that requires at least two op - amps .
another approach to this end is wien - bridge based chua s circuit variant @xcite , where a wien - bridge oscillator is cascaded properly with the chua s diode . in this context ,
@xcite and @xcite report a general algorithm for designing chua s circuit , in which it has been shown that a sinusoidal oscillator can be converted into chua s circuit by incorporating a proper type of nonlinearity .
different realizations of chua s diode , which is the only locally active nonlinearity in the circuit , have been reported , e.g. , chua s diode using voa @xcite , @xcite , cfoa @xcite , ic realization of chua s diode @xcite , etc . also , four element chua s circuit has been reported in ref.@xcite , which is the minimum component chua s circuit till date .
a detailed overview on the chua s circuit implementation can be found in ref .
@xcite and @xcite .
recently , an inductor - free realization of the chua s circuit based on electronic analogy has been reported @xcite , @xcite , which provided the advantage of exploring the system dynamics in the negative parameter space ( i.e. , negative values of @xmath0 and @xmath1 of a classical chua s circuit @xcite ) .
present paper reports an inductor - free realization of chua s circuit , in which we have properly cascaded a single amplifier biquad based active band pass filter with a chua s diode .
the system has been mathematically modeled with three - coupled first - order autonomous nonlinear differential equations .
it has been shown through numerical solutions of the mathematical model and real world hardware experiments that the circuit shows all the behaviors of a classical chua s circuit , e.g. , fixed point , limit cycle formation , period doubling cascade , chaotic spiral attractors , double scrolls , and boundary crisis .
occurrence of chaos has been established through lyapunov exponents and experimental power spectrum .
the paper is organized in the following manner : next section describes the proposed circuit and its mathematical modeling .
numerical simulations , computations of nonlinear dynamical measures , e.g. , lyapunov exponents are reported in section [ sec:3 ] .
section [ sec:4 ] gives an account of the experimental results .
finally , section [ sec:5 ] concludes the outcome of the whole study .
the proposed circuit is shown in fig.[f1](a ) .
this circuit has two distinct parts : ( i ) a second order narrow band active bandpass filter ( bpf ) , and ( ii ) parallel combination of a grounded capacitor ( @xmath2 ) and chua s diode ( @xmath3 ) .
@xmath4 node of the bpf is connected to the parallel combination of capacitor @xmath2 and chua s diode ( @xmath3 ) through a passive resistor @xmath5 .
note that , the inductor - capacitor parallel combination of a classical chua s circuit has been replaced by a resonator circuit , which is , in the present circuit , an active bpf ; resonant ( or center ) frequency of the bpf can be controlled by simply varying the resistors @xmath6 or / and @xmath7 ( instead of varying a capacitor , or an inductor ) .
to keep the number of active component lesser we have chosen the single amplifier biquad based narrow band active bpf proposed by deliyannis and friend @xcite , @xcite , which consists of only one single amplifier ( in the form of an op - amp ) , two capacitors having same values ( @xmath8 ) , and four resistors ( @xmath9 ) .
this active bpf is a second order system with the following transfer function : @xmath10 here , @xmath11 .
proper choice of @xmath12 and @xmath13 ( and hence @xmath14 ) makes the coefficient of @xmath15 in the denominator negative , which in turn brings the pole of the circuit to the right half of the @xmath15-plane , resulting a sinusoidal oscillation .
real time dynamics of the system can be expressed in terms of the following two coupled first - order autonomous differential equations @xcite : [ eq1b ] @xmath16 equation ( [ eq1b ] ) can be written as : + @xmath17 it can be seen from ( [ eqn1c ] ) that the bpf can be made to oscillate sinusoidally if one can ensure the following condition : @xmath18 .
this in turn gives the condition of oscillation of the sinusoidal oscillator as : @xmath19 ; subsequently , the frequency of oscillation of the sinusoidal oscillator can be obtained from the condition:@xmath20 .
thus the frequency of oscillation is obtained as : @xmath21 .
chua s diode is characterized by the function @xmath22 , where @xmath23 is the voltage drop across @xmath2 .
@xmath22 is a three segment piece wise linear function with @xmath24 and @xmath25 as the slopes , and @xmath26 is the break point voltage for those segments .
variation of @xmath22 with @xmath23 has been shown in fig.[f1](b ) ( showing only the three segment region ) .
@xmath22 is defined by : @xmath27 in the present work , we have chosen the voa implementation of chua s diode @xcite as shown in fig.[f1](c ) .
the dynamics of the proposed circuit can be described by three - coupled first - order autonomous nonlinear differential equations in terms of @xmath23 , @xmath4 and @xmath28 ( fig .
[ f1 ] ) : [ eq2 ] @xmath29 where , @xmath30 .
+ eq.([eq2 ] ) has been written in the following dimensionless form using dimensionless quantities : @xmath31 , @xmath32 ( @xmath33 ) , @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 : [ eq4 ] @xmath40,\\ \dot{y } & = & -kx+k(r_1 + 1)y-\frac{(2k+1)r_2}{(k+1)}z,\\ \dot{z } & = & -(k+1)x+(k+1)(r_1 + 1)y-2r_2z.\end{aligned}\ ] ] here , @xmath41 is defined as @xmath42 where , we have defined the following parameters : @xmath43 , and @xmath44 .
numerical integration has been carried out on ( [ eq4 ] ) using fourth - order runge - kutta algorithm with step size @xmath45 .
we have chosen the resistor @xmath6 as the control parameter ( remembering that @xmath6 controls the gain and center frequency of the bpf ) ; thus in the numerical simulations , @xmath46 acts as the control parameter keeping other parameters fixed at : @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath51 .
note that , the values of @xmath52 and @xmath53 are generally used in a classical chua s circuit .
it has been observed that , with increasing @xmath46 , for @xmath54 , the circuit shows a fixed point . for @xmath55 ,
the fixed point loses its stability through hopf bifurcation , and a stable limit cycle emerges . at @xmath56 ,
limit cycle of period-1 becomes unstable and a period-2 ( p2 ) cycle appears .
further period doubling occurs at @xmath57 ( p2 to p4 ) , and @xmath58 ( p4 to p8 ) . through period doubling bifurcation , the circuit enters into the regime of spiral chaos at @xmath59 . with further increase of @xmath46 , at @xmath60 ,
the circuit shows the emergence of a double scroll attractor .
finally , the system - equations show diverging behavior beyond @xmath61 , indicating boundary crises . phase plane representation ( in @xmath62 plane ) for different @xmath46
is shown in fig.[fphyx ] , which shows the following characteristics : period-1 ( @xmath63 ) , period-2 ( @xmath64 ) , period-4 ( @xmath65 ) , period-8 ( @xmath66 ) , spiral chaos(@xmath67 ) , double scroll ( @xmath68 ) .
the phase plane plots , for the same parameters , in @xmath69 plane is shown in fig.[fphzyx](a)-(e ) also , the same double scroll chaotic attractor in @xmath70 plane is shown in fig.[fphzyx](f ) .
these observations can be summarized through bifurcation diagram with @xmath46 as the control parameter .
bifurcation diagrams are obtained by using poincar section @xcite at @xmath71 with @xmath72 , excluding the transients .
fig.[fbif ] ( upper and middle trace ) shows bifurcation diagrams in @xmath73 and @xmath74 , respectively , for different @xmath46 .
clearly , it shows a period doubling route to chaos .
further , it shows the presence of a period-3 window at @xmath75 . among the other periodic windows ,
more prominent is a period-5 window near @xmath76 interspersed in a double scroll chaos .
plane ) for different @xmath46 : ( a)@xmath63 ( period-1 ) , ( b)@xmath64 ( period-2 ) , ( c)@xmath65 ( period-4 ) , ( d ) @xmath66 ( period-8 ) , ( e)@xmath67(spiral chaos ) , ( f)@xmath68 ( double scroll ) .
( @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath51 ) ] plane ( a)-(e ) , and @xmath70 plane ( f ) for different values of @xmath46 .
parameter values are same as fig.[fphyx ] . ]
( upper trace ) and @xmath74 ( middle trace ) with @xmath46 as a control parameter .
lower trace shows the largest lyapunov exponent ( @xmath77 ) with @xmath46 .
( @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath51).,title="fig : " ] ( upper trace ) and @xmath74 ( middle trace ) with @xmath46 as a control parameter .
lower trace shows the largest lyapunov exponent ( @xmath77 ) with @xmath46 .
( @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath51).,title="fig : " ] ( upper trace ) and @xmath74 ( middle trace ) with @xmath46 as a control parameter .
lower trace shows the largest lyapunov exponent ( @xmath77 ) with @xmath46 .
( @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath51).,title="fig : " ] for quantitative measure of the chaos generated by the circuit model , lyapunov exponents have been computed using the algorithm proposed in @xcite .
[ fbif ] ( lower trace ) shows the spectrum of largest lyapunov exponent in @xmath46 parameter space .
it agrees with the bifurcation diagram .
presence of a positive le and strange attractors ( as the double scroll attractor ) ensures the occurrence of chaotic behavior in the circuit .
the proposed circuit has been designed in hardware level on a breadboard .
since the circuit needs three voas we have chosen ic tl074 ( quad jfet op - amp ) for the present purpose with @xmath78 volt power supply .
bpf has been constructed with @xmath79 nf , @xmath80 k@xmath81 , @xmath82@xmath81 , and @xmath83k@xmath81 .
chua s diode is constructed with following parameters @xcite , @xcite : @xmath84 k@xmath81 , @xmath85 @xmath81 , @xmath86 @xmath81 , @xmath87 k@xmath81 , @xmath88 k@xmath81 , and @xmath89 k@xmath81 .
the grounded capacitor value is taken as @xmath90 nf .
coupling resistor @xmath5 is fixed at @xmath91k@xmath81 ( approx . ) using a @xmath92 k@xmath81 pot . to explore the dynamics of the circuit
we have varied the resistor @xmath6 through a @xmath93 k@xmath81 pot .
all the potentiometers are precision pot having thousand turns .
capacitors and resistors have 5% tolerances . for @xmath94@xmath81 the circuit shows a fixed dc value ( equilibrium point of the circuit ) . for @xmath95@xmath81 a stable limit cycle has been observed with frequency @xmath96hz . at @xmath97@xmath81 ( approx . )
the limit cycle of period-1 loses its stability and a period-2 oscillation emerges .
a period-4 behavior has been observed at @xmath98@xmath81 ( approx . ) .
period-8 is found for @xmath99@xmath81 ( approx . ) .
further decrease in @xmath6 results in spiral chaos in the circuit ( at @xmath100@xmath81(approx . ) ) .
the double scroll attractor is observed at @xmath101@xmath81 ( approx . ) .
the circuit shows a large limit cycle for @xmath102@xmath81 that indicates the occurrence of boundary crises , where the active resistor becomes eventually passive , i.e. , for a large voltage across its terminals , the @xmath103 characteristic of chua s diode is no longer a three - segment curve but , it becomes a five - segment curve .
two additional segments ( situated in two outer regions ) give positive resistance behavior , i.e. , now the instantaneous power consumption becomes positive @xcite .
all the above mentioned behaviors ( except the large limit cycle ) have been shown in fig.[fexpt1 ] ( in @xmath4-@xmath23 space ) and fig.[fexpt2 ] ( in @xmath28-@xmath4 and @xmath28-@xmath23 space ) , which depict the experimental phase plane plots recorded in a real time oscilloscope ( aplab make , two channel , 60mhz ) . apart from the experimental phase plane plots , another very useful tool for exploring the real circuit behavior for a range of circuit parameters is experimental bifurcation diagram @xcite , @xcite . to grab the behavior of the proposed circuit for a range of values of @xmath6 , experimental bifurcation diagram of @xmath28 has been plotted taking @xmath6 as the control parameter ( fig.[fexptbif ] ) .
for seventy different values of @xmath6 in the range of @xmath104 @xmath81 to @xmath105 @xmath81 , we have acquired the experimental time series data of @xmath28 using agilent make infinium digital storage oscilloscope ( 5000 data per set ) and find out the poincar section using the _ local minima _ of the time series data . plotting these for all the @xmath6s gives the experimental bifurcation diagram .
schematic difference between the numerical bifurcation diagram ( fig.[fbif ] ( middle trace ) ) and experimental bifurcation diagram occurs due to the difference in the sampling schemes employed in finding out the poincar section .
however , qualitatively they agree with each other , e.g. , both the diagrams show period doubling route to chaos .
blurring of points in the experimental bifurcation diagram occurs due to the inherent circuit noise that is reflected in the experimental time series data .
also , since we have to acquire the data manually , resolution of the plot is not high enough in comparison with the numerical bifurcation plot .
time series data and the corresponding ffts of @xmath23 are measured using agilent make infinium digital storage oscilloscope ( 500mhz , sampling rate 1gs / s ) and are shown in fig.[fex3 ] .
fig.[fex3](a ) ( upper trace is for time series and the lower trace is for fft ) shows the period-1 oscillation with frequency @xmath96hz ; fig.[fex3](b ) and ( c ) shows the results for spiral chaos and double scroll chaos , respectively .
power spectrum of fig.[fex3](b ) and ( c ) ( lower traces ) are continuous and broad in nature , indicating the occurrence of chaotic oscillations .
it can be seen that the experimental results agree with numerical simulations of the mathematical model of the circuit .
-@xmath23 space ( a ) period-1 at @xmath106@xmath81 ( b ) period-2 at @xmath107@xmath81 ( c ) period-4 at @xmath108@xmath81 ( d ) period-8 at @xmath99@xmath81 ( e ) spiral chaos at @xmath109@xmath81 .
( f ) double scroll attractor at @xmath110@xmath81 .
( other parameters are : bpf:@xmath79 nf , @xmath80 k@xmath81 , @xmath82@xmath81 , and @xmath83k @xmath81 .
chua s diode : @xmath84 k@xmath81 , @xmath85 @xmath81 , @xmath86 @xmath81 , @xmath87 k@xmath81 , @xmath88 k@xmath81 , and @xmath89 k@xmath81 .
the grounded capacitor @xmath90 nf .
@xmath111k@xmath81 ) .
( a)-(e)@xmath4 ( @xmath73 -axis ) : @xmath112 v / div , @xmath23 ( @xmath113-axis ) : @xmath93 v / div .
( f)@xmath4 ( @xmath73 -axis ) : @xmath114 v / div , @xmath23 ( @xmath113-axis ) : @xmath92 v / div . ]
-@xmath4 space ( ( a)-(d)@xmath28 ( @xmath73 -axis ) : @xmath93 v / div , @xmath4 ( @xmath113-axis ) : @xmath112 v / div .
( e)@xmath28 ( @xmath73 -axis ) : @xmath92 v / div , @xmath4 ( @xmath113-axis ) : @xmath114 v / div . ) , and ( f)@xmath28-@xmath23 space ( @xmath28 ( @xmath73 -axis ) : @xmath92 v / div , @xmath23 ( @xmath113-axis ) : @xmath92 v / div . ) ; parameter values are same as used in fig.[fexpt1 ] . ]
( v ) with @xmath6 ( @xmath81 ) as the control parameter .
other parameters are same as fig.[fexpt1 ] . ] for ( a ) period-1 oscillation ( frequency @xmath96hz ) ( b ) spiral chaotic oscillation and ( c ) double scroll chaos ( circuit parameters are same as described in fig.[fexpt1 ] ) .
( frequency span of the fft : @xmath115 khz).,title="fig : " ] for ( a ) period-1 oscillation ( frequency @xmath96hz ) ( b ) spiral chaotic oscillation and ( c ) double scroll chaos ( circuit parameters are same as described in fig.[fexpt1 ] ) .
( frequency span of the fft : @xmath115 khz).,title="fig : " ] for ( a ) period-1 oscillation ( frequency @xmath96hz ) ( b ) spiral chaotic oscillation and ( c ) double scroll chaos ( circuit parameters are same as described in fig.[fexpt1 ] ) .
( frequency span of the fft : @xmath115 khz).,title="fig : " ]
in this paper we have reported a single amplifier biquad based inductor - free implementation of chua s circuit , where we have suitably cascaded an active band pass filter with a chua s diode .
the proposed circuit has been modeled mathematically by three - coupled first - order autonomous nonlinear differential equations .
numerical simulations of the mathematical model of the circuit and the real world hardware experiments confirm that the proposed circuit shows the same behavior that is shown by a classical chua s circuit ( e.g. , fixed point behavior , limit cycle formation through hopf bifurcation , period doubling , spiral chaotic attractor , double scroll attractor , and boundary crisis ) .
the presence of chaos has been established through the lyapunov exponents and experimental power spectrum .
as the circuit is inductor - free it has got all the advantages of an inductor - free circuit ( e.g. suitable for ic design , robustness , etc . ) .
further , instead of varying an inductor or a capacitor , we have used a single resistor ( e.g. , @xmath6 ) as a control parameter to observe all the complex behaviors of the circuit . since the circuit has a large number of choices of control parameters ( viz .
@xmath7 , @xmath5 and @xmath0 ) , one can use these parameters to observe different behaviors of the circuit .
but , the basic bifurcation routes to chaos remains the same for each of the parameters .
the bpf circuit we have used is narrow band ( unlike wien - bridge based circuit@xcite , where the frequency selective network is not narrow band ) , thus one has more control over the center frequency of the circuit . the present circuit can be suitably designed ( with high frequency op - amps ) to generate chaotic oscillations in a high frequency region , and also may be useful for chaos based communication systems .
the author is indebted to prof .
sarkar ( dept . of physics , the university of burdwan , india ) for the useful suggestions and insightful discussions .
also , the author would like to thank the anonymous reviewers for their useful suggestions .
99 ogorzalek , m.j.:chaos and complexity in nonlinear electronic circuits . world scientific series on nonlinear science , series a - vol . 22 ( 1997 ) ramos , j.s . : introduction to nonlinear dynamics of electronic systems : tutorial .
nonlinear dyn . 44 .
314 ( 2006 ) kennedy , m.p .
, rovatti , r. and setti , g. ( eds . ) , chaotic electronics in telecommunications .
florida : crc press .
( 2000 ) banerjee , t. , sarkar , b.c . :
chaos , intermittency and control of bifurcation in a zc2-dpll .
j. electron .
96(7 ) , 717 - 731 ( 2009 ) matsumoto , t. , chua , l.o . ,
komuro , m. : the double scroll.ieee trans .
circuits syst .
32 , 797 - 818 ( 1985 ) kennedy , m.p.:three steps to chaos - part ii : a chua s circuit primer .
ieee trans . ckt .
and syst - i .
40(10 ) , 640 - 656 ( 1993 ) morgul , o. : inductorless realization of chua s oscillator . electron . lett .
31 , 1424 - 1430 ( 1995 ) rocha , r. , medrano - t .
, r.o.:an inductor - free realization of the chua s circuit based on electronic analogy .
nonlinear dyn .
56(4 ) , 389 - 400 ( 2009 ) elwakil , a.s , kennedy , m.p.:chua s circuit decomposition : a systematic design approach for chaotic oscillators .
journal of the franklin institute .
337 , 251 - 265 ( 2000 ) elwakil , a.s , kennedy , m.p.:generic rc realizations of chua s circuit .
j. bifurcation chaos .
10 , 1981 - 1985 ( 2000 ) kennedy , m.p .
: robust op - amp realization of chua s circuit .
frequenz 46 , 66 - 80 ( 1992 ) elwakil , a.s . ,
kennedy , m.p .
: improved implementation of chua s chaotic oscillator using current feedback op - amp .
ieee trans .
circuits syst .
i 47 , 289 - 306 ( 2000 ) cruz , j.m . ,
chua , l.o . : a cmos ic nonlinear resistor for chua s circuit .
ieee trans .
circuits syst .
i 39 , 985 - 995 ( 1992 ) barboza , r. , chua , l. o.:the four - element chua s circuit .
j. bifurcation and chaos 18 , 943 - 955 ( 2008 ) fortuna , l. , frasca , m. , xibilia , m.g .
: chua s circuit implementations : yesterday , today , and tomorrow .
world scientific series on nonlinear science series a - vol .
65 , ( 2009 ) kilic , r. : a comparative study on realization of chua s circuit : hybrid realizations of chua s circuit combining the circuit topologies proposed for chua s diode and inductor elements .
j. bifurcation and chaos 13 , 1475 - 1493 ( 2003 ) rocha , r. , andrucioli , g.l.d.,medrano - t .
, r.o . : experimental characterization of nonlinear systems : a real - time evaluation of the analogous chua s circuit behavior .
nonlinear dyn .
62(1 - 2),237 - 251 ( 2010 ) deliyannis , t. : high - q factor circuit with reduced sensitivity .
4(26 ) , 577 - 678 ( 1968 ) friend , j.j . : a single operational - amplifier biquadratic filter section . in : ieee int . symp .
circuit theory , pp .
189 - 190 ( 1970 ) banerjee , t. , karmakar , b. , sarkar , b.c.:single amplifier biquad based autonomous electronic oscillators for chaos generation .
nonlinear dyn .
62 , 859 - 866 ( 2010 ) nayfeh , a.h . ,
balachandran , b. : applied nonlinear dynamics : analytical , computational , and experimental methods .
wiley , new york ( 1995 ) wolf , a. , swift , j. b. , swinney , h. l. and vastano , j. a. : determining lyapunov exponents from a time series .
physica d. 16 , 285 - 317 ( 1985 ) buscarino , a. , fortuna , l. , frasca , m. and sciuto , g.:coupled inductors - based chaotic colpitts oscillator .
j. bifurcation and chaos . 2 , 569 - 574 ( 2011 ) viana , e.r . ,
rubinger , r.m . ,
albuquerque , h.a .
, de oliveira , a.g . and ribeiro , g.m.:high resolution parameter space of an experimental chaotic circuit .
chaos 20 , 023110 ( 2010 )
|
the present paper reports an inductor - free realization of chua s circuit , which is designed by suitably cascading a single amplifier biquad based active band pass filter with a chua s diode .
the system has been mathematically modeled with three - coupled first - order autonomous nonlinear differential equations .
it has been shown through numerical simulations of the mathematical model and hardware experiments , that the circuit emulates the behaviors of a classical chua s circuit , e.g. , fixed point behavior , limit cycle oscillation , period doubling cascade , chaotic spiral attractors , chaotic double scrolls and boundary crisis .
the occurrence of chaotic oscillation has been established through experimental power spectrum , and quantified with the dynamical measure like lyapunov exponents .
= 1
|
the concept of reaction mechanism plays a major role in chemistry and represents a synthesis of our understanding of the way in which different topological changes in the bonding structure of a reactant or product are correlated as the reaction proceeds .
recent advances in ultrafast lasers@xcite , x - ray@xcite and other spectroscopies as well as in computational chemistry@xcite have made possible the study of most gas - phase and some condensed - phase reactions in molecular detail . however , most experimental investigations of complex reaction mechanisms taking place in liquid environments are still inferred from isotope and solvent ( medium ) effects on the reaction rate@xcite .
consequently , the interpretation of the experimental results as well as the reaction mechanisms inferred are more controversial than those of gas - phase reactions .
computer studies can be useful as a complement to experimental data in cases where experiments alone can not provide a definitive picture of the mechanism of the chemical process .
it is therefore desirable to develop systematic computational approaches to carefully examine the relation between isotope effects and reaction mechanism in condensed phase systems .
however , computational calculations of kinetic isotope effects in condensed - phase reactions can become expensive due to a number of difficulties .
some of the practical challenges involving the calculation of kinetic isotope effects and reaction rates in solution are associated with the fact that accurate descriptions of transformations in which chemical bonds are broken and formed require time - consuming _ ab initio _ electronic structure methods .
computer time limitations become particularly relevant when investigating `` rare events '' such as chemical reactions , especially when the reactions are accompanied by substantial differences in the structure of the solvent . to further complicate matters , quantum effects such as zero - point vibrations and tunneling effects are important in some chemical processes , such as proton transfer reactions . another technical problem in the simulations of reactive systems
is that the statistical resolution of calculations of the reaction rate depends on how many statistically independent configurations are obtained during the simulation : simulations in which a large number of successive configurations have similar configurations of the reactive core or of the solvent molecules suffer from large uncertainties in the calculated reaction rates , precluding any definitive interpretation of the reaction mechanism .
it is therefore critical to develop methods which sample statistically independent configurations along the reaction path rapidly and correctly . in the case where the reaction path can be characterized by means of a small number of reaction
coordinates , accurate , statistically well - resolved calculations of reaction rates can be performed by developing improved methods for computing reaction free energy profiles along these reaction coordinates .
a number of techniques for computing free energy profiles along reaction coordinates have been proposed in the literature , including umbrella sampling@xcite , thermodynamic integration in conjunction with the blue - moon ensemble method@xcite , projection methods@xcite , variable transformation approaches@xcite and guiding potentials@xcite .
the use of molecular mechanics - based guiding potentials was proposed simultaneously and independently by iftimie et al.@xcite and vondele et al.@xcite and implemented in a monte carlo and a molecular dynamics framework , respectively .
the basic idea of the method consists of using a `` fast '' _ molecular mechanics _ potential to guide a computationally intensive _ ab initio _ simulation .
the monte carlo version of the `` guiding '' approach in reference [ was called the molecular mechanics based importance function method ( mmbif ) . it was demonstrated that the utilization of a reasonably accurate molecular mechanics potential as an importance function decreases the correlation of an _ ab initio _ monte carlo calculation by two orders of magnitude .
the method was illustrated on a gas - phase formic acid - water system in which the activated processes involved breaking and forming hydrogen bonds@xcite , and was successfully applied to calculate the kinetic isotope effects in a model gas - phase intramolecular proton transfer reaction@xcite .
one of the major challenges in _ ab initio _ simulations of reactions in condensed phase environments is to thoroughly sample configurations of the system when changes in the solvent occur on long time scales .
for instance , in molecular dynamics simulations of proton transfer reactions in which the collective behavior of the solvent can strongly influence the dynamics of the reaction , the sampling efficiency can be limited by long solvent dielectric relaxation time .
in essence , an independent configuration of the system requires that the equations of motion be propagated for a time which is longer than the dielectric relaxation time .
even simple organic solvents such as tetrahydrofuran , the relevant solvent in this study , have dielectric relaxation times on the order of @xmath0 ps@xcite , so that independent solvent configurations are only obtained after several thousand elementary propagation steps .
more structured solvents such as water , with a dielectric relaxation time of roughly @xmath1 ps@xcite , require even longer propagation for proper sampling of solvent configurations .
the long time scale of structural rearrangements in solvents pose a serious challenge to _ ab initio _ calculations even when the solvent is modeled using molecular mechanics since each propagation step in the dynamics involves a time - consuming _ ab initio _ calculation .
ideally , successive configurations in a simulation involve drastically different solvent and solute configurations .
this is only possible using an artificial dynamics to generate the sequence of configurations .
one way to generate relatively uncorrelated successive configurations is to apply importance sampling ideas . in this paper ,
the molecular mechanics - based importance sampling method is adapted to calculate reaction rates of chemical processes in condensed phases where collective motions of the environment can influence the quantitative features of the chemical process and , in some cases , play a critical role in determining the mechanism of a reaction .
the monte carlo procedure involves separating the task of sampling the configurations of the condensed phase system into two parts .
the first part involves an efficient scheme of updating the solvent configuration while the second focuses on the relatively slow _ ab initio _ calculation of the reactive core .
this approach allows extensive sampling of the molecular mechanical solvent without a significant increase in the overall computational work over a gas phase _ ab initio _ simulation .
the method is applied to study the double proton transfer reaction in an acetic acid - methanol complex solvated by tetrahydrofuran .
the computational study of proton - transfer reactions can be used to understand the conditions for the validity of a well - known conjecture proposed in the physical organic chemistry literature , stating that the breakdown of the rule of geometric mean for kinetic isotope effects , which is a relation@xcite involving ratios of kinetic isotope effects corresponding to different isotopic substitutions at the primary and secondary atoms , is a signature of tunneling for both primary and secondary atoms@xcite .
the consequences of applying the rule of geometric mean to interpret experimentally determined isotope effects are of far - reaching importance : the inferred relationship between the rule of geometric mean and reaction mechanism forms the basis for the proposal that _ multiple _ _ intramolecular _ proton transfer reactions are likely to proceed via a two - step mechanism , in contrast to _ multiple intermolecular _ proton transfers which are believed to proceed via a synchronous pathway@xcite .
the same relationship is at the heart of the recent suggestion that tunneling effects have played an important role in the design of the active sites of some proteins@xcite .
some of the most striking consequences of the rule of geometric mean@xcite appear when studying multiple proton transfer reactions in condensed phases .
the study of reactions involving the exchange of a pair of protons between two molecules may provide insight into the dynamics of certain types of enzymatic reactions in which several functional groups in the active center are properly aligned so that concerted catalysis can occur .
this type of catalysis mechanism is called _ bifunctional catalysis _ and is the principal mechanism responsible for the several orders of magnitude increase in the reaction rate in several important biochemical transformations @xcite .
the double proton transfer reaction between acetic acid and methanol is one of the simplest examples of reactions involving an intermolecular exchange of protons between two molecules and therefore is a good candidate for computationally investigating general aspects of bifunctional catalysis .
the chemical processes occurring in a solution of acetic acid - methanol in tetrahydrofuran ( thf ) have been studied experimentally by gerritzen and limbach@xcite .
the majority species in the system consist of complexes formed from either a single molecule of acetic acid or a single molecule of methanol hydrogen - bonded with a single molecule of solvent .
the minority species in the system consist of linear and cyclic clusters of acetic acid hydrogen - bonded with methanol and solvated by tetrahydrofuran .
the double - proton transfer reaction takes place along the hydrogen bonds of the cyclic clusters .
when the concentrations of acetic acid and methanol are reduced , the only cyclic cluster formed within which double - proton transfer is experimentally observed is formed from one molecule of acetic acid and a single molecule of methanol@xcite ( see fig .
1 ) . in this work
, we will focus only on the intermolecular double proton transfer which takes place in the cyclic cluster formed from one molecule of methanol and a single molecule of acetic acid in a solution of tetrahydrofuran ( thf ) .
the quantum nuclear effects due to the small mass of the hydrogen atoms will be neglected here , and we will concentrate only on the sampling issues . once the sampling method is tested for classical atoms , the method will be developed in a later article to include nuclear quantum effects via centroid transition state theory using sampling ideas similar to those used in our previous studies of malonaldehyde@xcite .
since it is impractical to treat large systems quantum mechanically , one is inevitably faced with the decision of how to combine _ ab initio _ electronic structure methods with more empirical approaches .
one alternative to full _ ab initio _ calculations consists of using a mixed quantum - mechanical / molecular - mechanical ( qm / mm ) description of the system .
however , in general it is problematic to clearly define the physics of the interface region between the quantum and molecular mechanical subsystems . in the best of circumstances ,
the separation can be made without `` cutting '' a covalent bond .
for reactive processes in which the reaction occurs in a small solute which can be simulated in isolation using _ ab initio _ methods , the logical separation between the quantum and molecular mechanical regions is at the solute - solvent level .
that is , the energies of the solute in which the reaction is occurring are calculated using electronic structure methods for the solute in the presence of the solvent , while the solvent energies are described by molecular mechanical potentials .
even after this separation has been defined , one is faced with another technical issue of how an infinite condensed phase system can be represented in a practical fashion .
one common approach utilized in simulations of condensed phase systems is to simulate explicitly a system consisting of a solute surrounded by a small number of solvent molecules periodically replicated in an infinite fashion .
unfortunately , such a periodic replication of the system introduces artifacts .
one such artifact is the existence of spurious long - ranged correlations due to the periodicity of the system . in principle
, such correlations could have a large impact on the reactive process .
the use of periodic boundary conditions also introduces complexity in the electronic structure calculation itself when localized basis sets are used .
an alternative approach is to use toroidal boundary conditions in which part of the solvent is explicitly represented while the influence of the solvent from regions far from the solute is implicitly incorporated . in the present study
we have utilized a quantum mechanics / molecular mechanics / implicit solvent ( qm / mm / continuum ) approach , in which the bond breaking and forming processes taking place in the solute are described by including all solute electron - electron , electron - nuclei and nuclei - nuclei interactions in an electronic hamiltonian .
the interactions between the solvent molecules which form the first few solvation shells ( in practice , @xmath2 thf molecules ) are described using a molecular mechanics potential .
the electrostatic interactions between the solute and the solvent molecules in the first few solvation shells are incorporated into this approach by solving the _ ab initio _ electronic structure equations for the solute in the presence of the electric field generated by the molecular mechanics solvent charges .
the total energy of the system also includes the lennard - jones interaction energy between solute and the solvent molecules in the first solvation shells , and the reaction field energy which accounts for the long - range electrostatic interactions between solute and solvent molecules ( see fig .
the simulations were performed using the toroidal boundary conditions approach@xcite , in which a cutoff distance is used that sets the boundaries of the region within which intermolecular interactions are explicitly counted .
therefore , the only quantum electron - electron , electron - nuclei and nuclei - nuclei interactions which were explicitly counted in the present treatment were those which correspond to atoms separated by less than the cutoff distance , here chosen to be @xmath3 angstroms .
the effects of the neglected interactions in the toroidal boundary conditions approach have been approximately accounted for by adding reaction field corrections@xcite to the electronic energy@xcite : @xmath4 where the sum is over all molecules i inside the primitive cell , @xmath5 is the dielectric constant of the solvent , @xmath6 is the radius of the spherical surface , @xmath7 is the dipole moment of molecule @xmath8 , and @xmath9 is the total dipole moment inside the spherical surface surrounding the molecule @xmath8 .
the correct energetics in hydrogen - bonded systems and in systems undergoing proton transfer reactions is difficult to describe even with _
ab initio _ methods .
in particular , dft studies of weak hydrogen - bonding systems have proved to be particularly difficult and only limited success has been achieved in predicting the geometries and energies for reactant and transition state configurations on the potential energy surface using most exchange - correlation functionals @xcite .
care should therefore be exercised when choosing a particular _ ab initio _ method to calculate proton transfer reaction rates .
the non - local exchange - correlation schemes developed by proynov , vela and salahub @xcite have shown particular promise for the description of hydrogen - bonded systems .
sirois et al .
@xcite have demonstrated that their kinetic - energy dependent exchange functionals ( blap and plap ) performed better than all gga options ( bp86 , pp86 , pw91 ) , blyp , or other hybrid methods ( b3lyp , b3pw91 ) on systems involving intra - molecular hydrogen bonds . the predictions for equilibrium and transition state geometries as well as the energetics was in agreement with high - quality post - hartree - fock calculations [ ccsd(t ) and g2]@xcite . specifically , for the gas - phase cyclic cluster of acetic acid and methanol in fig .
1 , the activation energy using the plap exchange correlation functional was found to be approximately @xmath10 kcal / mol , in excellent agreement with the qcisd value of @xmath11 kcal / mol@xcite . for the simulations described in the present work ,
the energies of different configurations were carried out using a modified version of the lcgto - dft program demon - ks3.4 @xcite using the plap exchange - correlation functional .
the dft electronic structure calculations were carried out as in reference [ ] , where the application of dft electronic structure methods to hydrogen - bonding systems is discussed in detail .
a double-@xmath12 plus polarization ( dzvp ) orbital basis set was used for all atoms and the convergence level for the scf ( self - consistent field ) energy using the auxiliary fitting basis sets @xcite was @xmath13 kcal / mol . in order to implement the qm / mm / continuum approach to compute reaction rates for the double proton transfer reaction in the acetic acid - methanol cluster solvated by tetrahydrofuran
, a sufficiently accurate molecular mechanics description of the interaction between the solvent molecules is needed@xcite . in this work , the opls all atom ( opls - aa ) force field of jorgensen et al.@xcite with a modified electrostatic interaction term has been used to describe the interactions between thf solvent molecules .
the modifications to the electrostatics were designed to improve the gas - phase distribution of the partial charges in a thf molecule as well as to improve the description of polarization effects in condensed phases .
since the local electrostatic environment as well as long ranged polarization effects can influence the proton transfer process , it is important to properly account for the permanent and induced charges in the solvent . to describe all such electrostatic phenomena , all solvent molecules
have been assigned permanent and induced charges .
it can be demonstrated@xcite using second - order perturbation theory that the electrostatic interaction energy between two polarizable molecules @xmath14 and @xmath15 , each of which carries a set of atomic permanent and induced charges , @xmath16 and @xmath17 , with @xmath18 corresponding to the charges on molecule @xmath19 , and @xmath20 and @xmath21 , where @xmath22 are the site charges on molecule @xmath23 , respectively , can be written to a good approximation as : @xmath24 where @xmath25 is the distance between sites @xmath26 and @xmath27 on molecules @xmath19 and @xmath23 . in principle , the induced charge appearing on a given solvent molecule is dependent on its local environment .
one simple way of incorporating solvent polarization effects is to assign each solvent molecule the same average induced charge in a `` mean - field '' fashion .
more sophisticated methods of including polarization effects either assign site polarizabilities or use fluctuating charges distributed at specific locations on each solvent molecule@xcite . such methods when combined with _
ab initio _ electronic structure methods either necessitate an iterative solution of the electronic structure and fluctuating charge distributions or involve dynamical methods in an extended lagrangian system@xcite .
unfortunately , each of these approaches has shortcomings which make them impractical to implement in conjunction with importance sampling monte - carlo methods . in principle
, the iterative minimization of the kohn - sham and fluctuating charge functionals can be implemented within a monte - carlo sampling approach at the cost of additional computational effort .
however , provided the solvent is not very polarizable , the variation of the induced charges on the solvent molecules from their mean due to the presence of the solute is expected to be small .
for this reason , we have utilized fixed induced charges for all solvent molecules .
this corresponds to computing ground state energies for the reactive core @xmath28 based on a kohn - sham functional@xcite @xmath29 $ ] which depends on the ground state electron distribution @xmath30 of the solute in the presence of the fixed permanent and induced external solvent charges . note that this functional includes the electrostatic energy of interaction between the quantum solute and the charges on the molecular - mechanical solvent molecules . in general ,
one complication must be considered when discussing electrostatic interactions in mixed qm / mm systems that arises from the fact that the quantum mechanical electron density can become over - polarized by the molecular mechanical point charges due to the absence of considerations of the fermi repulsion between quantum and molecular mechanical charges@xcite .
such effects are particularly severe when using delocalized basis sets to represent the quantum subsystem , but are less significant when gaussian or other local basis sets are utilized . in the present work , the full coulomb interaction potential has been used to describe electrostatic interaction terms between the solute and the solvent charges without any screening modifications at short distances since all dft calculations for the quantum subsystem use localized basis sets . in our study of the acetic acid - methanol system in a thf solution ,
the permanent charges have been assigned the value @xmath31 , whereas the induced charges are set to a mean - field value of @xmath32 for all indices @xmath8 and @xmath33 , where @xmath34 are the standard charges in the opls force field .
it can be verified that the electrostatic interaction energy calculated using eq .
( [ totalinteractionenergypermanentplusinducedtext ] ) with this set of permanent and average induced charges is precisely the electrostatic energy calculated using the standard set of fixed opls charges . on the other hand , by assigning a mean induced charge to each solvent molecule , the computed values for the gas - phase dipole moment and condensed phase dielectric constant are in better agreement with the corresponding experimental values ( see table 1 ) .
it should be emphasized that a less satisfactory means of describing electrostatic interactions between the solute and the solvent would consist of calculating the electrostatic interactions between the gas phase electron distribution with the solvent charges . in such a scenario ,
the ground state electron distribution for the solute interacts with the solvent via a coulomb interaction of the form @xmath35 where @xmath36 is the gas - phase electron distribution and @xmath37 is defined to be the distance between the ith charge @xmath38 in the solvent and the point @xmath39 .
this approximation corresponds to calculating the solute energy in the condensed phase by a zeroth order approximation for the electronic distribution of the solute , that is , @xmath40
\approx f\left [ \rho _ { 0}(\mathbf{x})\right ] + v.\ ] ] such a description neglects the fact that the ground state solute electron distribution is influenced by the presence of the solvent charges . the influence of the solvent charges on the ground state energy can be accounted for by incorporating a polarization energy of the solute by the solvent .
although such a crude level of description of the electrostatic interactions may be incomplete , it is useful in developing importance sampling monte - carlo schemes based on guiding potentials , such as the molecular mechanics based importance function method ( mmbif ) described in the next section .
the mmbif method@xcite consists of utilizing an auxiliary markov chain with a known asymptotic molecular mechanical distribution to propose trial configurations for an _ ab initio _ based monte carlo simulation . in the method ,
each trial configuration is obtained as the last state in a classical markov chain generated from the current configuration in the _ ab initio _ simulation using various updating schemes .
the proposed configurations are then accepted or rejected in the _ ab initio _
chain according to the usual metropolis - hastings algorithm@xcite . if the previous and new trial configurations in the _ ab initio _
mc chain are denoted by @xmath41 and @xmath42 respectively , the proposed state is accepted with the probability @xmath43 , where @xmath44 is defined to be @xmath45 where @xmath46 and @xmath47 are the potential energies of configuration @xmath48 calculated by _ ab initio _ methods ( here density functional theory , abbreviated dft ) and the classical potential , respectively , and @xmath49 is boltzmann s constant .
it is straightforward to show that this acceptance criterion guarantees that the _ ab initio _ markov chain has the correct limiting boltzmann distribution@xcite , regardless of the number of classical updates used to generate the proposed configuration .
the efficiency of the mmbif approach relies on constructing a molecular mechanics potential for the entire system which approximates the true interactions of the system at a qualitative level . at first glance
, the construction of a molecular mechanics potential for a condensed phase system appears a daunting task given that the electron distribution of the solute changes considerably during the reactive process and is influenced in a complicated fashion by its local environment .
however , it is relatively straightforward to construct a molecular mechanics potential based on simple approximate forms of the interaction potentials , such as that in eq .
( [ interactionsolutesolventcorrect ] ) . for such forms of the potential ,
the construction of the potential is reduced to modeling the reaction in gas phase and calculating effective charges on the gas - phase solute which mimic the correct electron distribution . the approximate form for the potential
can be corrected for using importance sampling methods .
for example , a monte - carlo chain of states generated using an approximate expression for the energy can be manipulated by re - weighting the configurations appearing in the chain by an appropriate factor@xcite .
the efficiency of this approach is highly dependent on the quality of the approximation for the true energy of the system .
one might anticipate that the crude of level of description of the electrostatic interactions in eq .
( [ interactionsolutesolventcorrect ] ) which neglects any polarization effects of the solute by the solvent molecules would introduce large statistical uncertainties at the re - weighting step .
however , the polarization of the solute by the solvent can be approximated by adjusting the charge of the solvent molecules in the expression for the interaction between the solute and the solvent . as will be discussed , the effective charge on the solvent molecules
can be designed to approximate solute polarization effects and thereby improve the statistical resolution of the re - weighting procedure .
the task of constructing a molecular mechanics potential for the gas - phase proton transfer reaction is facilitated by using bond evolution theory considerations .
following these lines , the molecular mechanics description of the acetic acid - methanol complex in the absence of the solvent was created as suggested in reference [ .
the total molecular mechanics energy was decomposed into two components .
the first component of the total energy was written as a sum of harmonic potentials representing the variation of the potential with bond length , bond angle or bond dihedral displacements from their minimum energy values at a fixed value of the control parameter @xmath50 , defined by @xmath51 where the numbering of the atoms is that from fig .
the second component of the total energy was written as a sum of four morse potentials depending on the four @xmath52 bond lengths , plus two effective potentials depending on two parameters , @xmath53 and @xmath54 defined as @xmath55 these effective potentials account for the flow of electronic charge during the reaction , as well as for the fermi repulsion between the oxygen atoms @xmath56 and @xmath57 , and between @xmath58 and @xmath57 .
this second component of the total energy was implemented using the same functional forms as in reference [ . as in the case of the malonaldehyde study , no potential which depends explicitly on the angles @xmath59 or @xmath60
was utilized , although some dependence on the angle @xmath61 was explicitly introduced using a functional form which interpolates between a harmonic potential for transition state values of the parameter @xmath50 , and zero for values of @xmath50 characteristic for the reactant or product configurations .
the complete details for the construction of the guiding potential for the solute can be found in reference [ .
the simplest practical means of incorporating the electrostatic interactions between the solute and the solvent molecules in the molecular mechanics potential is to fit partial charges to atomic sites in the solute to reproduce the gas - phase electronic distribution .
however , since the electron distribution of the solute varies appreciably with the configuration of the solute , the guiding potential must incorporate solute charges which vary as the reaction proceeds . in their bond evolution theory analysis of the tautomerization of malonaldehyde ,
krokidis et al.@xcite found that the total charge in the basins of attraction of the proton and oxygen atoms varies approximately linearly with a control parameter similar to the parameter @xmath50 .
therefore , it can reasonably be assumed that most of the variation of the charges on atomic sites in the solute can be explained by a linear variation with @xmath50 .
an alternative approach to incorporate the solute - solvent interactions can be constructed using simulated tempering methods@xcite .
the advantage of this procedure is that it does not rely on any approximation for the variation of the fitted charges on the solute during the reaction .
the method essentially consists of using an extended state space to gradually turn on electrostatic interactions between the solute and the solvent .
this approach is represented schematically in figure 4 . in the simulated tempering method ,
a markov chain is constructed whose states ( @xmath62 ) are defined on the space formed by the direct product between a finite set of `` electrostatic '' indices , @xmath63 and the entire solvent plus the solute configurational space . in vector notation ,
the states @xmath64 will be denoted @xmath65 , where @xmath66 and @xmath67 represent the solute and solvent degrees of freedom , respectively .
this markov chain , whose generic state is denoted ( @xmath62 ) , is constructed to produce states asymptotically distributed according to the probability density @xmath68 where @xmath69 are constants which will be referred to as the weights for the unnormalized probability densities @xmath70 defined by @xmath71 where @xmath72 .
the potentials @xmath73 contain five components : the _ ab initio _
potential @xmath74 calculated for the gas - phase solute configuration , the molecular mechanics potential @xmath75 describing the interaction between the explicit solvent atoms , the lennard - jones potential @xmath76 describing the dispersive and short - ranged interactions between solute and solvent atoms , the coulomb electrostatic interaction @xmath77 between some charges on the solute atoms and the permanent and induced charges on the solvent atoms , and the reaction field energies @xmath78 describing the long - range solute - solvent electrostatic interactions : @xmath79 the coulomb solute - solvent electrostatic interaction potential between charges @xmath80 on the solute and @xmath81 on the solvent , given by @xmath82 where @xmath83 and @xmath84 are the number of charge sites on the solute and solvent molecules , respectively , and @xmath85 is a scaling factor henceforth called the charge fraction . note that this interaction potential depends on the choice of charges assigned to the solvent and solute molecules . in our calculations ,
the charges on the solute were fitted using the kollman - singh procedure@xcite from the gas - phase electron distribution of the solute .
the use of charges fitted from the gas - phase calculation of the electron distribution does not account for the polarization of the solute by the solvent . to partially compensate for the neglect of this effect , the total charge on the solvent molecules used in eq .
( [ electrosolutesolvent ] ) is set to the sum of the permanent and induced charges on the solvent , @xmath86 .
this is in contrast to the calculation of the electrostatic interaction between solvent molecules in which the effective charges are set to @xmath87 .
this approximation is consistent with first - order perturbation theory in which the polarization energy of the solute by the solvent is approximated by the polarization energy of the solvent by the solute@xcite .
note that the only difference between the potentials @xmath88 for different electrostatic indices @xmath8 consists of the form of the electrostatic interaction energy @xmath77 between the solute and the explicit solvent atoms .
the energy @xmath89 is chosen to be zero in this work implying that @xmath90 , which amounts to turning off solute charges . for this system
all electrostatic interactions between the solvent and the solute are scaled to zero although the solvent and solute still interact through lennard - jones potentials . by design ,
the last electrostatic index is set to unity , @xmath91 so that the energy @xmath92 corresponds to calculating the electrostatic interaction between the solvent and solute atoms using the fitted configuration - dependent solute charges obtained _ via _ the kollman - singh method .
the additional dimension of the solute plus solvent state space represented by the electrostatic index @xmath93 ensures that the solvent configuration adapts smoothly to the solute via a stepwise process in which the charges of the solute atoms gradually interact with the other charges in the system .
it should be noted that an equivalent result can be achieved by using a parallel tempering scheme in which the label @xmath93 corresponds to a stepwise decrease of a `` sampling temperature '' associated with the solute - solvent electrostatic interaction . in our implementation of the importance sampling ,
a markov chain of extended state space configurations was generated by two types of transitions . in transitions of the first type
the electrostatic index @xmath8 was kept fixed while the configuration of the system * @xmath94 * was updated using a transition matrix which leaves @xmath95 invariant .
the way in which the configurations were updated for fixed electrostatic index was dependent on the electrostatic scaling factor .
when the scaling factor was zero both the configuration of the solute and the solvent were updated simultaneously using the mmbif approach .
the background molecular mechanics simulations used to guide the updates were run so as to produce effectively independent but energetically reasonable configurations of the entire system .
the simultaneous update of both solute and solvent configuration is possible in the absence of electrostatic interactions between the solute and solvent since the fitted charges are not used in the molecular mechanics auxiliary chain .
however , after the calculation of the _ ab initio _ dft energy in the acceptance - rejection step of the mmbif method , the fitted charges for the given solute configuration can be calculated . when the electrostatic scaling factor is nonzero and solute - solvent electrostatic interactions occur , the solvent was allowed to adjust to the charge distribution on the solute by updating the solvent configuration while maintaining the configuration of the core reactive region unchanged .
transitions of the second type move the system through the auxiliary parameter space by applying a transition matrix that changes the electrostatic index while leaving the configuration of the solute and solvent unchanged . in this study ,
the method for updating the electrostatic index @xmath8 consisted of using a metropolis algorithm , with a proposal distribution in which the proposed indices @xmath96 and @xmath97 are equally probable . the proposed change of electrostatic index is rejected if @xmath98 is outside the valid range @xmath93 , otherwise it is accepted with probability @xmath99 = \min \left [ 1,\frac{w_{i_{\textrm{new}}}}{\textrm{w}_{i_{\textrm{old } } } } \exp \left ( -\beta \left ( e_{i_{\textrm{new}}}-e_{i_{\textrm{old}}}\right ) \right ) \right ] .\ ] ] the marginal distribution of @xmath8 with respect to the equilibrium distribution is given by @xmath100 where the configuration partition function @xmath101 is defined as @xmath102 if the distributions @xmath103
are all to play a useful role in sampling , the weights @xmath69 should be chosen such that a roughly uniform distribution over @xmath8 is obtained .
since the @xmath101 are initially unknown , suitable values for the weights are found through a trial and error process using preliminary runs . to do this
, an iterative procedure can be used in which the markov chain is simulated using the current values for @xmath69 , and the frequencies @xmath104 with which each distribution is visited are recorded . next , new and improved weights @xmath105 are calculated as @xmath106 for electrostatic index @xmath8 .
if some of the frequencies @xmath104 are zero , various elaborations of the estimation procedure can be used and some of them are summarized in reference [ .
the number @xmath107 of values of the electrostatic scaling parameter @xmath85 used and the actual values of @xmath69 , @xmath93 are chosen by minimizing the average computer time necessary for a new solute configuration to appear with an electrostatic scaling parameter @xmath91 . a good starting estimate for the numbers @xmath69 and @xmath107 can be obtained by optimizing the weights and the number of intermediate chains using only the molecular mechanics guiding potential with a reasonable set of atomic charges on the solute atoms .
to explore the issue of how importance sampling methods can be effectively utilized in simulations of reactions in condensed phases , two different implementations of the mmbif sampling method were analyzed . in the first implementation , the dependence of the solute charges on the reactive state of the system was taken to vary linearly with the control parameter @xmath108 defined in eq .
( [ a_b_1andb_2definitions ] ) . for this simulation ,
the electrostatic interactions between the interpolated charges on the solute and the charges in the solvent were taken to be coulombic . in the second simulation , the simulated tempering method described above was applied to the solvated acetic acid - methanol cluster .
both simulations generate chains of states asymptotically distributed according to a boltzmann distribution based on eq .
( [ interactionsolutesolventcorrect ] ) in which the solute - solvent electrostatic interactions are modeled by calculating the coulomb interaction between the gas phase electron distribution with the charges in the solvent . in both simulations , the desired distribution for the chain of states based on the universal kohn - sham functional @xmath109 $ ] can be recovered at the end of the simulation by re - weighting each of the @xmath110 total configurations by a configuration dependent factor @xmath111 where @xmath112 - \left ( f[\rho_0 ( x_i)]+v \right ) = f[\rho_s(x_i)]-e^{s}(r_{i}^{s } ) - e^{ss}(r_{i}^{s},r_{i}^{s})\ ] ] is the difference in the polarization energy of the solute by the solvent estimated by calculating the energy of the ground state electron distribution in the presence of the solvent charges and the energy of a gas - phase electron distribution interacting with the optimized solvent charges . in all simulations , the calculations for the solvated cluster ( shown in fig .
3 ) have been carried out by treating all nuclei as classical point particles .
the calculations were conducted in the isobaric - isothermal ensemble at @xmath113 atm and @xmath114 k. in order to improve the sampling along the reaction coordinate , an umbrella potential@xcite was constructed for the guiding potential using a self - adaptive scheme .
simulations biased by the converged umbrella potential yielded a uniform sampling of the important regions of the reaction coordinate even though the activation barriers for the proton transfer reaction was on the order of @xmath115 @xmath116 .
two thirds of the _ ab initio _ markov chain transition steps in the simulation using the linearly - interpolated charges on the core were generated using the mmbif method .
the same fraction of base transitions were generated using the mmbif updates when the electrostatic index is zero in the simulated tempering simulation ( see fig .
the rest of the _ ab initio _ markov chain transitions were performed utilizing metropolis single - variable updates using the _ ab initio _ dft energy . as demonstrated in a previous study [ , the role of the metropolis dft updates is to prevent the guiding molecular mechanics markov chain of states from spending a large number of successive steps in those regions of the configuration space where the molecular mechanics density of states in the solute configuration space substantially underestimates the corresponding _ ab initio _ dft density of states . in the previous studies of the formic acid - water@xcite and of the malonaldehyde@xcite systems
it was demonstrated that a useful strategy for optimizing the mmbif method was to separate the variables to be updated in a classical mc step into several groups , with strongly correlated variables grouped together . in the case of the double proton transfer in the cyclic cluster formed by acetic acid and methanol , the vibrations of the two methyl groups
should be relatively uncoupled from the motions of the other atoms in the cluster . applying this separation of which
variables are updated in the mmbif method , the percent of rejections of proposed configurations obtained with the mmbif method was about @xmath117 .
the percent of rejections of the transitions which employed single - variable metropolis updates using the _ ab initio _ potential was @xmath118 .
it is important to compare the computational effort of performing liquid - phase versus gas - phase simulations of chemical reactions within the mmbif approach .
the potentials of mean force obtained using the reaction coordinate @xmath119 describing the double proton - transfer in the acetic acid - methanol system in tetrahydrofuran obtained in the first simulation and in a gas - phase simulation are represented in fig .
although equal cpu times were spent in computing the two potentials of mean force depicted in fig .
5 , the statistical uncertainties for the activation energy calculated for the double proton reaction rate in the solvated system is approximately @xmath120 times larger than the error bar for the activation energy calculated from the gas - phase simulation .
this decrease in the statistical resolution of the computations in the solvated system with respect to the gas - phase system is due to a larger integrated correlation time of the overall markov chain as well as to the fact that @xmath121 of the cpu time in the simulation of the solvated system is dedicated to calculating the molecular mechanics interactions between the _ solvent _ molecules .
although a single _ ab initio _ calculation is roughly @xmath0 orders of magnitude slower than a single update of the configuration of the molecular mechanical solvent , the long dielectric relaxation of the solvent required that many updates be carried out on the solvent before an independent solvent configuration was generated .
this translated into an approximately equal amount of cpu time for the molecular mechanical and quantum mechanical components of the simulation .
the approximation of the variation of the charges fitted using the kollman - singh procedure with the solute configuration by a linear variation with the parameter @xmath50 in equation ( [ a_b_1andb_2definitions ] ) proved quite accurate .
the standard deviation of the differences between the electrostatic interaction energies calculated using these fitted charges obtained via the kollman - singh procedure and the same energies calculated using solute charges which vary linearly with the parameter @xmath50 defined in equation ( [ a_b_1andb_2definitions ] ) is @xmath122 kcal / mol .
this small deviation should be contrasted with the value of @xmath123 kcal / mol representing the standard deviation of the values of the electrostatic interaction energy between the solute and solvent molecules ( see fig .
the distribution of the values of the polarization energy difference @xmath124 defined in eq .
( [ deltaepol ] ) relevant for the final re - weighting of data points is plotted in fig .
as expected , the values of @xmath124 are distributed over a small range of energy values of approximately @xmath125 kcal / mol .
in fact , calculation of the activation energy in the double proton reaction rate in tetrahydrofuran without re - weighting yields an activation energy which differs from the activation energy in fig . 5 only by a statistically irrelevant value of @xmath126 kcal / mol .
it should be emphasized that the small values of @xmath124 in fig .
7 are in part a consequence of approximating the polarization energy of the solute by the solvent by the polarization energy of the solvent by the solute .
we estimated that if the solvent charges used in eq .
( 11 ) were set to @xmath127 , the standard deviation for @xmath124 would have been approximately @xmath128 kcal / mol . the potentials of mean force for the double proton transfer reaction in tetrahydrofuran calculated using the linearly - interpolated charge approach and using the simulated tempering method are identical within error bars .
however , if the error bars for the activation energy in the linearly - interpolated charge method is @xmath125 kcal / mol , they are roughly four times larger in the simulation using the parallel tempering algorithm for comparable cpu times .
the optimal number of electrostatic indices in the simulated tempering approach was found to be @xmath129 . in the optimized
setting for the simulated tempering method in which equal amounts of time were spent at all values of the electrostatic index , the gibbs free energy differences @xmath130 computed for consecutive values of the index were found to be approximately @xmath131 kcal / mol .
the gibbs free energy @xmath132 of the electrostatic interaction between the solute and solvent charges have been estimated using equation([partition ] ) and are plotted in fig .
8 as a function of the fraction of solute charge @xmath85 being turned on .
several comments are worth making about the results in fig .
first , the partition functions @xmath133 for the umbrella potential - biased ensemble from which the free energies were computed correspond to a particular value of the electrostatic fraction parameter @xmath85 .
it should also be noted that although a difference in free energy between adjacent values of the electrostatic index of the order of @xmath116 seems to suggest a low probability of rejecting swaps between consecutive values of the index , the actual rejection rate was found to be significantly higher .
the relatively large rejection rate is due to the fact that differences in free energy reflect the _ average _ energetic _ and _ entropic differences between thermodynamic states with different @xmath134 , whereas the acceptance probability in equation ( [ acceptancecriterion ] ) involves only the difference in the energies of the actual configurations to be swapped . in particular , it was observed that the enthalpic @xmath135 and entropic @xmath136 contributions to the gibbs free energy @xmath132 vary in opposite directions as the fraction of the solute charge @xmath137 increases from zero to one .
these variations of enthalpic and entropic terms with the charge fraction can be visualized by comparing the results plotted in fig . 8 and fig .
9 . from the two figures it appears that @xmath135 increases and @xmath136 decreases with @xmath137 .
the opposite directions in which enthalpic and entropic terms vary with @xmath137 is reflected in the mobility of the state @xmath138 of the simulated tempering algorithm in the @xmath27 ( or @xmath137 ) subspace . turning our attention to equation ( [ acceptancecriterion ] ) , note that an increase of @xmath136 with @xmath137 suggests that for a large fraction of configurations ( i,*r * ) transitions in which @xmath27 is _ increased _
are accepted only if the ratio @xmath139 is approximately unity , @xmath140 .
such is the case for an important number of transitions from electrostatic index @xmath141 to electrostatic index @xmath142 , for example , for which an increase in @xmath137 is accompanied only by a small decrease and occasionally even an increase in energy . on the other hand , for a @xmath137 near unity
, a decrease of @xmath135 with @xmath137 suggests that a large fraction of transitions in which @xmath27 is _ decreased _ are rejected unless @xmath143 .
such is the case in a large number of transitions which are attempted from the final index @xmath144 to @xmath145 for example , which are accompanied by a significant increase in energy .
the choice of the weights @xmath146 suggested in equation ( [ choicew ] ) represents a good compromise in the sense that transitions which increase and which decrease @xmath27 have an equal probability of being accepted on average .
nevertheless , this analysis points to the fact that if one uses the simulated tempering approach for the study of chemical reactions in solution , one should try to avoid turning on the charges of the reactive core all the way from zero to their final values . as enthalpic and entropic terms
will always vary in opposite directions during the charging process , the sampling could become quite inefficient , especially when studying reactions in which there is a difference in the net charge between reactants and transition states , and not just in their dipole moments as in the present case .
however , this conclusion does not mean that the mmbif approach used in the simulation with linearly interpolated charges will always be more efficient than the simulated tempering method .
the efficiency of the mmbif approach relies heavily on the appropriateness of the postulated variation of the charges in a reactive system with the reaction coordinate . in the present study , the approximation of the variation of the fitted kollman - singh charges with the solute configuration by a linear variation with the parameter @xmath50 in equation ( [ a_b_1andb_2definitions ] ) proved quite accurate .
however , such a simple relation could break down , especially when studying reactions with a net transfer of charge . in this case
, we suggest using an approach which combines the benefits of both methods illustrated in the present study : the use of a simulated tempering approach in which the solute charge is gradually modified from a linear variation with the reaction coordinate to their actual values obtained via the kollman - singh approach .
in this article , two important issues on the applicability of the molecular mechanics based importance sampling method to the study of reactive events in condensed phase environments have been addressed .
one major concern in developing a successful implementation of the mmbif approach is the ease of development of a sufficiently accurate molecular mechanics potential to guide the sampling . although a fully automated approach to generating guiding potentials for general reactions is still not available , it is encouraging to note that the use of the same principles of bond - evolution theory@xcite as in our earlier study of the malonaldehyde system@xcite were adequate for designing a molecular mechanics potential for the acetic acid - methanol system .
this success is particularly impressive in light of the fact that the malonaldehyde and acetic acid - methanol systems differ substantially not only in their barrier heights ( by a factor of @xmath0 ) , but also in the qualitative nature of the chemical event ( one proton versus two protons transferred ) .
it suggests that bond - evolution theory guidelines are likely to be practical in developing molecular mechanics potentials for other proton transfer reactions , and for possibly other types of chemical events .
the study of the double proton transfer reaction between acetic acid and methanol in tetrahydrofuran also demonstrates that the mmbif method can be applied in reaction rate calculations of chemical transformations in solvents of medium polarity .
in particular , an increase in the cpu time of factors of @xmath0 and @xmath147 with respect to gas - phase calculations were obtained using two different sampling methods .
hence , we conclude that it should be possible in many instances to compute solvent mediated reaction rates with statistical accuracies comparable to those obtained in gas - phase calculations , even though the complexity of the calculation is increased enormously by the presence of the solvent . it should be emphasized that the contribution of the solvent to the total activation energy in the present study is only on the order of a few factors of @xmath116 at room temperature . as a result ,
the actual mechanism of the proton transfer event is virtually unchanged from the process in gas phase .
this simplification allowed the separation of the reactive _ ab initio _ core and the solvent degrees of freedom into effectively disjoint sets which were updated in isolation in the parallel tempering implementation of the sampling .
in many cases , the structure of the solvent plays a more substantial role in the reactive process . under such circumstances
, care must be taken to devise methods in which the reactive core and the solvent structure are updated in a more correlated fashion .
for example , for a reaction in a solvent with a larger dielectric constant , the simulated tempering approach can be implemented by incorporating simultaneous mmbif updates of the solute and solvent degrees of freedom at each electrostatic index .
it is informative to compare and contrast the advantages and disadvantages of using the mmbif method versus using present day molecular dynamics based methods for calculating reaction rates in solution using a hybrid qm / mm approach .
the practical use of the simple molecular dynamics methods for sampling configurations of a system containing a reactive core which is described using _ ab initio _ electronic structure methods and the solvent molecules which are described using a molecular mechanics potential is limited by the fact that calculating the time - evolution of the system necessitates the computation of the time - consuming _ ab initio _ forces acting on the core atoms .
these time - consuming calculations must be performed in traditional molecular dynamics calculations even when only the solvent degrees of freedom change significantly .
several methods have been proposed for circumventing the inefficiency of the traditional qm / mm molecular dynamics calculations which are based on an `` artificial '' separation of time scales associated with the solute and solvent atoms@xcite .
the central idea of these methods is to use a large mass in conjunction with a high temperature thermostat for the solute atoms , whereas the solvent atoms have usual masses and are in contact with a room temperature thermostat@xcite .
the large mass of the core atoms is chosen in such a way that the solvent degrees of freedom relax on a time scale which is much smaller than the time scale of the massive core atoms .
multiple - time scale arguments can be utilized to demonstrate that the integration of hamilton s equations describing the evolution of the solute plus solvent system can be performed by computing the forces acting on the solvent degrees of freedom significantly more often than the _ ab initio _ forces acting on the solute without altering the asymptotic boltzmann distribution of the configurations of the system .
in addition , the decoupling between the solute and solvent time scales enables the use of a large temperature thermostat coupled to the solute atoms without introducing an irreversible heat flow , thereby ensuring that the average kinetic energy of the core atoms is comparable with the magnitude of the reaction barrier which separates reactant and product configurations on the potential energy surface .
therefore , a relatively small number of _ ab initio _
calculations must be performed before a reactive event occurs .
both the mmbif and the modified molecular dynamics approach succinctly described above have a number of advantages and shortcomings when studying chemical reactions in solution which are essentially driven by fluctuations in the structure of the _ solute _ using a qm / mm approach . the main disadvantage of the mmbif method consists of the fact that a reasonable molecular mechanics description of the reactive event in the _ gas - phase _ solute is required , and this molecular mechanics potential must usually be created from scratch . nevertheless , once such molecular mechanics potential has been constructed , the cpu time necessary for calculating reaction rates in solution is essentially independent of the characteristics of the solvent such as its dielectric relaxation time .
in contrast , the molecular dynamics approach does not necessitate prior information with respect to the potential energy surface of the cluster , although if such information exists , it can be used to improve the efficiency of the sampling@xcite .
however , the average time needed to observe a reaction event increases as the square root of the effective mass of the reactive degree of freedom . on the other hand , given the requirement of the separation of time scales between the solvent and solute degrees of freedom , the characteristic time scale of the solute atom motion and therefore the lower bound for the value of the mass needed for the core atoms is _ determined _ by the duration of the solvent dielectric relaxation time .
therefore , it appears that the efficiency of the above mentioned molecular dynamics scheme decreases with the increase in the solvent dielectric relaxation time .
the methodology proposed in the present work can be combined with the ideas presented in reference [ to include nuclear quantum effects via centroid transition state theory with a supplementary increase in the cpu time by a factor of @xmath148 times the number of path - integral beads .
such a combination of procedures provides a rigorous and practical platform for calculation of kinetic isotope effects . an important goal for future
work is to clarify the mechanistic origin of the relation between the breakdown of the rule of geometric mean in multiple proton transfer reactions and tunneling effects .
10 g. r. fleming and p. g. wolynes , physics today * 43 * , 36 ( 1990 ) .
d. zhong and a. h. zewail , j. phys .
chem . a , * 102 * , 4031 ( 1998 ) .
b. perman , v. srajer , z. ren , t. teng , c. pradervand , t. ursby , d. bourgeois , f. schotte , m. wulff , r. kort , k. hellingwerf and k. moffat , science * 279 * , 1946 ( 1998 ) .
p. j. rossky and j. d. simon , nature * 370 * , 263 ( 1994 ) .
d. c. clary , science * 279 * , 1879 ( 1998 ) .
a. kohen , r. cannio , s. bartolucci and j. p. klinman , nature * 399 * , 496 ( 1999 ) ; b. j. bahnson , t. d. colby , j. k. chin , b. m. goldstein and j. p. klinman , proc .
94 * , 12797 ( 1997 ) ; y. cha , c. j. murray and j. p. klinman , science * 243 * , 1325 ( 1989 ) . m. page and a. williams , _ organic and bio - organic mechanisms _ , ( addison wesley longman , harlow , england , 1997 ) . g. m. torrie and j. p. valleau , chem .
* 28 * , 578 ( 1974 ) .
e. a. carter , g. ciccotti , j. t. hynes and r. kapral , chem .
phys . lett . *
156 * , 472 ( 1989 ) .
s. melchionna , phys .
e * 62 * , 8762 ( 2000 ) .
z. zhu , m. e. tuckerman and g. j. martyna , phys .
rev . lett . * 88 * , 100201 ( 2002 ) .
r. iftimie , d. salahub , d. wei and j. schofield , j. chem . phys .
* 113 * , 4852 ( 2000 ) .
j. van de vondele and u. rothlisberger , j. chem .
* 113 * , 4863 ( 2000 ) .
r. iftimie and j. schofield , j. chem . phys . *
114 * , 6763 ( 2001 ) .
r. iftimie and j. schofield , j. chem .
phys . * 115 * , 5891 ( 2001 ) .
r. iftimie and j. schofield , int .
. chem . * 91 * , 404 ( 2003 ) .
n. nandi , s. roy and b. bagchi , j. chem . phys . * 102 * , 1390 ( 1995 ) .
a. chaudhari , p. khirade , r. singh , s. n. helambe , n. k. narain and s. c. mehrotra , j. mol .
liq * 82 * , 245 ( 1999 ) .
a. kohen and j. h. jensen , j. am .
* 124 * , 3858 ( 2002 ) .
h. h. limbach , in _ electron and proton transfer in chemistry and biology _ , ( a. m@xmath149ler , h. ratajczak , w. junge and e. diemann , eds . ,
elsevier , new york , 1992 ) .
m. j. knapp , k. rickert and j. klinman , j. am .
soc . * 124 * , 3865 ( 2002 ) . for a simplified treatment of the problem of the rule of geometric mean
see also reference [ .
h. dugas , _ bioorganic chemistry , a chemical approach to enzyme action _ , ( @xmath150 edition , springer - verlag , new york , 1996 ) .
one of the textbook examples of bifunctional catalysis in solution chemistry is c. g. swain and j. f. brown , j. am .
* 74 * , 2534 ( 1952 ) .
d. gerritzen and h. h. limbach , j. am .
chem . soc . * 106 * , 869 ( 1984 ) .
m. neumann , mol .
phys . * 50 * , 841 ( 1983 ) .
r. o. watts , mol .
* 28 * , 1069 ( 1974 ) . v. barone and c. adamo , j. chem .
phys . * 105 * , 11007 , ( 1996 ) .
e. i. proynov , a. vela and d. r. salahub , chem .
230 * , 419 , ( 1994 ) ; * 234 * , 462(e ) , ( 1995 ) .
s. sirois , e. i. proynov , d. t. nguyen and d. r. salahub , j. chem .
phys . * 107 * , 6770 , ( 1997 ) .
a. fernndez - ramos , z. smedarchina and j. rodrguez - otero , j. chem .
phys * 114 * , 1567 ( 2001 ) .
e. i. proynov , s. sirois and d. r. salahub , int .
j. quantum chem . * 64 * , 427 , ( 1997 ) . for details on the demon quantum chemistry package , see http://cvs.demon-software.com/public_html/ w. l. jorgensen , d. s. maxwell and j. tirado - rives , j. am .
soc . * 118 * , 11225 ( 1996 ) .
r. iftimie , thesis , university of toronto , ( 2001 ) .
s. w. rick , s. j. stuart and b. j. berne , j. chem .
* 101 * , 6141 , ( 1994 ) .
gonzlez lebrero , d.e .
bikiel , m. dolores elola , d.a .
estrin and a.e .
roitberg , j. chem . phys . * 117 * , 2718 ( 2002 ) .
p. hohenberg and w. kohn , phys . rev . *
136 * , b864 ( 1964 ) .
w. kohn and l. j. sham , phys . rev . * 140 * , a1133 ( 1965 ) .
m. levy , proc .
usa * 76 * , 6062 ( 1979 ) .
a. laio , j. vandevondele , and u. rothlisberger , j. chem . phys . * 116 * , 6941 ( 2002 ) . c. reichardt , _ solvents and solvent effects in organic chemistry _ , ( vch , new york , 1988 , appendix a , table a-1 ) .
x. krokidis , v. goncalves , a. savin , b. silvi , j. phys .
a , * 102 * , 5065 ( 1998 ) ; see in particular the discussion in page 5070 and figure 8 . ; x. krokidis , s. nouri , b. silvi , j. phys . chem . a , * 101 * , 7277 ( 1997 ) .
e. marinari and g. parisi , europhys . lett .
* 19 * , 451 ( 1992 ) .
c. j. geyer and e. a. thompson , j. am .
assoc . * 90 * , 909 ( 1995 ) .
u. c. singh and p. a. kollman , j. comp .
chem . * 5 * , 129 ( 1984 ) .
l. russo , p. minry , z. zhu and m. e. tuckerman , j. chem .
phys . * 116 * , 4389 ( 2002 ) . j. v. vondele and u. rothlisberger , j. phys . chem .
b * 106 * , 203 ( 2001 ) .
p. w. atkins , _ molecular quantum mechanics _ , ( chapter 8 , oxford university press , new york , 1987 ) .
g. tabacchi , c. j. mundy , j. h@xmath149tter and m. parrinello , j. chem . phys .
* 117 * , 1416 ( 2002 ) . c. j. f. b@xmath151ttcher , theory of electric polarization , vol .
i , ( @xmath152 edition , elsevier , new york , 1973 ) , especially chapter 4 . figure 1 : the formation of two cyclic clusters involving one or two acetic acid molecules and one molecule of methanol .
the double - proton transfer reaction studied here takes place along the hydrogen bonds of the cyclic cluster formed from one molecule of acetic acid and one molecule of methanol .
figure 2 : a pictorial view of the qm / mm / continuum solvent method and of the reaction field approach .
the `` reactive core '' region contains a quantum representation ( i.e : nuclei + electrons ) of the atoms which are involved in the actual covalent bond - breaking and bond - forming events .
the `` explicit solvent '' region contains an atomic representation of the first few shells of solvent molecules .
the effects of the solvent molecules which are far from the reaction center are included in an implicit manner in the qm / mm / continuum approach using the reaction field method . in our implementation , the reaction field method consists of calculating an effective electrostatic interaction between the dipole moments of the molecules inside the first two regions ( see equation [ [ reactionfieldenergy ] ] ) .
figure 4 : a schematic representation of the simulated tempering method which uses the mmbif approach to generate configurations of the system in a simulation where solute polarizability is neglected .
the values of the electrostatic scaling parameter @xmath85 for electrostatic indices @xmath153 are gradually increased from zero to one .
figure 5 : the calculated potentials of mean force for the double - proton transfer reaction in the acetic acid - methanol cluster in gas - phase and in a solution of tetrahydrofuran using the reaction coordinate @xmath154 defined in equation ( [ reactioncoordinatemalonaldehyde2 ] ) .
note that the difference between the activation energies is approximately @xmath155 kcal / mol .
this difference is larger than the width of the @xmath156 confidence intervals for the activation energies , which have been estimated to be @xmath157 kcal / mol for the gas - phase and @xmath158 kcal / mol for the liquid - phase simulations ( inset ) .
figure 6 : the values of the electrostatic interaction energy between the solute and solvent molecules obtained for @xmath159 in the simulated tempering method in which the solute charges are the gas - phase kollman - singh charges .
note that the values of the solute - solvent electrostatic interaction energies are scattered over an energy range of approximately @xmath160 kcal / mol .
in contrast , the differences in the electrostatic interaction energies calculated using the kollman - singh charges and the same energies calculated using solute charges which vary linearly with the parameter @xmath161 defined in equation ( [ a_b_1andb_2definitions ] ) are scattered over an interval of only @xmath162 kcal / mol ( data not shown ) .
figure 7 : the values of the polarization energy difference @xmath163 defined in equation ( [ deltaepol ] ) .
note that the distribution of values is centered roughly at @xmath164 kcal / mol and has a small standard deviation of @xmath158 kcal / mol .
figure 8 : the free electrostatic interaction energy between an acetic acid - methanol complex and molecules of tetrahydrofuran solvent as a function of the charge state of the complex . @xmath165 and @xmath159 correspond to fully uncharged and charged solute .
the stars and diamonds represent the estimated values of the free energy obtained from the preliminary and from the optimized simulated tempering runs , respectively .
the continous line represents a spline interpolation between the computed values .
figure 9 : the normalized probability density of the solvent - solute electrostatic interaction energy on for electrostatic indices @xmath142 ( solid curve ) , @xmath166 ( long dashed curve ) and @xmath167 ( dashed curve ) . note that the enthalpies , calculated as the average solute - solvent electrostatic interaction energies , are approximately @xmath168 , @xmath169 and @xmath170 kcal / mol . comparing these results with the corresponding free energies in fig .
8 one obtains the corresponding entropies as being @xmath171 , @xmath172 and @xmath173 kcal / mol .
the variations in the mechanical work @xmath174 with the number of chain are negligible .
.[oplsaaexperimentalmodoplsaa]the values of the gas - phase dipole moment @xmath175 expressed in debyes , and of the static dielectric constant @xmath176 , obtained by considering the opls - aa charges as permanent charges , obtained from our modified version of the opls - aa force field ( see text ) which accounts approximately for electronic polarization effects , and from experimental data from reference [ . [ cols="^,^,^,^",options="header " , ]
|
in this article , we propose an efficient method for sampling the relevant state space in condensed phase reactions . in the present method ,
the reaction is described by solving the electronic schrdinger equation for the solute atoms in the presence of explicit solvent molecules .
the sampling algorithm uses a molecular mechanics guiding potential in combination with simulated tempering ideas and allows thorough exploration of the solvent state space in the context of an ab initio calculation even when the dielectric relaxation time of the solvent is long .
the method is applied to the study of the double proton transfer reaction that takes place between a molecule of acetic acid and a molecule of methanol in tetrahydrofuran .
it is demonstrated that calculations of rates of chemical transformations occurring in solvents of medium polarity can be performed with an increase in the cpu time of factors ranging from 4 to 15 with respect to gas - phase calculations .
* h * * s * * u * *p * * h * * s * * u * *p *
|
consider a classical object ( a degree of freedom ) trapped in a metastable potential minimum ; no decay out of this metastable state is possible at low temperatures , where thermal activation over the barrier is exponentially suppressed .
however , if the classical object is a composite one , with a quantum degree of freedom coupled to the classical one , then the quantum object may tunnel out of the metastable minimum and exert a pulling force on the classical object .
once the latter is large enough to completely suppress the trapping barrier , the classical object is able to leave the potential well hence a classical object may escape from a metastable state even at zero temperature if helped by a coupled quantum degree of freedom .
the above situation can be realized in a dc - squid ( superconducting quantum interference device ) featuring asymmetric dynamical parameters ; i.e. , with two josephson junctions of equal critical currents @xmath0 but strongly different ( shunt ) capacitances @xmath1 and ( shunt ) resistances @xmath2 , see fig .
[ fig : setup ] ; choosing large and small parameters @xmath1 and @xmath3 for the two junctions allows to place one of the junctions in the ` classical ' and the other into the quantum domain .
the tunneling of the quantum degree of freedom entails a distortion of the trapping potential of the classical junction , which might be sufficiently large to transform the metastable state of the classical junction into an unstable one .
the appearance of this complex decay channel depends critically on the applied bias current @xmath4 and the squid s loop inductance @xmath5 coupling the two junctions . )
biased superconducting loop with inductance @xmath5 .
the two junctions feature equal critical currents @xmath0 but strongly asymmetric ( shunt ) capacitances @xmath6 and ( shunt ) resistances @xmath7 .
we assume that @xmath6 and @xmath8 are chosen such that quantum effects are present for junction 2 but are strongly suppressed for junction 1 , hence @xmath9 and/or @xmath10 . ]
the gauge invariant phase differences @xmath11 , @xmath12 , across the two josephson junctions@xcite define our dynamical degrees of freedom : assuming equal critical currents @xmath13 , the junctions potential energies @xmath14 $ ] , @xmath12 , involve the josephson energy @xmath15 ( with @xmath16 the flux unit , @xmath17 and @xmath18 denote the unit charge and light velocity).@xcite their kinetic energies @xmath19 are determined by the junction capacitances @xmath6 playing the role of effective masses ( the relevant energy scale is given by the charging energy @xmath20)a dynamically asymmetric squid with one large and one small junction capacitance then provides us with the desired classical and quantum degrees of freedom ( we choose @xmath9 ; additional normal resistances @xmath8 introduce a dissipative dynamical component , see below ) .
the coupling of the two junctions via the loop inductance @xmath5 produces the interaction energy @xmath21 ( \varphi_2-\varphi_1)^2 $ ] involving the relative coordinate @xmath22 , whereas the external driving current @xmath4 couples to the absolute coordinate , @xmath23 . while large - capacitance ( ` classical ' ) junctions are easily fabricated , small ( ` quantum ' ) junctions are more difficult to realize . nevertheless , experimental techniques to fabricate small junctions are available today and their quantum behavior in the form of quantum tunneling , @xcite quantized energy levels @xcite and even quantum coherence @xcite has been demonstrated .
the quantum decay of the biased , dynamically symmetric dc - squid has been discussed before both experimentally@xcite and theoretically,@xcite also in the context of instanton splitting.@xcite in the squid discussed here , @xcite the large asymmetry of the dynamical parameters blocks the tunneling of the @xmath24 degree of freedom ; the ` heavy ' junction then remains frozen during the quantum decay of the ` light ' degree of freedom @xmath25 .
a trajectory of this kind describes the entry of flux into the squid loop which rearranges the current flow in the two arms in a way as to redirect more current through the heavy junction .
this increase in current produces an enhanced tilt @xmath26 in the potential of the heavy junction , which then may decay through a classical trajectory , see fig .
[ fig : corrpot ] .
we call this nontrivial decay sequence the ` mnchhausen decay '
. it is the aim of this work to determine the effective critical current @xmath27 for which the mnchhausen decay becomes possible , see figs . [
fig : phase_diagram_vv ] and [ fig : phase_diagram_adiabatic]-[fig : pe ] . in the following ,
we define our system in full detail , including a dissipative component in the junction dynamics ( sec .
[ sec : setup ] ) .
sections[sec : overdamped ] and [ sec : underdamped ] are devoted to the derivation of the effective critical currents for the various cases with junctions governed by massive or dissipative dynamics . in section [ sec : remarks ] we add remarks concerning the experimental realization of the system described here .
finally , we draw conclusions in section [ sec : conclusion ] .
within the resistively and capacitively shunted junction ( rcsj ) model ( at @xmath28 ) , the classical dynamics of the two phase differences @xmath24 and @xmath25 is governed by the equations of motion @xmath29 with the plasma frequency @xmath30 of an unbiased single junction and the damping coefficients @xmath31 ( @xmath8 denote the normal ohmic junction resistances ) . the potential ( see fig .
[ fig : corrpot ] ) is given by @xmath32 with the dimensionless current @xmath33 and the coupling constant @xmath34 ( @xmath35 denotes the usual screening parameter of the squid ) .
eqns . and
describe a dc - squid with symmetric inductance @xmath5 in a vanishing external magnetic field and driven by a bias current @xmath4 or , equivalently , the massive ( mass @xmath36 ) and/or dissipative ( @xmath37 ) dynamics of two harmonically ( @xmath38 ) coupled particles in a tilted ( @xmath39 ) and corrugated ( @xmath40 ) potential .
quantum effects of the light junction 2 are accounted for via the relevant tunneling and decay processes , see below . and coupling @xmath41 .
for an undercritical current @xmath42 , the symmetric minima @xmath43 are all metastable .
the stability of the side - minima @xmath44 depends crucially on the parameters @xmath39 and @xmath38 .
the line in ( b ) shows the decay path in an overdamped setup : the system starts out in a relaxed local ground state near the initial minimum at @xmath45 .
the ` light ' degree of freedom @xmath46 then tunnels ( dotted line ) and the system relaxes to the bottom of the next minimum near @xmath47 , with a classically stable but quantum mechanically metastable ground state . through an additional quantum phase slip , the light phase reaches the minimum near @xmath48 , which is _ not _ classically stable and henceforth the system can decay along a classically allowed path to @xmath49 .
the system then has turned unstable and enters a resistive state through iteration of the last two steps.,title="fig : " ] and coupling @xmath41 . for an undercritical current @xmath42 ,
the symmetric minima @xmath43 are all metastable .
the stability of the side - minima @xmath44 depends crucially on the parameters @xmath39 and @xmath38 .
the line in ( b ) shows the decay path in an overdamped setup : the system starts out in a relaxed local ground state near the initial minimum at @xmath45 .
the ` light ' degree of freedom @xmath46 then tunnels ( dotted line ) and the system relaxes to the bottom of the next minimum near @xmath47 , with a classically stable but quantum mechanically metastable ground state . through an additional quantum phase
slip , the light phase reaches the minimum near @xmath48 , which is _ not _ classically stable and henceforth the system can decay along a classically allowed path to @xmath49 .
the system then has turned unstable and enters a resistive state through iteration of the last two steps.,title="fig : " ] the potential gives rise to two types of relevant frequencies .
one is the plasma frequency @xmath50 , the small - amplitude frequency in the direction of @xmath11 around a local minimum of the potential @xmath51 . with the effective potential @xmath52(\varphi_i ) = v(\varphi_l=\mathrm{const.},\varphi_i),\ i\neq l , \label{eq : effpot}\ ] ] cf .
[ fig : effpot ] , @xmath53 , evaluated at a local minimum @xmath54 , and depends on the parameters @xmath39 and @xmath38 as well as on @xmath54 . for the heavy junction (
junction 1 ) , @xmath55 can become arbitrarily small upon approaching criticality , while for the quantum junction ( junction 2 ) @xmath56 becomes small only for @xmath57 and @xmath58 .
the other frequency is given by the @xmath59 constant of the ` superwell ' in @xmath60 , cf .
fig.[fig : effpot ] , and is relevant only in the regime @xmath61 and for the quantum junction ( junction 2 ) , @xmath62 .
( solid line ) and the parabola remaining after dropping @xmath63 ( dashed line ) for @xmath64 , @xmath65 and @xmath66 .
the bullets and arrows illustrate the sequential decay of the quantum phase @xmath25 to the ground state for the case of strong damping . ]
these characteristic frequencies delineate the regimes where the two junctions behave classically and quantum mechanically , respectively .
the relevant physical parameters are given by the ratios @xmath67 .
for the heavy junction , we have to make sure that the number @xmath68 of states in the local well remains large . choosing a large ratio @xmath69 then guarantees that quantum effects ( i.e. , @xmath70 of order unity ) become relevant only very close to criticality and we choose to ignore them in our discussion below .
the system develops the most interesting behavior if the quantum degree of freedom develops well localized states within all relevant local minima in @xmath71 , fig .
[ fig : effpot ] , and hence parameters should be chosen such that the quasi - classical description applies .
on the other hand , we require tunneling and coherence effects to manifest themselves on reasonably short ( measurable ) timescales . we then choose a ratio @xmath72 , such that the local wells in @xmath73 contain a few quasi - classical states each .
the strength of dissipation can be quantified by the dimensionless damping parameters , @xmath74 below , we are interested in the two limiting cases of strong and weak damping . for a strongly damped quantum junction with @xmath75 , the quantum decay of @xmath25 out of a metastable
well of @xmath76 is incoherent @xcite and its subsequent relaxation is fast ( as compared to the dynamics of @xmath24 ) . for weak
damping , @xmath77 , @xmath78 , the kinetic energy stored in the motion of the heavy junction has to be accounted for ; in addition , the finite lifetime of the quantum states of the light junction due to the residual dissipation has to be considered , see the discussion in sec .
[ sec : underdamped ] .
we start by analyzing the situation for strong damping , @xmath75 and @xmath79 , see ref . for an earlier short report .
we bear in mind an experiment with a dc - squid characterized by a fixed inductance @xmath80 and biased with a current @xmath81 .
the task is to determine whether the mnchhausen decay can take place ; in the experiment , the latter manifests itself through the transition to a finite voltage state . in the strong damping case ,
no kinetic energy is stored in the system .
furthermore , the evolution is not sensitive to the way the current is ramped . after current ramping ,
the system starts out in a relaxed state where the phases @xmath11 are localized in the diagonal metastable minimum at @xmath45 ( up to an arbitrary multiple of @xmath82 ) .
for sufficiently large @xmath39 , the quantum degree of freedom @xmath25 undergoes tunneling to a new local minimum nearby @xmath83 , while the classical degree of freedom @xmath84 remains localized , thus allowing a flux @xmath85 to enter the squid loop , cf .
[ fig : effpot ] .
if the resulting force on the classical phase @xmath24 is sufficiently large , the mnchhausen decay is enabled with a classical decay of @xmath24 and successive iteration of quantum decay ( directed along @xmath25 , flux entry ) and classical relaxation ( directed mainly along @xmath24 , flux exit ) , cf .
[ fig : corrpot ] . from the segments
@xmath86 , eq . , and @xmath87 , eq .. in ( c ) , the relevant region of the @xmath39-@xmath38-plane is shown . at a
, the quantum junction ( see ( b ) ) has relaxed to the initial minimum ( @xmath88 ) and the classical junction is localized in a stable minimum ( see ( a ) ) . increasing the bias current @xmath39 and approaching b , the potential barrier confining the classical phase in @xmath89 is only slightly reduced . as @xmath90 ( c ) , the increase in @xmath39 has deformed the effective potential @xmath76 as to allow a phase slip of @xmath25 ( see ( b ) ) .
the additional force it exerts on the classical junction immediately removes the barrier ( see ( a ) ) and delocalizes @xmath91 at @xmath92 . on the other hand ,
an increase in bias @xmath90 from a@xmath93 to b@xmath93 at lower @xmath38 leads to a phase slip of @xmath25 ( see ( e ) ) , without delocalizing @xmath24 ( see ( d ) ) . a further increase in @xmath39 triggers a classical decay ( see ( d ) ) when crossing the critical line at @xmath94 ( c@xmath93 ) . ]
the strong coupling at large @xmath38 keeps the local minimum at @xmath95 lower in energy than the adjacent local well at @xmath96 for all bias currents @xmath39 , hence tunneling is inhibited and no mnchhausen decay takes place . at lower @xmath38 , a bias current @xmath42 can sufficiently lower the adjacent well such as to bring both minima to equal height .
this condition is reached once the minimum of the parabola in @xmath76 ( at @xmath97 ) is aligned with the midpoint between the two corresponding minima of @xmath98 ( at @xmath99 ) ; i.e. , if @xmath100 .
thus , for @xmath101 with@xcite @xmath102 the minimum at @xmath47 is lower and a quantum decay is enabled ( cf .
[ fig : explanation](b ) , point c in the diagram ) .
the jump of @xmath25 by roughly @xmath82 then pulls the heavy junction out of its minimum and the mnchhausen decay is initiated , cf
. figs .
[ fig : explanation](a ) and ( b ) associated with the sequence of points a , b , and c in fig.[fig : explanation](c ) .
the smaller @xmath38 becomes , i.e. , the weaker @xmath25 is bound to @xmath24 , the less current is needed to lower the adjacent minimum such as to allow for a decay of @xmath25 , hence the positive slope of @xmath103 .
decreasing @xmath38 too far , however , the pulling force exerted by the quantum junction may not be sufficient to drag the heavy junction out of its minimum .
hence , we have to investigate the shape of the potential @xmath104 after the phase slip in @xmath25 , i.e. , nearby the point @xmath105 , and check whether the barrier against a classical decay ( mainly along @xmath24 ) has disappeared ; this is identical to the calculation of the critical current of a squid with a trapped flux.@xcite we first determine the position of the minimum by solving the equations @xmath106 for @xmath107 and @xmath96 . at the critical coupling @xmath108
the minimum should merge with a saddle and define an inflection point along some direction in the @xmath109-plane .
the resulting system of equations requires numerical solution and the result is shown in the inset of fig .
[ fig : phase_diagram_vv ] .
however , within the interesting region at small coupling @xmath38 we can find an approximate analytical solution : for @xmath61 , the side minimum of @xmath51 becomes unstable predominantly along the @xmath24-direction .
thus , the minimum disappears if @xmath110 we choose the solution @xmath111 ( we set @xmath112 and assume @xmath113 ; the other solution @xmath114 can not solve eq . and is discarded ) .
inserting @xmath24 into the sum of eqs . and yields the relation @xmath115 from which we find @xmath116 $ ] ( the other solution @xmath117 $ ] is excluded since it describes a maximum along the @xmath25-direction ) . inserting @xmath24 and @xmath25 into eq . , we find the condition @xmath118 for @xmath119 no barrier blocks the motion of @xmath24 after the phase slip in @xmath25 and a classical decay of @xmath24 is enabled , thus completing the mnchhausen decay .
this scenario is described by the sequence a , b , c in fig .
[ fig : explanation](c ) . for @xmath120
the force after the phase slips is too small and an additional increase in @xmath39 is necessary to drive the system overcritical , as illustrated by the sequence a@xmath93 , b@xmath93 , c@xmath93 in fig .
[ fig : explanation](c ) and figs .
[ fig : explanation](d ) and ( e ) .
the increase in critical current with decreasing coupling defines a negative slope for @xmath108 .
we define the effective critical current @xmath121 as the phase boundary between the stable region ( @xmath122 ) , where the mnchhausen decay is prohibited , and the delocalized phase ( @xmath123 ) .
this critical line is assembled from the segments @xmath124 , the inverse functions of @xmath125 , eqs . and , respectively : ramping up the current @xmath39 at large values of @xmath38 , we eventually cross @xmath126 .
the phase slip of @xmath25 immediately enables the classical decay of @xmath24 and the system enters a running state , hence , @xmath92 . at lower @xmath38 ,
increasing @xmath39 beyond @xmath86 triggers a phase slip of @xmath25 , but the resulting force is too weak to delocalize @xmath24 .
a further increase in @xmath39 is necessary until , at @xmath127 , the minimum disappears and the mnchhausen decay proceeds , hence , @xmath94 .
following the above discussion , the system is stable for @xmath128 . upon further decreasing @xmath38 , however ,
more and more side minima in @xmath129 become accessible to the quantum junction . from fig .
[ fig : effpot ] we notice , that the different local minima of @xmath129 are located near @xmath130 .
they describe the state where a flux @xmath131 has entered the loop and we label them by the index @xmath132 .
eq . is then straightforwardly generalized to the critical line @xmath133 describing the entry of the @xmath132-th fluxon .
the @xmath132-th phase slip of @xmath25 occurs when the global minimum in @xmath76 shifts from @xmath134 to @xmath130 .
we can proceed analogously to the derivation of eq .
: let @xmath135 be the solution of eqs . and near @xmath136 .
in order to find the crossing in the height of the two minima we account for the relaxation of @xmath24 and align the minimum of the shifted parabola ( @xmath137 ) with the midpoint between the two minima of the @xmath138-potential ( @xmath139 ) , hence @xmath140 and @xmath141 a convenient approximation is made by ignoring the relaxation of @xmath24 , resulting in the expression @xmath142 similarly , we can generalize eq . to a critical coupling @xmath143 , determining whether the resulting force at given fluxon index @xmath132 is sufficient to delocalize @xmath24 .
we then solve eqs . and near @xmath144
, proceed as in the derivation of eq .
, and find ( for small @xmath61 ) @xmath145 and inductive coupling @xmath146 . here
, we assume strong damping @xmath147 and @xmath75 .
the effective critical current @xmath121 ( solid line ) marks the boundary between a localized classical junction ( lower bias @xmath148 ) and a delocalized classical junction ( @xmath149 ) , corresponding to a finite voltage state of the squid .
branches with negative slope are determined by a classical instability ( mainly along @xmath150 ) , while those with positive slope are determined by a quantum instability of the light junction . for @xmath122 , the dotted lines @xmath151 mark
the entry of flux through the quantum junction ( the integer @xmath132 approximately quantifies the flux through the ring in the stable state ) ; these lines can be measured via monitoring of the flux threading the loop . for @xmath123 , the dashed lines @xmath152
mark the minimum number @xmath153 of flux units necessary to delocalize the classical junction .
the inset shows a comparison between the approximate result , eqs . and ( solid line ) and the exact numerical result ( dashed ) . ] the critical current line @xmath121
is constructed from interchanging segments of @xmath151 and @xmath154 , resulting in the dynamical phase diagram , fig .
[ fig : phase_diagram_vv ] .
note that the different nature of the decay , classical or quantum , associated with the two types of critical lines may allow for an experimental distinction : ramping the current past a @xmath155-type segment of @xmath121 triggers a _ quantum _ decay with a broad histogram describing multiple measurements .
a @xmath156-type segment of @xmath121 triggers a _ classical _ decay with a sharp histogram ( the quantum decay of @xmath25 needs to have occurred already , which limits the ramping speed before reaching the critical line ) .
note that the lines @xmath151 are detectable throughout all the stable portion of the phase diagram , e.g. , via a measurement of the flux threading the loop ( the flux increases by approximately one flux unit upon crossing the dotted lines in fig.[fig : phase_diagram_vv ] ) .
the phase diagram in fig .
[ fig : phase_diagram_vv ] shows that the critical line @xmath121 approaches the value @xmath157 for @xmath158 .
this is understood from analyzing eqs . and
for large @xmath132 and small @xmath38 , providing the relations @xmath159 and @xmath160 , respectively .
equating the two conditions gives @xmath161 . from a physical point of view , this result can be easily explained : the decay of @xmath25 proceeds towards the bottom of the parabola in @xmath76 ; for @xmath158 , the current through the quantum junction then approaches zero , @xmath162 .
consequently , all current is redirected through the classical junction 1 , with the effective bias now increased to @xmath163 .
junction 1 thus turns dissipative at @xmath164 , the critical current of a single junction but half the critical current of the dc - squid .
and @xmath25 at @xmath165 , cf .
fig.[fig : corrpot](b ) .
the flux increases by approximately one flux unit each time the quantum junction decays ; this process then involves the tunneling rate @xmath166 of @xmath25 , eq . .
the additional flux leaves the loop again within the time interval @xmath167 as the classical junction relaxes dissipatively to the next minimum . ] the dynamical phase diagram , fig .
[ fig : phase_diagram_vv ] , can be tested experimentally through a measurement of the dc - voltage drop across the device .
the second josephson relation@xcite tells that the time - averaged voltage @xmath168 . inspecting fig .
[ fig : corrpot](b ) , we understand that the continuous iteration of quantum and classical decays of the light and heavy phase variables generate the nonzero averages @xmath169 for @xmath123 . for the overdamped setup
as described above , the voltage drop is small as strong damping reduces the tunneling rate determining @xmath170 .
this is particularly true close to the critical line , where the variable @xmath25 needs to tunnel in the ` flat ' part of @xmath76 .
the decay process is faster for weak damping and thus generates a larger voltage signal we discuss this situation in the next section .
another characteristic signal of the mnchhausen decay is the time - trace of the magnetic flux threading the squid loop during the alternating decay of the two phases ( cf .
[ fig : corrpot](b ) ) , see fig .
[ fig : flux ] for an illustration .
such time - traces exhibit two characteristic time scales , one due to the tunneling of the quantum junction involving the rate @xmath166 separating subsequent peaks of flux entry into the loop ; the other is due to the classical relaxation of the heavy junction and involves the dissipative time @xmath171 describing the flux exit and return to the low - flux state of the loop . at small @xmath38 and small effective bias between two minima , @xmath172 with @xmath173.@xcite
the behavior of the system for weak damping @xmath174 is quite similar to the one encountered before , while the ( small ) differences strongly depend on the many system parameters and their associated timescales . to begin with we neglect dissipative effects .
bearing in mind the limit of large @xmath175 , we can exploit the adiabatic separation of fast and slow degrees of freedom ; i.e. , we first consider the fast problem for the quantum junction , @xmath176 with the reduced hamiltonian ( cf .
@xmath177 ( \varphi_{2})\ ] ] defined at _ fixed _ @xmath24 .
. establishes the dynamics of the ` light ' variable @xmath25 and we can find the energies @xmath178 .
these act as effective potentials for the ` heavy ' degree of freedom @xmath24 ; e.g. , assuming the quantum junction to reside in a state @xmath179 , @xmath180 , the effective potential for @xmath24 is given by @xmath181 .
typical examples for the effective potentials @xmath182 ( ground state ) and @xmath183 ( first excited state ) are shown in fig .
[ fig : p1 ] . in the dissipative situation , we could derive the phase diagram from simply analyzing the potential @xmath51 ; here , instead , we first determine which states @xmath184 of the light junction are relevant and then analyze the dynamics of @xmath24 in the resulting effective potential ( e.g. , fig.[fig : p1](a ) ) defined via @xmath178 , thus determining whether the heavy junction stays localized or enters a running state .
both tasks strongly depend on the various timescales involved .
we start with the timescale of the ` external control ' , the rate at which the bias @xmath39 is ramped .
whereas the case of strong damping excluded the presence of kinetic energy in the system , this is no longer the case here .
the importance of the ramping rate then comes about through the amount of energy which is transferred to the two degrees of freedom .
the change of @xmath39 translates to a change of the potential @xmath51 with a rate of the order of @xmath185 .
the ramping can be adiabatic , in which case no energy is given to the system , or instantaneous , where the amount of transferred energy is maximal .
intermediate types of ramping will not be discussed . , generated by the two lowest levels of the ` light ' phase . in ( a ) , the situation for @xmath186 is shown .
the avoided level crossing @xmath187 in the spectrum of the light junction is reflected in the effective potentials @xmath188 ( solid ) and @xmath183 ( dotted ) for the heavy junction .
the inset illustrates the 2-dimensional path @xmath189 $ ] in @xmath104 corresponding to @xmath188 ( solid , a / b ) and @xmath183 ( dotted , c / d ) ; the upward / downward jump along @xmath25 to the next minimum corresponds to the avoided crossing in @xmath190 .
( b ) series of @xmath188 ( solid ) and @xmath183 ( dotted ) for various values of @xmath39 at @xmath191 .
increasing @xmath39 shifts the position of the avoided crossing in @xmath192 to the left .
its transition past @xmath193 ( between @xmath194 and @xmath195 ) corresponds to crossing the line @xmath86 and implies the decay of the quantum junction to the next local minimum .
note that in the present case the newly formed minimum is stable ( a point of type b@xmath93 in fig .
[ fig : explanation ] ) . ]
the adiabatic separation of time scales requires a large capacitance ratio @xmath196 . for @xmath197 the intravalley motion of the two junctions
are separated .
the ( tunneling ) motion of the light junction across valleys produces a stronger condition : consider the tunneling splitting @xmath198 in the spectrum of @xmath199 at an avoided level crossing , cf . fig .
[ fig : p1](a ) , @xmath200 with @xmath201 depending on the actual shape of the potential , @xmath202 in the quasi - classical approximation . here , @xmath203 is the ground state energy in the well and @xmath204 and @xmath205 are the left and right classical turning points , respectively , @xmath206 .
the tunneling gap @xmath198 then is relevant when analyzing the motion of the classical junction ( @xmath24 ) in the effective potential given by @xmath178 . if the motion of @xmath24 is sufficiently slow , its passage along an avoided crossing in the spectrum of @xmath199 is adiabatic and @xmath24 follows @xmath178 through the anticrossing ( cf .
the trajectory a@xmath207b in fig . [ fig : p1](a ) ) ; i.e. , the phase @xmath25 tunnels to the next adjacent minimum . on the other hand ,
if the motion of @xmath24 is fast , the state of the quantum junction undergoes landau - zener tunneling ( cf .
the trajectory a@xmath207d in fig.[fig : p1](a ) ; the light junction then remains trapped in its local well ) and @xmath24 follows the effective potential @xmath208 after the avoided crossing .
the probability @xmath209 for landau - zener tunneling at an anticrossing of the two lowest levels of @xmath199 is given by @xmath210 where @xmath211 and @xmath212 is the total time derivative .
to estimate the rate of change in energy @xmath213 , we start from @xmath214\ ] ] and find the time derivative @xmath215\dot\varphi_1 \nonumber\\ & = & ke_j ( \varphi_2^{\mathrm{min},n}-\varphi_2^{\mathrm{min},n+1})\dot \varphi_1,\end{aligned}\ ] ] where we have used that @xmath216 vanishes at the minima @xmath217 and only the term originating from the coupling term in @xmath51 is relevant .
since @xmath218 is at most of order @xmath219 , we obtain the estimate @xmath220 .
the smallness of @xmath221 , guaranteeing adiabatic motion of the classical junction ( this requires a correspondingly large @xmath222 ) then follows from @xmath223 the relaxation of the quantum degree of freedom @xmath25 is another important element in our discussion . the dissipation described by the normal resistance @xmath2 in the rcsj - equation of motion , cf .
eq . , leads to typical finite lifetimes @xmath224 of the excited states @xmath225 of the quantum junction.@xcite a much longer lifetime @xmath226 shows up if the quantum junction is trapped in a local ground state of a side - well in @xmath76 , cf .
[ fig : decay ] .
the decay then is protected through a large barrier @xmath227 , enhancing the typical lifetime to a value @xmath228 , with @xmath229.@xcite thus , there are two ( extreme ) ways how a highly exited state in the ` superwell ' of @xmath76 can decay , see fig.[fig : decay ] , either via states within the superwell involving the typical lifetime @xmath230 , or via states involving a tunneling process , resulting in an exponentially larger decay time of the order of @xmath226 . , eq . .
solid arrows indicate a decay sequence involving only states within the ` superwell ' with a typical decay time @xmath230 .
the dashed arrows illustrate another ( extreme ) decay sequence , involving localized states in the side - minima of @xmath60 . the decay time then involves the typical lifetime @xmath231 of local ground states within side - wells . ] in the following , we investigate three regimes and derive the corresponding effective critical current @xmath121 .
we begin with the case of adiabatic ramping ( sec .
[ sec : adiabatic ] ) , where we assume that the ramping is slower than the typical relaxation time of the heavy junction and the timescales between @xmath24 and @xmath25 separate ; i.e. , inequality holds .
we show that the critical current @xmath121 obtained in the dissipative case , sec.[sec : overdamped ] , is only slightly altered due to the ( small ) finite amount of kinetic energy the classical junction can store .
next , we discuss the case of fast ramping ( sec .
[ sec : fast ] ) , where the phase @xmath24 remains effectively ` frozen ' during the current ramping and quite an appreciable amount of potential energy is converted to kinetic energy in the motion of the classical junction , leading to a pronounced reduction of @xmath121 .
both discussions require a very large value of @xmath232 to guarantee the absence of landau - zener tunneling due to the motion of @xmath24 .
subsequently , we discuss the consequences of a moderate ratio @xmath233 accessible in today s experiments in the last part of this section ( sec .
[ sec : farthest ] ) .
then , @xmath25 can remain trapped in a localized state in a side well during the motion and liberation of the heavy junction .
the probabilistic effects emerging for this situation change the nature of the phase diagram qualitatively .
we start from the system with fixed coupling @xmath38 at @xmath234 and consider a state localized at @xmath235 .
we assume slow current ramping on the dissipative timescale of the classical junction s motion , @xmath236 where @xmath237 denotes the classical relaxation time of the heavy junction .
the inequalities ( slow ramping ) and ( slow motion of @xmath24 ) guarantee that the quantum junction remains in the global ground state , while the classical junction follows its local ground state near @xmath238 .
the effective potential for the heavy junction then is determined by the ground state energy @xmath239 . in our estimate @xmath240
\label{eq : e_0}\ ] ] we neglect the correction due to the ground state energy @xmath241 ( as measured from the bottom of the potential ) of the light junction;@xcite the phase @xmath242 refers to the global minimum of the light junction .
furthermore , we ignore the small splitting @xmath187 at the avoided level crossing ( cf.fig . [
fig : p1](a ) ) , replacing it with a sharp kink . with the above approximations , the force acting on the heavy junction originates from the classical force exerted by the quantum junction residing in the global minimum .
the situation is thus the same as in the strong damping regime : the ` flux - entry lines ' @xmath151 , marking the current where the light junction is allowed to tunnel , are still given by eq .. when the classical junction has relaxed to the bottom of a minimum , that minimum has to disappear in order for @xmath24 to enter the running state .
given the expression for @xmath188 , the local minimum in @xmath188 corresponding to the @xmath132-th side minimum in @xmath243 turns into an inflection point at @xmath244 , eq .. this condition again agrees with the one for strong damping . ) .
the critical line @xmath121 is nearly identical with the one in the strong damping case , but is hysteretic along extensions of the critical lines @xmath151 .
these are due to the nonzero kinetic energy acquired by the classical junction when sliding down in the new well formed after the tunneling of @xmath25 at @xmath151 and allowing @xmath24 to surmount a residual potential barrier in the effective potential @xmath188 .
the inset illustrates the condition determining the termination point of the hysteretic line along @xmath86 .
the three solid dots mark points of equal potential energies ; the energy gained by the classical junction due to the phase slip of @xmath25 just suffices to climb the top of the remaining barrier along @xmath24 . ]
a difference to the previous dissipative situation is given by the possibility to transform potential into kinetic energy .
ramping @xmath39 at fixed @xmath38 past the flux - entry line @xmath151 triggers a phase slip in @xmath25 and deforms the potential well in @xmath188 for the classical junction .
the latter finds itself on the slope of the well and gains kinetic energy while sliding down towards the new minimum .
the classical junction then may enter a running state if this kinetic energy gain is sufficient to surpass the barrier blocking the newly formed ( @xmath132-th ) minimum .
different from the overdamped case , the system then exhibits hysteretic behavior , as illustrated in the dynamical phase diagram fig.[fig : phase_diagram_adiabatic ] , where the running state can be entered along extensions of the @xmath151 line away from the phase boundary .
the point where these extensions terminate is determined by the condition that the energy right after the tunneling of @xmath25 is equal to the energy at the top of the barrier , as illustrated in the inset of fig.[fig : phase_diagram_adiabatic ] . if the coupling @xmath38 is too small to lower the barrier sufficiently ( cf . fig .
[ fig : p1 ] ( b ) ) , the state of the heavy junction upon reaching @xmath245 is transformed into a localized excited state of the newly formed potential well in @xmath182 .
given the slow ramping @xmath246 , the heavy junction relaxes to the ground state of the well and a further increase in @xmath39 is necessary to remove the barrier completely , hence the critical line jumps forward to @xmath247 .
we now turn to the case @xmath248 , where the ramping of @xmath39 is instantaneous with respect to the ( massive ) dynamics of the heavy junction ; the final current @xmath39 has to remain below @xmath249 , since fast ramping beyond this value allows the classical junction to overcome the potential barriers through conversion of potential to kinetic energy .
furthermore , we study the setting where @xmath250 ; the quantum junction then has relaxed to the ground state ( at @xmath217 ) in the effective potential @xmath71 before any motion of the heavy junction sets in .
hence , initially , the effective potential for the classical junction is given by the ground state energy @xmath188 .
the absence of landau - zener tunneling ( due to the inequality ) then assures that the effective potential is given by @xmath188 during the entire motion of the classical junction .
we then can find the criterion for @xmath24 to enter a running state : during the fast ramping of the current @xmath39 , the heavy junction remains frozen at @xmath251 .
we have to check whether the potential energy of the heavy junction is sufficient to overcome all barriers in @xmath252 ; i.e. , we have to inspect if the maximum of @xmath188 in the interval @xmath253 is realized at @xmath251 .
as soon as this condition applies , the classical junction becomes delocalized .
this analysis has to be performed numerically , and the result is displayed in fig .
[ fig : mmfa ] . as a first overall result
, we note that the critical current is shifted to lower values due to the instantaneous change of the effective potential for @xmath24 by the fast ramping of the current and by the quantum decay of @xmath25 , allowing the heavy junction to transform potential into kinetic energy .
another prominent change is in the form of the transition line , specifically , the ` cut ' of the first tip of @xmath121 : the first segment at large values of @xmath38 , cf .
[ fig : mmfa ] , corresponds to the conventional case for a @xmath155-type line ; the criterion for delocalization involves the full conversion of potential to kinetic energy of the heavy junction and the subsequent tunneling of the quantum junction right at the turning point of the classical junction , cf .
inset ( 1 ) of fig .
[ fig : mmfa ] . following the @xmath254-line further with decreasing @xmath39
, we reach a point where the potential energy of the classical junction after the phase slip event is no longer sufficient to overcome the next barrier and an additional increase in @xmath39 is required to provide the missing energy .
the segment ( 2 ) then involves partial conversion of the initial potential energy within the two wells before and after the phase slip event . finally , moving along the segment ( 2 ) towards lower @xmath38 the value of @xmath24 where the phase slip occurs decreases and finally reaches @xmath251 ( cf .
[ fig : mmfa ] , inset ( 3 ) ) .
the critical line then continues along the third segment which is of the typical @xmath156-type , with the quantum junction undergoing tunneling before the classical junction starts moving .
after tunneling , a large current is required to lower the barrier in the cosine potential of the classical junction , cf.fig .
[ fig : mmfa ] , inset ( 3 ) .
note that the differentiation between positively sloped segments of type ( 1 ) and type ( 2 ) persists to segments determined by a higher fluxon index @xmath132 , although the difference is invisible on the scale of fig .
[ fig : mmfa ] .
( solid ) of the current - biased dc - squid with strongly asymmetric parameters as a function of bias current @xmath39 and inverse inductance @xmath255 .
we assume instantaneous ramping of the bias @xmath256 and a strictly classical and adiabatically slow ( as compared to the dynamics of the quantum junction ) motion of the weakly damped heavy degree of freedom @xmath150 .
for comparison , the critical line of the dissipative case is also displayed ( dashed line ) .
the inset illustrates the three segments constituting the first part of @xmath121 , where solid dots indicate points of equal potential energy .
inset ( 3 ) shows the situation right at the tip between segments ( 2 ) and ( 3 ) . ]
above , we have assumed the limit of a very large ratio @xmath175 . in the experiment , the dynamical asymmetry of the squid is implemented by shunting one of the two junctions with an external capacitance.@xcite this technique allows to reach large capacitance ratios of the order of @xmath233 .
given this finite mass ratio , one may enter a new dynamical regime where the classical junction is pulled out of its metastable state more efficiently .
this is realized when the quantum junction attains a high energy state residing at the opposite side of the parabolic potential ( large values of @xmath25 in fig .
[ fig : decay ] ) and holds on to it while dragging the heavy junction out of its metastable state .
this requires a fast@xcite current ramping @xmath257 ( on the scale of the intervalley motion of the light junction ) , pushing the quantum junction to high potential energies during the ramp - up process .
furthermore , the condition @xmath258 has to be satisfied to allow for motion of the heavy junction before relaxation of the quantum junction ( from a side - well local ground state ) back to the global minimum . in order to simplify the situation
, we assume a ( weak ) finite damping of the classical junction , @xmath259 , assuring that the classical junction relaxes before tunneling of the quantum junction ( this assumption is merely a convenience , allowing us to ignore the availability of kinetic energy for the classical junction upon fast ramping ) .
we then start from the local ground state of the original potential well at @xmath260 .
the decay proceeds in the direction of @xmath25 in the effective potential @xmath261(\varphi_2)$ ] and can end up in any of the lower local ground states for the quantum junction .
we denote the energy of the local ground state of the quantum junction in the @xmath132-th well by @xmath262 ( not to be confused with @xmath263 , the @xmath132-th excited state in the spectrum of the quantum junction ) .
we estimate these energies as @xmath264 where @xmath265 is the solution to eq . near @xmath266 at fixed @xmath24 and we have neglected the zero point energy @xmath267 .
the dynamics of @xmath24 depends strongly on the fluxon index @xmath132 of the state of the quantum junction after the decay : the larger @xmath132 , the stronger is the force on @xmath24 facilitating its escape .
we can define two extreme values for the critical current , a lower limit @xmath268 arising from the largest attainable fluxon index @xmath132 , and an upper limit @xmath269 associated with the index @xmath132 of the global minimum of the quantum junction .
we estimate the maximal index @xmath132 by comparing the energy of the state of the quantum junction before its inital decay with the height of the barrier blocking the access to the @xmath132-th minimum ; the numerical solution of the relation @xmath270(\varphi_2=\arcsin j)=v_\mathrm{eff } ( \varphi_2^{\mathrm{max},n } ) , \label{eq : farthest}\ ] ] with @xmath271 the solution to eq . near @xmath272 at @xmath66
provides us with the line @xmath273 .
the subsequent stability analysis of @xmath24 generates the lines @xmath274 ; their combination into @xmath268 is displayed in fig .
[ fig : farthest ] . and fast ramping .
the system always turns resistive for @xmath275 but never turns resistive for @xmath276 . for @xmath277 the delocalization of the system
is determined by the statistical nature of the decay process of the quantum junction involving the side minima . ]
the shape of @xmath269 is found as described before : we determine the index @xmath132 of the global minimum of @xmath73 and deduce the motion of @xmath24 in the effective potential @xmath278 .
the reduction in critical current due to the massive dynamics of @xmath24 is not as pronounced as in fig .
[ fig : mmfa ] , since the maximal critical current @xmath279 involves the process where the classical junction dissipates most of its kinetic energy : the quantum junction decays to the @xmath280 fluxon state , the classical junction relaxes to its local minimum near @xmath281 , whereupon the quantum junction decays further to its global ground state with fluxon index @xmath132 . only the potential energy gained by the classical junction in the last step can be transformed to kinetic energy , making the resulting line shown in fig .
[ fig : farthest ] differ from @xmath121 in the overdamped case ( fig .
[ fig : phase_diagram_vv ] ) .
the two critical lines @xmath268 and @xmath269 define a broad intermediate regime , see the grey area in fig .
[ fig : farthest ] , where the decay of the system is of probabilistic nature : whereas no mnchhausen decay is possible for @xmath282 , the decay of the system in the intermediate regime depends on which minimum the quantum junction @xmath25 decays to . for @xmath283 the system is always bound to decay . the critical line @xmath269 crosses twice the critical line @xmath268 near @xmath284 and for @xmath285 .
this special situation arises when the most distant accessible minimum and the global minimum are neighbors ( with indices @xmath285 and @xmath286 ) and hence strongly interrelated . with decreasing @xmath38 , both the height of the maximum near @xmath287 ( determining @xmath288 ) and the position of the minimum at @xmath48 ( @xmath285 ) with respect to the one at @xmath47 ( @xmath286 ; the competition between these minima defines @xmath289 )
are lowered .
the latter effect is slightly stronger , leading to the crossing of the lines under a small angle . upon further lowering of @xmath38 , the relevant fluxon index for @xmath288 increases twice as fast as for @xmath289 , preventing any further crossings of the critcal lines . to guarantee reasonable time scales in the experiment
requires a fast quantum dynamics near the global minimum , hence the ratio @xmath290 should be small , of order unity . for a fast ramping ,
this is in conflict with the requirement of large @xmath226 in all side - wells ; as a result , not all reachable side minima may be relevant for the dynamics of @xmath24 ( i.e. , @xmath291 does not hold ) , effectively shifting the lower critical current @xmath288 to larger values .
so far we have neglected any quantum corrections for the variable @xmath24 ; however , with a finite ratio @xmath175 such corrections may become relevant .
modifications due to quantum fluctuations in the heavy degree of freedom are twofold : on the one hand , we have disregarded the tunneling of @xmath24 through a small residual barrier , leading to a broadening of all lines of the @xmath156-type . on the other hand ,
we have assumed the dynamics of @xmath25 to occur in the effective potential @xmath292(\varphi_2)$ ] , with @xmath24 fixed , while for a finite value of @xmath222 the ground state wave function of the heavy junction acquires a finite width of the order of @xmath293 .
this leads to a broadening of all lines of the @xmath155-type , which can be interpreted as @xmath24 taking part in the tunneling of @xmath25 .
the running state following a mnchhausen decay is associated with a finite time - averaged voltage across the device .
the above results thus can be tested experimentally by measuring the voltage drop as a function of the applied bias current @xmath4 and of the inductance @xmath5 of the squid .
mapping any part of @xmath121 requires changing the coupling @xmath294 .
the latter is determined by the total inductance of the loop @xmath295 , comprising the geometric ( @xmath296 ) and kinetic ( @xmath297 ) inductances of the loop .
the former is of the order of the loop perimeter and accounts for the magnetic field energy .
the kinetic inductance relates to the geometric quantity via @xmath298 , with @xmath299 the radius of the wire and @xmath300 denoting the superconducting penetration depth .
usually , the inductance is dominated by its geometric part . however , a large inductance , as desired for our setup , can be installed by using a thin wire . furthermore , the inductance is fixed after fabrication and its modification ( e.g. , via installing a diamagnetic shield ) is rather difficult .
another way to trace the critical current line is obtained by applying an external magnetic flux @xmath301 to the sample , in which case the potential eq .
has to be replaced by @xmath302 the critical current @xmath303 then can be studied as a function of @xmath301 .
the analysis proceeds in the same way as before : a change in @xmath38 altering the opening angle of the parabola in @xmath51 is replaced by a shift in the parabola s position due to the applied flux @xmath301 . and bias current @xmath39 . here
, the coupling @xmath304 is fixed at a value @xmath305 ; the critical lines in the diagram correspond to @xmath285 .
the critical line is periodic in @xmath301 with a period @xmath306 , a consequence of the invariance under the replacements @xmath307 and @xmath308 . ] as an illustration , we discuss here the most simple case of an overdamped dynamics , analogous to sec .
[ sec : overdamped ] . in a first step ,
the system relaxes to the ground state at the minimum of the potential @xmath51 , eq .
, as given by the solutions @xmath309 of @xmath310 near @xmath311 .
the critical line @xmath312 then derives from the condition @xmath313 where both the shift of the parabola @xmath76 by @xmath314 and the flux dependence of @xmath315 are accounted for .
similarly , eq . for @xmath143 is changed to @xmath316 from which we find the critical lines @xmath317 .
the results of eqs . - are invariant under the shifts @xmath318 and @xmath319 .
the mnchhausen decay thus can be studied within a finite interval , e.g. @xmath320 $ ] , and the resulting phase diagram is displayed in fig .
[ fig : pe ] .
biasing the squid with an external magnetic flux has a convenient side effect : for negative flux @xmath301 , the crossing point of @xmath151 and @xmath154 is shifted to larger values of @xmath38 and the tunneling barrier for @xmath25 is lowered , allowing to study the mnchhausen effect using a squid with considerably smaller inductance and hence smaller size .
we have shown how a system consisting of a ` quantum ' degree of freedom coupled to a ` classical ' one , implemented with a two - junction squid , can perform a complex decay process out of a zero - voltage state involving quantum tunneling of the ` light ' junction and classical motion of the ` heavy ' junction .
we have analyzed the cases of strong and weak damping , for different methods of preparation ( fast and slow ramping ) , and for various ratios of the capacitances @xmath175 and have found the effective critical current @xmath121 exhibiting a common characteristic shape .
its zig - zag like structure originates in two competing effects : while the number @xmath132 of flux units which can enter the squid is increased with decreasing @xmath38 , the current redirected over the heavy junction at given @xmath132 is decreased . for junctions with equal critical currents , the critical line approaches the value @xmath321 in the dissipative case ; asymmetries in the critical currents of the two junctions lead to a shift of the critical line @xmath121 to values straddling the critical current of the heavy junction . going over from dissipative to massive dynamics , the kinetic energy stored in the system allows to better overcome the potential barriers and the critical line shifts to smaller values of @xmath39 ; furthermore , hysteretic effects , similar to the situation encountered for underdamped junctions , and even probabilistic behavior may show up . when the ramping is slow , the modifications with respect to the discussion in sec .
[ sec : adiabatic ] are negligible : due to the slow ramping , the quantum junction remains always in its ground state and hence the classical junction moves adiabatically in the effective potential @xmath188 created by the quantum junction .
the resulting critical current @xmath121 is identical to the one presented in fig .
[ fig : phase_diagram_adiabatic ] .
|
a classical system can not escape out of a metastable state at zero temperature .
however , a composite system made from both classical and quantum degrees of freedom may drag itself out of the metastable state by a sequential process .
the sequence starts with the tunneling of the quantum component which then triggers a distortion of the trapping potential holding the classical part .
provided this distortion is large enough to turn the metastable state into an unstable one , the classical component can escape .
this process reminds of the famous baron mnchhausen who told the story of rescuing himself from sinking in a swamp by pulling himself up by his own hair
we thus term this decay the ` mnchhausen effect ' .
we show that such a composite system can be conveniently studied and implemented in a dc - squid featuring asymmetric dynamical parameters .
we determine the dynamical phase diagram of this system for various choices of junction parameters and system preparations .
|
knots @xmath0 and @xmath1 in the 3sphere @xmath2 are said to be _ concordant _ if there exists a ( topologically ) locally flat properly embedded annulus @xmath3 $ ] in @xmath4 $ ] whose boundary is the union of @xmath5 and @xmath6 . here
@xmath7 denotes the mirror image of @xmath1 with reversed string orientation . a knot which is concordant to the unknot
is said to be _
equivalently , a knot is slice if and only if it bounds a locally flat 2disk in the 4ball . it is well - known that @xmath0 is concordant to @xmath1 if and only if the connected sum @xmath8 is slice . concordance classes , which are the equivalence classes of knots modulo concordance , form an abelian group under connected sum , and this abelian group is called _ the knot concordance group @xmath9_. in the group @xmath9 , the identity is the class of slice knots , and the inverse of the concordance class of @xmath0 is that of @xmath10 . in the 1960 s ,
levine @xcite defined _ the algebraic knot concordance group @xmath11 _ using seifert forms of knots , and showed that there is a surjective homomorphism @xmath12 .
a knot is _ algebraically slice _ if it is mapped to the identity in @xmath11 , or equivalently , if it has a metabolic seifert form .
using characters on the first homology of prime power order cyclic covers of the 3sphere branched along a knot , casson and gordon @xcite exhibited the examples of nonslice knots which are algebraically slice .
it is known that in some cases we can distinguish knots up to concordance if the knots have coprime alexander polynomials .
levine @xcite proved that if the connected sum of two knots with coprime alexander polynomials is algebraically slice , then so are both knots . in other words , if two knots are not algebraically slice and have coprime alexander polynomials , then they are not algebraically concordant to each other , and hence they are not concordant . regarding casson gordon invariants , the first author showed a similar result in @xcite : if the connected sum of two knots with coprime alexander polynomials has vanishing casson gordon invariants , then so do both knots .
in @xcite , cochran , orr and teichner revealed more structures of the group @xmath9 by constructing the filtration @xmath13 of @xmath9 which is indexed by nonnegative half - integers : @xmath14 for each @xmath15 , where @xmath16 , the group @xmath17 is the subgroup of _
@xmath18solvable _ knots .
roughly speaking , a knot @xmath0 is _
@xmath18solvable _ if the 0surgery on @xmath0 in @xmath2 , denoted @xmath19 , bounds a certain @xmath20manifold called an _
@xmath18solution _ , which can be considered as an approximation of a slice disk complement of order @xmath21 ( see definition [ def : solvability ] ) .
the filtration @xmath13 reflects classical concordance invariants at low levels : a knot is algebraically slice if and only if it is @xmath22solvable .
if a knot is @xmath23solvable , then all topological concordance invariants prior to their work in @xcite , including casson gordon invariants , vanish . for each nonnegative integer @xmath21 , cochran , orr and
teichner constructed obstructions for a knot being @xmath24solvable using von neumann @xmath25invariants ( * ? ? ?
* theorem 4.2 ) ( also see theorem [ thm : rho invaiants](1 ) ) .
note that slice knots are @xmath18solvable for all @xmath15 , and therefore these are also obstructions to a knot being slice . a basic idea of the obstructions in (
* theorem 4.2 ) is as follows . for an @xmath24solvable knot @xmath0 ,
if a representation @xmath26 for a group @xmath27 satisfying certain conditions extends to an @xmath24solution for @xmath0 , then the von neumann @xmath25invariant @xmath28 associated to @xmath29 vanishes .
therefore , it is essential to find representations on @xmath30 that extend to @xmath24solutions ( or slice disk complements ) .
for instance , to get the desired representations one can use certain quotient groups of the fundamental group of a putative @xmath24solution for @xmath0 .
this method was introduced in @xcite and used by many authors . on the other hand ,
the original idea in @xcite for getting representations that extend to an @xmath24solution is to use _ rationally universal representations_. in @xcite , for each nonnegative integer @xmath21 , they defined _ the rationally universal group @xmath31 _ and representations @xmath32 via higher - order blanchfield linking forms ( see section [ subsec : universal ] for more details ) .
they also showed that if a knot @xmath0 is @xmath24solvable , then some of these rationally universal representations @xmath33 extend to @xmath24solutions for @xmath0 and the von neumann @xmath25invariants associated to such @xmath33 vanish ( * ? ? ?
* theorem 4.6 ) ( or see definition [ def : vanishing ] and theorem [ thm:4.6 ] ) . in this paper
, we say that a knot has _ vanishing @xmath25invariants of order @xmath21 _ if it satisfies a criterion in ( * ? ? ?
* theorem 4.6 ) .
therefore , theorem 4.6 in @xcite can be rephrased that an @xmath24solvable knot has vanishing @xmath25invariants of order @xmath21 .
although it is not easy to show that an arbitrarily given @xmath24solvable knot @xmath0 has vanishing @xmath25invariants of order @xmath21 , these obstructions obtained using rationally universal representations have an advantage that they are described in terms of information only from the 3manifold @xmath19 , not using information from a putative @xmath24solution for @xmath0 , which is a 4manifold bounded by @xmath19 .
we remark that using these obstructions the first examples of nonslice knots with vanishing casson - gordon invariants were found @xcite .
regarding the approach for knot concordance using coprimeness of alexander polynomials , the authors showed that if the connected sum of two knots with coprime alexander polynomials has vanishing @xmath25invariants of order 1 , then so do both knots ( * ? ? ? * theorem 1.1 ) .
the aim of this paper is to generalize this splitting property of @xmath25invariants of order 1 to all orders @xmath34 , and obtain obstructions to knots being concordant .
our results are described in terms of higher - order alexander modules and blanchfield linking forms , which are explained below . for a knot @xmath0 ,
let @xmath35 denote the ( rational ) alexander module @xmath36)$ ] and @xmath37 the blanchfield linking form defined on @xmath38 . for knots @xmath0 and @xmath1 , let @xmath39 .
it is well - known that @xmath40 splits as @xmath41 , and similarly we have @xmath42 .
generalizing @xmath38 and @xmath37 to higher - orders , for each positive integer @xmath21 and a representation @xmath43 , the _
@xmath21-th order alexander module @xmath44 associated to @xmath33 _ and the _ @xmath21-th order blanchfield linking form @xmath45 associated to @xmath33 _ were defined in @xcite .
we note that the construction of @xmath33 depends on an inductive choice of elements @xmath46 for @xmath47 , and hence so does that of @xmath44 and @xmath45 .
in contrast to the case of @xmath48 , for @xmath49 the module @xmath50 and the form @xmath51 do not split as a direct sum in general .
notheless , in theorem [ thm : splitting h_1 ] we give a criterion for @xmath52 under which @xmath50 and @xmath51 associated to @xmath33 split as a direct sum .
loosely speaking , they split if @xmath33 is constructed by choosing @xmath53 ` along @xmath0 ' for @xmath54 , which means that we have @xmath55 where splitting of @xmath56 as a direct sum can be assumed for @xmath57 from the construction of @xmath33 .
see theorem [ thm : splitting h_1 ] for more details .
similarly , if we choose @xmath53 ` along @xmath1 ' , then @xmath50 and @xmath51 split as a direct sum . in ( * ? ? ?
* theorem 1.1 ) , coprimeness of the alexander polynomials of @xmath0 and @xmath1 was needed only to guarantee that every self - annihilating submodule @xmath58 of @xmath40 splits as @xmath59 where @xmath60 and @xmath61 are self - annihilating submodules of @xmath38 and @xmath62 , respectively .
recall that a @xmath63$]-submodule @xmath60 of @xmath38 is _ self - annihilating with respect to @xmath37 _ if @xmath64 , where @xmath65 we generalize the condition of having coprime alexander polynomials to higher - orders : for each nonnegative integer @xmath21 and a knot @xmath66 where @xmath67 , we define the knot @xmath66 to have _ splitting self - annihilating submodules of order @xmath21 for @xmath0 _ , roughly speaking , if every self - annihilating submodule @xmath68 of @xmath50 splits as @xmath69 ( where @xmath70 and @xmath71 are self - annihilating submodules of @xmath44 and @xmath72 , respectively ) whenever @xmath73 is constructed by choosing @xmath53 ` along @xmath0 ' for @xmath74 in such a way that each @xmath53 is contained in a self - annihilating submodule of @xmath75 .
see definition [ def : splitting sas ] for the precise definition .
we remark that when @xmath49 , this condition is not symmetric with respect to @xmath0 and @xmath1 .
we generalize the coprimeness condition to higher - orders as described above for technical reasons . instead of the condition for higher - orders defined above , one might want to use the condition of having coprime _ higher - order alexander polynomials_. however , compared to the case of @xmath48 , we get more technical difficulties for using higher - order alexander polynomials since higher - order alexander polynomials are defined in _ noncommutative rings_. for example , for a noncommutative ring @xmath76 and @xmath77 , the element @xmath78 does not annihilates the right @xmath76-module @xmath79 in general . also note that when @xmath48 , in the case of classical alexander polynomials and blanchfield linking forms , the condition of having splitting self - annihilating submodules is weaker than that of having coprime alexander polynomials .
now we are ready to state our main theorem .
[ main ] let @xmath21 be a positive integer .
let @xmath0 and @xmath1 be knots such that @xmath80 has splitting self - annihilating submodules of order @xmath81 for @xmath0 . if @xmath80 has vanishing @xmath25invariants of order @xmath21 , then so does @xmath0 .
note that the case @xmath82 in theorem [ main ] results in theorem 1.1 in @xcite .
since an @xmath24solvable knot has vanishing @xmath25invariants of order @xmath21 , theorem [ main ] can be used as a concordance obstruction . as an application of theorem [ main ] , we show the following : in @xcite , cochran , orr and teichner gave the first examples of knots which are @xmath83solvable but not @xmath84solvable .
these knots are also the first examples of nonsilce knots with vanishing casson - gordon invariants .
let @xmath0 be one of these knots .
recall that _ the concordance genus of @xmath0 _ is defined to be the minimum of the genus of @xmath85 among all knots @xmath85 concordant to @xmath0 . in theorem
[ thm : example ] , we show that @xmath0 is not concordant to any knot whose alexander polynomial has degree less than or equal to two , which implies that @xmath0 is not concordant to any knot of genus 1 .
since @xmath0 has genus 2 , this shows that @xmath0 has concordance genus 2 .
note that these are the first examples of knots that are @xmath83solvable that are proved to be of concordance genus greater than 1 .
we remark that in @xcite cochran , harvey and leidy defined a refinement of the filtration @xmath13 via sequences of ( alexander ) polynomials , and using it they distinguished , up to concordance , knots which have coprime alexander polynomials or coprime higher - order alexander polynomials .
they also showed that the aforementioned knots in @xcite are not concordant to their knots in @xcite , which are not restricted to genus one knots .
our results are related to the following more general question which motivated our work : if two nonslice knots have coprime alexander polynomials , are they mutually nonconcordant ? one can see that the answer is no in smooth concordance due to the existence of knots with trivial alexander polynomial that are not smoothly slice .
in topological concordance , however , the answer is unknown to the authors knowledge .
theorem [ main ] above and the work of levine , cochran
leidy , and the authors in @xcite can be considered as evidence that the answer might be yes . in @xcite one can also find interesting results along this direction which involve the quotients of successive terms of the filtration @xmath13 and higher - order alexander polynomials . also regarding the question above , one may ask whether or not @xmath80 has splitting self - annihilating submodules of order @xmath21 for @xmath0 ( and for @xmath1 ) for all @xmath86 if @xmath0 and @xmath1 have coprime alexander polynomials , but the authors do not know the answer .
this paper is organized as follows . in section [ sec : preliminaries ] , we give background materials such as the notion of @xmath18solvability , rationally universal groups , higher - order alexander modules and blanchfiled linking forms .
section [ sec : vanishing ] is devoted to the proof of theorem [ main ] .
we give examples in section [ sec : examples ] . in this paper ,
unless mentioned otherwise all manifolds are assumed to be connected and oriented .
in this section , we briefly review work of cochran , orr and teichner on the filtration @xmath13 of @xmath9 in @xcite to have the paper self - contained .
we claim no originality in this section except subsection [ subsec : vanishing ] .
let @xmath87 be a group and @xmath21 a nonnegative integer .
@xmath21th derived group of @xmath87 _ , denoted @xmath88 , is defined inductively as follows : @xmath89 and @xmath90 $ ] for @xmath91 .
let @xmath92 be a closed 3manifold and @xmath93 a spin 4manifold such that @xmath94 .
let @xmath95 .
for @xmath96 , we have the intersection form @xmath97)\times h_2(w;{\mathbb{z}}[\pi/\pi^{(n ) } ] ) \to { \mathbb{z}}[\pi/\pi^{(n)}].\ ] ] also there is the self - intersection invariant @xmath98 .
refer to ( * ? ? ?
* section 7 ) for the definition of @xmath98 .
[ def : solvability ] let @xmath21 be a nonnegative integer . a closed 3manifold @xmath92 is _
@xmath18solvable via @xmath93 _ if there exists a spin 4manifold @xmath93 with @xmath99 satisfying the following : 1 .
@xmath100 is an isomorphism , 2 . for @xmath101 and @xmath102 ,
there exist elements @xmath53 and @xmath103 in @xmath104)$ ] such that @xmath105 and @xmath106 for all @xmath107 .
then @xmath93 is said to be an _
@xmath18solution _ for @xmath92 .
the 3manifold @xmath92 is _ @xmath24solvable via @xmath93 _ , and @xmath93 is called an _ @xmath24solution _ for @xmath92 , if @xmath93 satisfies the following additional condition : * for @xmath108 , there exist elements @xmath109 mapping to @xmath53 under the canonical homomorphism @xmath110)\to h_2(w;{\mathbb{z}}[\pi/\pi^{(n)}])$ ] such that @xmath111 for all @xmath107 . for a knot @xmath0 and @xmath112 , we say that @xmath0 is _
@xmath113solvable via @xmath93 _ if @xmath19 , the 0surgery on @xmath0 in @xmath2 , is @xmath113solvable via @xmath93 . in this case
, @xmath93 is said to be an _
@xmath113solution for @xmath0_. we denote by @xmath114 the set of all ( concordance classes of ) @xmath113solvable knots . in @xcite , cochran , orr and teichner showed that @xmath115 if @xmath116 , and they also showed that slice knots are @xmath113solvable for all @xmath117 .
there are other variations of the notion of @xmath113solvability such as rational @xmath113solvability which uses rational coefficients instead of integer coefficients in its definition .
it is easy to see that @xmath113solvable knots are rationally @xmath113solvable . for a nonnegative integer @xmath21 ,
we define _ rationally universal groups _ @xmath31 as follows .
* definition 3.1)[def : ru groups ] we define _ the 0-th rationally universal group @xmath118 _ to be the infinite cyclic group @xmath119 generated by @xmath120 . for a nonnegative integer @xmath21 , we let @xmath121-\{0\})^{-1}\ ] ] and define _ the @xmath122-st rationally universal group @xmath123 _ to be @xmath124 , where @xmath125 is the right ring of quotients of @xmath126 .
that is , the ring @xmath125 is the ( skew ) quotient field of @xmath127 . in @xcite , these @xmath31 and @xmath128 are denoted by @xmath129 and @xmath130 , respectively .
note that @xmath125 are equipped with involutions induced from taking @xmath131 for @xmath132 .
we denote the involution of @xmath133 by @xmath134 .
one can easily see that @xmath31 is @xmath18solvable , that is , @xmath135 , and that @xmath31 embeds into @xmath128 .
furthermore , the groups @xmath31 are poly - torsion - free - abelian ( abbreviated ptfa ) , that is , each @xmath31 admits a normal series of finite length whose successive quotient groups are torsion - free abelian .
note that @xmath136 $ ] and @xmath137 , where @xmath138 is the quotient field of @xmath139 $ ] . for higher order cases
, we have the following proposition .
* corollary 3.3)[prop : ru ] for @xmath96 , the rings @xmath128 are left and right principal ideal domains and isomorphic to ( skew ) laurent polynomial rings @xmath140 $ ] , where @xmath141 are the ( skew ) quotient fields of @xmath142 $ ] . in particular ,
the ( skew ) fields @xmath125 are isomorphic to @xmath143 .
we explain how to obtain representations @xmath144 for a closed 3manifold @xmath92 with first betti number 1 .
the example for @xmath92 we should keep in mind is @xmath19 , the 0surgery on a knot @xmath0 in @xmath2 . to this end
, we define higher - order alexander modules and blanchfield linking forms .
we recall the classical invariants .
let @xmath145 be the homomorphism induced by the abelianization .
for example , if @xmath92 is the 0surgery @xmath19 , then we can choose @xmath146 to be the abelianization sending a meridian of a knot @xmath0 to @xmath120 .
then we obtain a coefficient system @xmath147 , and from this we obtain @xmath148 as a ( right ) @xmath149module , which is _ the 0-th order alexander module @xmath150 _ of @xmath0 . this alexander module is the same as the classical rational alexander module .
it is well - known that we have the blanchfield form @xmath151 , which is nonsignular in the sense that it provides an isomorphism @xmath152 .
now for a nonnegative integer @xmath21 , suppose that we have a representation @xmath153 . then _
the @xmath21-th order alexander module of @xmath92 associated to @xmath29 _ , denoted @xmath154 , is defined to be the first homology @xmath155 with the twisted coefficient system given by @xmath156 .
we write @xmath157 for @xmath154 when @xmath158 .
we suppress superscripts when they are understood from the context .
* proposition 2.11 , theorem 2.13)[thm : blanchfield form ] let @xmath92 be a closed 3manifold with first betti number 1 . if @xmath153 is a nontrivial representation , then @xmath159 is a ( right ) @xmath128-torsion module , and there is a nonsingular symmetric linking form @xmath160 .
the linking form @xmath161 is _ the @xmath21-th order blanchfield form on @xmath159_. when @xmath48 , this is the same as the classical rational blanchfield form .
we write @xmath162 for @xmath161 when @xmath158 , and sometimes suppress superscripts for simplicity .
we describe how @xmath163 is defined : for @xmath164,[y]\in { \mathcal{a}}_n$ ] , there exists @xmath165 such that @xmath166 for some 2-chain @xmath167
. then @xmath168,[y ] ) = \frac{1}{r}(c\cdot y)$ ] where @xmath169 is the equivariant intersection over @xmath128 .
now we explain how to obtain a representation @xmath144 assuming we are given a representation @xmath170 .
recall that @xmath171 .
we define @xmath172 to be the set of representations @xmath144 modulo @xmath173-conjugations such that @xmath174 where @xmath175 is the canonical projection . in the theorem
below , we show how to obtain representations @xmath144 by inductively choosing elements from alexander modules .
* theorem 3.5)[thm : extension ] let @xmath92 be a closed 3manifold with first betti number 1 .
let @xmath176 be the abelianization .
suppose that @xmath34 and we have constructed a representation @xmath177 , the @xmath178-st order alexander module @xmath179 associated to @xmath180 , and the @xmath178-st order blanchfield form @xmath181 . then the following hold . 1 .
there is a bijection @xmath182 which is natural with respect to continuous maps .
there is an isomorphism @xmath183 with @xmath184 giving a natural bijection @xmath185 . in particular
, each choice of an element @xmath186 induces a representation @xmath144 such that @xmath174 .
this @xmath33 will also be denoted by @xmath187 to indicate its dependency of @xmath188 .
3 . if @xmath186 , then the image of @xmath188 under the map @xmath189 is the map given by @xmath190 . in particular , for @xmath191 and @xmath192 with @xmath193 ( hence @xmath194 can be considered as an element in @xmath179 ) , if @xmath195 , then @xmath196 for the representation
@xmath191 constructed as in theorem [ thm : extension](2 ) , we say that @xmath33 is _ associated to @xmath180 and @xmath188_. theorem [ thm : extension](2 ) implies that for @xmath74 , by inductively choosing @xmath197 , we can obtain a representation @xmath144 , which is denoted by @xmath198 , or more simply , @xmath199 . in this case
, we also say that @xmath33 , @xmath200 , and @xmath201 are _ associated to @xmath202_. for a later use , we consider the case that @xmath33 is obtained by taking @xmath203 where @xmath204 is the canonical injection to the right summand , that is , @xmath33 is associated to @xmath180 and @xmath205 . in this case , the coefficient system @xmath206 factors through @xmath207 and we have the following theorem .
* theorem 5.16)[thm : factoring through ] let @xmath21 be a positive integer .
if @xmath144 factors through @xmath170 as above , then the following hold . 1
. @xmath208 as ( right ) @xmath128-modules .
2 . for the canonical homomorphism @xmath209 and @xmath210 and @xmath211 , @xmath212 now if @xmath144 factors through @xmath176 , then by applying theorem [ thm : factoring through ] repeatedly we obtain that @xmath213 and @xmath214 where @xmath215 is the canonical homomorphism .
if @xmath144 factors through @xmath176 , then we write @xmath216 for @xmath217 .
let @xmath92 be a closed @xmath218manifold , and let @xmath219 be a representation where @xmath27 is a countable discrete group
. then one can define the von neumann @xmath25invariant @xmath220 of @xmath92 associated to @xmath29 ( see @xcite ) .
if @xmath29 extends to @xmath221 for a compact oriented @xmath20manifold @xmath93 with @xmath99 , then the invariant @xmath222 can be computed as @xmath223 where @xmath224 denotes the @xmath225signature of the intersection form on @xmath226 and @xmath227 the ordinary signature of @xmath93 .
( the existence of such @xmath228 and @xmath93 can be assumed for the computation of @xmath222 by theorem [ thm : rho invaiants](2 ) below and weinberger s work . ) for a knot @xmath0 and a representation @xmath229 , we write @xmath230 for @xmath28 . recall that @xmath18solvable 3manifolds ( or knots ) are rationally @xmath18solvable .
* theorem 4.2 ) ( * ? ? ?
* section 2)[thm : rho invaiants ] let @xmath92 be a closed 3manifold and @xmath27 a group . 1 .
let @xmath27 be an @xmath18solvable ptfa group .
if @xmath92 is ( rationally ) @xmath24solvable via @xmath93 to which the coefficient system @xmath231 extends , then @xmath232 .
( subgroup property ) if @xmath219 factors through @xmath233 , then @xmath234 .
3 . for a knot
@xmath0 , if @xmath235 is a nontrivial homomorphism , then @xmath230 equals the integral of the levine - tristram signature function @xmath236 of the knot @xmath0 , integrated over the circle normalized to length one .
in addition , if @xmath0 is algebraically slice , then @xmath237 .
4 . if @xmath219 is a trivial homomorphism , then @xmath238 .
cochran , orr and teichner investigated in which cases representations to @xmath31 extend to @xmath24solutions , and obtained their obstructions to a knot being @xmath24solvable , which are given below . recall that an @xmath128submodule @xmath239 of @xmath159 is said to be _ self - annihilating with respect to @xmath163 _ if @xmath240 where @xmath241 definition [ def : vanishing ] and theorem [ thm:4.6 ] below can be found in ( * ? ? ?
* theorem 4.6 ) .
[ def : vanishing ] let @xmath21 be a positive integer . a knot @xmath0 has _ vanishing @xmath25invariants of order @xmath21 _ if it satisfies the following conditions @xmath242@xmath18 .
* there exists a self - annihilating submodule @xmath243 of @xmath244 such that for each @xmath245 and the corresponding representation @xmath246 , the invariant @xmath247 .
* for each @xmath245 , there exists a self - annihilating submodule @xmath248 of @xmath249 such that for each @xmath250 and the corresponding representation @xmath251 , the invariant @xmath252 . * for each @xmath253 with @xmath245 , there exists a self - annihilating submodule @xmath254 of @xmath255 such that for each @xmath256 and the corresponding representation @xmath257 , the invariant @xmath258 .
+ * for each @xmath259 with @xmath260 , @xmath261 , there exists a self - annihilating submodule @xmath262 of @xmath263 such that for each @xmath264 and the corresponding representation @xmath265 , the invariant @xmath266 .
below we state the obstruction to a knot being @xmath24solvable .
* theorem 4.6)[thm:4.6 ] let @xmath21 be a positive integer . if a knot @xmath0 is ( rationally ) @xmath24solvable , then @xmath0 has vanishing @xmath25invariants of order @xmath21 .
we give another description of the condition of having vanishing @xmath25invariants of order @xmath21 .
for a knot @xmath0 and a submodule @xmath239 of @xmath267 , we say that @xmath239 is _ associated to @xmath268_. [ def : vanishing 2 ] let @xmath21 be a positive integer . for a knot @xmath0 ,
let @xmath269 be a nonempty set of @xmath21-tuples @xmath270 where each @xmath271 is a self - annihilating submodule of @xmath267 associated to a representation @xmath272 for @xmath74 .
the set @xmath269 is a _ vanishing system of submodules of order @xmath21 for @xmath0 _ if for each @xmath273 the following hold . 1 . for each @xmath274 such that @xmath275 , the representation @xmath268 is associated to @xmath276 and some @xmath277 .
2 . if @xmath278 for some @xmath279 , then there exists @xmath280 with @xmath281 associated to @xmath282 for @xmath283 such that the following hold : 1 . for @xmath284
, we have @xmath285 and @xmath286 , 2 . the representation @xmath287 is associated to @xmath268 and @xmath188 .
3 . for @xmath74 , the invariant @xmath288 . furthermore , for each @xmath264 and @xmath265 , the invariant @xmath266 . it is easy to see the following proposition , of which proof is omitted .
[ prop : vanishing ] let @xmath21 be a positive integer .
a knot @xmath0 has vanishing @xmath25invariants of order @xmath21 if and only if it has a vanishing system of submodules of order @xmath21 .
in this section , we investigate @xmath25invariants for a knot which is given as the connected sum of a pair of knots , and prove theorem [ main ] .
let @xmath0 and @xmath1 be knots in @xmath2 , and let @xmath67 .
we denote by @xmath19 the 0surgery on @xmath0 in @xmath2 , and denote by @xmath289 and @xmath290 the 0surgery on @xmath1 in @xmath2 and the 0surgery on @xmath66 in @xmath2 , respectively . throughout this section , unless mentioned otherwise , the symbol @xmath92 denotes @xmath290 .
we write @xmath291 for @xmath292 , the exterior of @xmath0 , and write @xmath293 and @xmath294 for those of @xmath1 and @xmath66 , respectively .
let us denote meridians of @xmath0 , @xmath1 , and @xmath66 by @xmath295 , @xmath296 , and @xmath297 , respectively .
we also write @xmath298 , @xmath299 , and @xmath300 for longitudes of @xmath0 , @xmath1 , and @xmath66 , respectively .
we abuse notations so that they also represent their homology classes and homotopy classes as well . for a group @xmath87 and elements @xmath301
, we denote the normal subgroup of @xmath87 generated by @xmath302 , @xmath303 , by @xmath304 .
we aim to find certain conditions under which the vanishing property of @xmath25invariants splits via connected sum operation . in doing so
, we need to verify that certain representations @xmath305 , @xmath306 , induce representations @xmath307 so that we can define higher - order alexander modules and @xmath25invariants of @xmath0 .
the following lemma is well - known .
for instance , see @xcite .
[ lem : pi_1 of m ] @xmath308 .
note the following equalities : @xmath309 one can prove that @xmath310 similarly . therefore , if we have a representation @xmath311 for some @xmath274 with @xmath312 , then one can obtain a representation @xmath313 by taking compositions @xmath314 .
similarly , for a given representation @xmath311 for some @xmath274 with @xmath312 , the representation @xmath315 is induced from @xmath268 .
we note the restrictions of @xmath316 to @xmath291 is the same as that of @xmath268 to @xmath291 , that is , @xmath317 . since @xmath318 , the homomorphism @xmath319 extends to @xmath30 , and this extension of @xmath319 to @xmath30 is the same as @xmath316 , and similarly for @xmath289 and @xmath320 .
henceforth we drop the superscripts from @xmath316 and @xmath320 if they are clearly understood in the context . for @xmath176 ,
since @xmath321 and @xmath118 is @xmath322solvable , we have @xmath323 .
therefore , the representations @xmath146 are induced for @xmath30 and @xmath324 , and it is well - known that @xmath41 and @xmath325 .
but this direct sum decomposition does not hold in general for higher - order cases .
in fact , @xmath326 and sometimes the inequality becomes strict .
nonetheless , we can obtain the direct sum decompositions @xmath327 and @xmath328 if the elements @xmath53 are inductively chosen from the summands corresponding to @xmath0 while constructing representations @xmath329 .
this procedure is stated in the following theorem .
[ thm : splitting h_1 ] under the notations introduced above , the following hold .
* let @xmath176 be the abelianization .
then the following hold . * * @xmath330 , and we have induced representations @xmath146 on @xmath30 and @xmath324 . *
* @xmath331 and @xmath325 .
* let @xmath332 and @xmath333 be the representation associated to @xmath334 .
then the following hold : * * @xmath335 , and we have induced representations @xmath336 on @xmath30 and @xmath324 . * * @xmath337 factors through @xmath146 , and hence @xmath338 and @xmath339 . *
* @xmath340 and @xmath341 .
* let @xmath342 and @xmath343 be the representation associated to @xmath344 .
then the following hold : * * @xmath345 , and we have induced representations @xmath346 on @xmath30 and @xmath324 . * * @xmath347 factors through @xmath146 , and hence @xmath348 and @xmath349 . * * @xmath350 and @xmath351 .
+ * let @xmath352 and @xmath73 be the representation associated to @xmath353 .
then the following hold : * * @xmath354 , and we have induced representations @xmath33 on @xmath30 and @xmath324 . * * @xmath355 factors through @xmath146 , and hence @xmath356 and @xmath357 . * * @xmath358 and @xmath328 .
+ we use an induction argument .
the statement @xmath322 is a well - known classical result . for @xmath49 , assuming the statements @xmath242 through @xmath178 hold , we prove that the statement @xmath18 holds . for @xmath359 , the representation @xmath360 is defined by theorem [ thm : extension ] . since @xmath361 by the assumption , we can consider @xmath299 as an element in @xmath362 . by the assumption that @xmath363
, we can express @xmath299 as @xmath364 .
then @xmath365 , and by theorem [ thm : extension](3 ) we have @xmath354 .
we show that @xmath366 for all @xmath367 , which will imply that @xmath355 factors through @xmath146 . by theorem [ thm : factoring through ]
, this will also show that @xmath356 and @xmath357 .
let @xmath367 .
then we can regard @xmath194 as an element in @xmath368 .
since @xmath331 and @xmath325 and @xmath369 , it follows that @xmath370 . then by theorem [ thm : extension](3 ) , @xmath371 , and @xmath194 can be considered as an element in @xmath372 . using this argument repeatedly and by the assumption that the statements @xmath242@xmath178 hold
, we obtain that @xmath373 . to prove that @xmath358 , we first show that @xmath374 .
note that @xmath375 where the meridians @xmath376 and @xmath377 are identified as the meridian @xmath378 .
then with the coefficient systems induced from @xmath33 , we obtain the following exact sequence using a mayer - vietoris argument .
@xmath379 note that since @xmath378 and @xmath293 are subspaces of @xmath289 , the maps @xmath33 on @xmath380 and @xmath381 factor through @xmath146 .
since theorem [ thm : factoring through](1 ) works for any spaces as well , we have @xmath382 and @xmath383 . in particular , @xmath384 as right @xmath128-modules , and the map @xmath274 in the above exact sequence is injective .
( in fact , if @xmath34 and @xmath385 does not factor through @xmath146 , then @xmath386 and the map @xmath274 is an isomorphism , but we do not need this fact in this paper . ) since @xmath387 and @xmath388 , we have @xmath389 .
thus @xmath390 . using the above direct sum decomposition of @xmath391 and the fact that @xmath392
, we can deduce that @xmath393 here the second equality follows from a mayer - vietoris sequence .
moreover , since @xmath394 , we have @xmath395 in @xmath396 and @xmath397 . therefore , @xmath398 finally
, we show that @xmath328 .
let @xmath164 $ ] and @xmath399 $ ] be classes in @xmath400 .
then @xmath401,[y])=\frac{1}{r}(c\cdot y)$ ] , where @xmath402 is a chain in @xmath403 such that @xmath404 for some @xmath165 and @xmath169 is the equivariant intersection over @xmath128 . using that @xmath405 , @xmath406 , @xmath407 , and @xmath408 , one can see that a class in @xmath155 can be represented by a sum of chains supported by @xmath291 or @xmath293 .
therefore we can obtain the desired splitting of @xmath51 as in the case of classical blanchfield form . in showing that a knot @xmath0 has vanishing @xmath25invariants of higher - orders , we need existence of certain self - annihilating submodules of higher - order alexander modules of @xmath0 .
indeed , in @xcite the authors showed that @xmath0 has vanishing @xmath25invariants of order 1 under the condition that @xmath0 and @xmath1 have coprime alexander polynomials .
this coprimeness condition was necessary only to guarantee that every self - annihilating submodule of @xmath40 can be decomposed as a direct sum of self - annihilating submodules of @xmath35 and @xmath368 .
though we can define higher - order alexander polynomials up to certain indeterminacy , over a noncommutative ring such as @xmath128 for @xmath49 it is not easy to use them .
moreover , eventually we need splitting of self - annihilating submodules of @xmath50 .
therefore , we choose to generalize the condition of coprimeness of classical alexander polynomials to higher - order cases in terms of splitting of self - annihilating submodules . since the construction of @xmath50 depends on the choice of @xmath409 , @xmath74 , we need an assumption that for _ every _ @xmath50 associated to certain @xmath53 for @xmath74 , _ every _ self - annihilating submodule of @xmath50 is decomposed as a direct sum .
we state this condition carefully below .
we recall that @xmath410 where @xmath67 .
[ def : splitting sas ] under the notations introduced above : * let @xmath176 be the abelianization . recall that by theorem [ thm : splitting h_1 ] we have @xmath331 and @xmath325 . * * a connected sum @xmath66 has _ splitting self - annihilating submodules of order 0 for @xmath0 _ if every self - annihilating submodule @xmath58 of @xmath40 is decomposed as @xmath411 where @xmath60 and @xmath61 are self - annihilating submodules of @xmath35 and @xmath368 .
* let @xmath66 be a connected sum that has splitting self - annihilating submodules of order 0 for @xmath0 .
* * an element _ @xmath412 belongs to @xmath0 _ if there exists a self - annihilating submodule @xmath58 of @xmath40 such that @xmath413 .
* * recall that if @xmath414 is a representation associated to @xmath334 which belongs to @xmath0 , then by theorem [ thm : splitting h_1 ] we have @xmath340 and @xmath341 .
a connected sum @xmath66 has _ splitting self - annihilating submodules of order 1 for @xmath0 _ if for every @xmath336 associated to some @xmath334 which belongs to @xmath0 , every self - annihilating submodule @xmath415 of @xmath416 is decomposed as @xmath417 where @xmath418 and @xmath419 are self - annihilating submodules of @xmath420 and @xmath372 . + * let @xmath66 be a connected sum that has splitting self - annihilating submodules of order @xmath81 for @xmath0 . * * for @xmath421 , let @xmath409 be elements which belong to @xmath0 . for @xmath422 where @xmath362 is associated to @xmath423 , the elements _
@xmath424 belong to @xmath0 _ if there exists a self - annihilating submodule @xmath425 of @xmath362 such that @xmath426 . *
* recall that if @xmath144 is a representation associated to @xmath427 which belong to @xmath0 , then by theorem [ thm : splitting h_1 ] we have @xmath358 and @xmath328 . a connected sum
@xmath66 has _ splitting self - annihilating submodules of order @xmath21 for @xmath0 _ if for every @xmath33 associated to some @xmath424 which belong to @xmath0 , every self - annihilating submodule @xmath68 of @xmath50 is decomposed as @xmath428 where @xmath70 and @xmath71 are self - annihilating submodules of @xmath429 and @xmath430 .
+ note that @xmath66 has splitting of self - annihilating submodules of order @xmath431 for @xmath0 and @xmath1 if @xmath0 and @xmath1 have coprime alexander polynomials ( for example , see @xcite ) .
now we are ready to give a proof of theorem [ main ] .
recall that we let @xmath67 and @xmath410 . by proposition [ prop : vanishing ]
, the knot @xmath66 has a vanishing system of submodules of order @xmath21 , denoted @xmath269 . using @xmath269
, we will construct a vanishing system of submodules of order @xmath21 for @xmath0 , which will complete the proof .
let @xmath432 , where for @xmath275 each @xmath271 is a self - annihilating submodule of @xmath433 such that @xmath268 is associated to @xmath276 and @xmath277 .
if the elements @xmath434 belong to @xmath0 , we say that _
@xmath239 belongs to @xmath0_. now if @xmath239 belongs to @xmath0 , then by theorem [ thm : splitting h_1 ] , for @xmath74 we obtain the representations @xmath435 and @xmath315 which are induced from @xmath268 , and we have @xmath436 and @xmath437 .
furthermore , since @xmath66 has splitting self - annihilating submodules of order @xmath81 for @xmath0 , for @xmath74 we have @xmath438 where @xmath439 and @xmath440 are self - annihilating submodules of @xmath441 and @xmath442 , respectively . for each @xmath443 which belongs to @xmath0 , let @xmath444 .
now we let @xmath445 . we show that @xmath446 is nonempty .
pick an arbitrary element @xmath447 in @xmath269 where each @xmath268 is associated to @xmath276 and @xmath448 for @xmath275 . since @xmath66 has splitting self - annihilating submodules of order @xmath81 ,
we have @xmath449 . pick @xmath450 .
since property ( 2 ) in definition [ def : vanishing 2 ] holds for @xmath239 , after re - labeling if necessary , we may assume that @xmath336 is associated to @xmath146 and @xmath334 .
since @xmath334 belongs to @xmath0 , and since @xmath66 has splitting of self - annihilating submodules of order @xmath81 , we have @xmath451 .
now pick @xmath452 . by using the argument above again
, we may assume that @xmath346 is associated to @xmath336 and @xmath453 , and that @xmath454 . by repeating this argument as necessary ,
we can find @xmath443 such that each @xmath268 is associated to @xmath276 and @xmath455 for @xmath275 , which implies that @xmath239 belongs to @xmath0 .
we show that @xmath446 is a vanishing system of submodules of order @xmath21 for @xmath0 .
we need to show that each @xmath456 in @xmath446 satisfies properties ( 1)(3 ) in definition [ def : vanishing 2 ] .
let @xmath456 be an element in @xmath446 as defined above . since @xmath427 belong to @xmath0 , there exist @xmath457 for @xmath74 such that @xmath458 .
note that each @xmath75 is associated to @xmath316 .
therefore , to show that @xmath456 satisfies property ( 1 ) , it suffices to show that each @xmath316 is associated to @xmath459 and @xmath460 for @xmath275 . for every @xmath461 , since @xmath462 , we have @xmath463 where @xmath464 . also note that @xmath316 is induced from @xmath268 which is associated to @xmath276 and @xmath465 for @xmath275 . combining this observation with theorem [ thm : extension ]
, one can deduce that the representation @xmath307 associated to @xmath459 and @xmath460 is the same as @xmath316 .
we show that @xmath456 satisfies property @xmath83 .
fix an integer @xmath274 such that @xmath421 .
let @xmath466 , and let @xmath467 . since @xmath239 satisfies property ( 2 ) , there exists @xmath468 with each @xmath281 associated to some @xmath469 such that the following hold : * for @xmath284 , we have @xmath285 and @xmath286 , * the representation @xmath287 is associated to @xmath268 and @xmath470 . in particular , we have @xmath471 where @xmath472 and @xmath473 for @xmath474 .
then since @xmath66 has splitting self - annihilating submodules of order @xmath81 , and since @xmath470 belongs to @xmath0 by its definition , we have @xmath475 for some self - annihilating submodules @xmath476 and @xmath477 of @xmath478 and @xmath479 . now , choose an arbitrary element @xmath480 , and we let @xmath481
. then @xmath482 belongs to @xmath0 . using the argument above again with @xmath482 , after re - labeling if necessary
, we may assume that @xmath483 splits as @xmath484 .
now by repeating this argument as necessary , we may assume that for @xmath485 we also have @xmath486 and the representation @xmath487 is associated to @xmath488 and some @xmath489 such that @xmath490 .
( recall that @xmath281 are equal to @xmath491 for @xmath284 . )
therefore , @xmath492 belongs to @xmath0 , and we obtain @xmath493 where @xmath494 . for @xmath495 , we have @xmath496 for @xmath474 and @xmath497 is associated to @xmath316 and @xmath498 , and hence property ( 2 ) holds for @xmath456 . for property ( 3 ) , we construct a cobordism @xmath499 between the disjoint union @xmath500 and @xmath92 .
this construction is well - known , and it is briefly described below .
( for details , refer to ( * ? ? ?
* section 4 ) . )
( 291,87 ) ( 0,0 ) ( 22,43)(0,0)@xmath0 ( 273,43)(0,0)@xmath1 ( 145,19)(0,0)@xmath501 ( 145,62)(0,0)[b]@xmath431 ( 73,11)(0,0)@xmath431 ( 219,11)(0,0)@xmath431 attach a 1handle at the top level between @xmath502 $ ] and @xmath503 $ ] so that the upper boundary is the 0surgery on the split link @xmath504 . then attach a 2handle along @xmath501 with zero framing as indicated in figure [ fig:1 ] .
the resulting 4manifold is the desired @xmath499 .
its boundary at the bottom , @xmath505 , is the disjoint union @xmath506 and the boundary at the top , @xmath507 , is @xmath92 .
also one can easily see that @xmath508 , and by lemma [ lem : pi_1 of m ] we obtain that @xmath509 . for @xmath443 which belongs to @xmath0 and @xmath510
as defined above , we need to show that @xmath511 for @xmath74 and @xmath512 for every representation @xmath513 which is associated to @xmath514 and some @xmath515 .
let @xmath274 be an integer such that @xmath74 .
since @xmath516 , the representation @xmath268 extends to @xmath499 .
denote the extension of @xmath268 to @xmath499 by @xmath517 .
moreover , on @xmath500 which is a subspace of @xmath499 , this extension @xmath517 agrees with the representations @xmath316 on @xmath30 and @xmath320 on @xmath324 induced from @xmath311 .
then we have @xmath518 where @xmath519 is the @xmath225-signature on @xmath520 and @xmath521 is the ordinary signature of @xmath499 . by (
* lemma 4.2 ) , we get @xmath522 , and therefore @xmath523 . since property
( 3 ) holds for @xmath239 , we have @xmath524 . by the construction of @xmath239 , the representation @xmath320 factors through @xmath146 . by the assumption ,
the knot @xmath1 has a self - annihilating submodule of @xmath368 , and therefore @xmath1 is algebraically slice . by theorem [ thm : rho invaiants ] , it follows that @xmath525
. therefore @xmath511 .
let @xmath513 be a representation which is associated to @xmath514 and some @xmath515 .
then @xmath526 is the same as the representation induced from @xmath144 which is associated to @xmath180 and @xmath527 where @xmath528 .
note that @xmath529 since property ( 3 ) holds for @xmath239 .
then we can show that @xmath512 similarly as above , and this shows that @xmath456 satisfies property ( 3 ) .
in this section , using theorem [ main ] we show that the knots constructed by cochran , orr and teichner in @xcite , which are @xmath83solvable but not @xmath84solvable , are not concordant to any knot whose alexander polynomial has degree less than or equal to two .
since the knots of cochran orr teichner have genus 2 , this will imply that they are not concordant to genus 1 knots , and hence they have concordance genus 2 .
we refer the reader to @xcite for the detailed construction of the knots of cochran , orr and teichner . instead , in the next lemma we list some properties of the knots which we will need later .
before listing the properties , we discuss some technical issues .
let @xmath0 be one of the knots of cochran orr teichner .
for each @xmath530 where @xmath60 is a self - annihilating submodule , there is a representation @xmath531 associated to @xmath146 and @xmath188 . using the arguments in @xcite , we can replace @xmath532 by a subgroup @xmath533 of @xmath532 such that @xmath533 contains the image of @xmath336 and @xmath534 where @xmath535/(t^2 - 3t+1)^m$ ] for some positive integer @xmath536 .
( this @xmath533 is denoted by @xmath532 in @xcite . also see @xcite . )
that is , re - labeling @xmath533 by @xmath532 , theorem [ main ] and related theorems in this paper are all available .
henceforth , in this section we use this re - labeled @xmath532 .
we list properties of @xmath0 in lemma [ lem : cot knot property ] below , which can be found in @xcite .
recall from proposition [ prop : ru ] that @xmath537 is the quotient field of @xmath538={\mathbb{z}}[{\mathbb{q}}[t^{\pm 1}]/(t^2 - 3t+1)^m]$ ] . 1 .
the alexander polynomial @xmath539 is @xmath540 .
2 . as @xmath149-modules , we have @xmath541/(t^2 - 3t+1)^2 $ ] , and @xmath35 has a unique nontrivial proper @xmath149-submodule , denoted @xmath60 .
the module @xmath60 is also a unique self - annihilating submodule of @xmath35 .
3 . let @xmath542 be associated to @xmath543 .
then as right @xmath544-modules , we have @xmath545 for some @xmath546 , and @xmath542 has a unique nontrivial proper @xmath544-submodule
. this unique submodule is generated by @xmath547 , and it is a unique self - annihilating submodule of @xmath548 .
4 . the knot @xmath0 does not have vanishing @xmath25invariants of order @xmath549 .
we will also need the following proposition and corollary later . recall that a ring is _ simple
_ if it has no nontrivial proper ( two - sided ) ideals . also recall that by proposition [ prop : ru ] we have @xmath550 $ ] .
note that @xmath537 is commutative .
however , @xmath544 is noncommutative : note that @xmath551/(t^2 - 3t+1)^m$ ] is a right @xmath552module , and that the semi - direct product @xmath553/(t^2 - 3t+1)^m\rtimes\gamma_0 $ ] , which can be embedded into @xmath554\rtimes { \gamma}_0 $ ] as a submodule , is defined by choosing the right multiplication ( see @xcite )
. therefore , for @xmath555 we have @xmath556 for the automorphism @xmath557 which is multiplication by @xmath558 on the right . also observe that @xmath559 is not the identity map for any @xmath49 . to show that @xmath544 is simple
, we should show that @xmath544 has no nontrivial proper ( two - sided ) ideal .
suppose to the contrary : let @xmath560 be a nontrivial proper ideal of @xmath544 . since @xmath561 is an automorphism ,
the ring @xmath544 is a pid . since @xmath544 is a pid , we have @xmath562 for some nonzero @xmath563 .
in fact , the element @xmath564 has least degree among the elements in @xmath560 and we have @xmath565 . since @xmath566 , one can see that @xmath564 is not a unit in @xmath544 .
therefore the degree of @xmath564 is greater than 0 , and hence we can write @xmath567 where @xmath568 , @xmath569 , and @xmath570 .
let @xmath571 be a nonzero element in @xmath537 .
since @xmath572 , there is @xmath573 such that @xmath574 .
therefore , we have @xmath575 and @xmath576 .
since neither @xmath577 nor @xmath578 is 0 , we obtain @xmath579 .
since @xmath571 is arbitrary , it follows that @xmath580 and @xmath581 is the identity map , which is a contradiction .
[ thm : example ] if @xmath0 is one of the knots in @xcite which are @xmath83solvable but not @xmath84solvable , then @xmath0 is not concordant to any knot whose alexander polynomial has degree less than or equal to two .
in particular , the knot @xmath0 has concordance genus 2 .
let @xmath1 be a knot whose alexander polynomial has degree less than or equal to two .
if @xmath1 has trivial alexander polynomial , then by freedman s work @xmath1 is slice .
since @xmath0 is not @xmath84solvable , it implies that @xmath0 is not concordant to @xmath1 . from now on we assume that @xmath1 has degree 2 alexander polynomial .
suppose that @xmath1 is concordant to @xmath0 .
since @xmath0 is @xmath83solvable , so is @xmath1 , and this implies that @xmath1 is algebraically slice ( see ( * ? ? ?
* remark 1.3.2 ) ) .
therefore , we may assume that the alexander polynomial of @xmath1 is @xmath582 for some integer @xmath583 . since @xmath0 is concordant to @xmath1 , the connected sum @xmath584 is slice .
for convenience , we re - label the knot @xmath7 by @xmath1 . note that this re - labeling does not change @xmath585 .
let @xmath67 .
now we have that @xmath66 is slice , and in particular it is @xmath84solvable . by theorem [ thm:4.6 ] , the knot @xmath66 has vanishing @xmath25invariants of order 2 .
we will show that @xmath66 has splitting self - annihilating submodules of order @xmath586 for @xmath0 . by theorem [ main ] , this will imply that @xmath0 has vanishing @xmath25invariants of order @xmath549 , which is a contradiction by lemma [ lem : cot knot property](4 ) , and it will complete the proof . henceforth , we use the notations given in the first paragraph of section [ sec : vanishing ] .
since @xmath587 and @xmath582 are coprime , by ( * ? ?
* theorem 3.1 ) the knot @xmath66 has vanishing self - annihilating submodules of order 0 for @xmath0 .
that is , if @xmath243 is a self - annihilating submodule of @xmath40 , then @xmath588 where @xmath60 and @xmath61 are self - annihilating submodules of @xmath38 and @xmath62 , respectively .
note that @xmath60 is the unique nontrivial proper submodule of @xmath38 by lemma [ lem : cot knot property](2 ) .
let @xmath243 be a self - annihilating submodule of @xmath40 and let @xmath450 .
let @xmath589/(t^2 - 3t+1)^m\rtimes\gamma_0 $ ] be a representations associated to @xmath146 and @xmath334 .
then @xmath590 by theorem [ thm : splitting h_1 ] .
let @xmath591 be a self - annihilating submodule of @xmath416 which is associated to @xmath336 .
let @xmath592 and @xmath593 .
we will show that @xmath451 , and that @xmath418 and @xmath419 are self - annihilating submodules of @xmath594 and @xmath595 , respectively .
this will imply that @xmath66 has vanishing self - annihilating submodules of order 1 for @xmath0 .
it is obvious that @xmath596 . to show @xmath597 , we assume lemma [ lem : splitting p_1 ]
below which asserts that @xmath418 is the unique nontrivial proper submodule of @xmath594 ( see lemma [ lem : cot knot property ] ( 3 ) ) , which is generated by @xmath547 .
note that lemma [ lem : splitting p_1 ] , combined with lemma [ lem : cot knot property ] ( 3 ) , also implies that @xmath598 .
let @xmath599 .
then for every @xmath600 , we have @xmath601 since @xmath602 and @xmath603 . since @xmath604
, we have @xmath605 for all
@xmath600 , and hence @xmath606 .
since @xmath598 , we have @xmath607 . by the definition of @xmath418
, it follows that @xmath608 . since @xmath609
, we also have @xmath610 .
this shows that if @xmath611 , then @xmath607 and @xmath612 .
therefore @xmath597 , and it follows that @xmath613 .
now we show that @xmath614 .
it is clear that @xmath615 since @xmath616 and @xmath617 .
let @xmath618 .
then , for every @xmath619 , we have @xmath620 , and @xmath621 since @xmath622
. therefore @xmath623 , and hence @xmath612 .
this shows @xmath624 .
recall that @xmath591 is a self - annihilating submodule of @xmath416 and @xmath625 . by theorem
[ thm : splitting h_1 ] and our choice of @xmath334 , we have @xmath626 .
therefore , we have @xmath627 as @xmath544-modules where @xmath550 $ ] and @xmath628 . by lemma [ lem : cot knot property](3 )
, we have @xmath629 . since @xmath625 , we have @xmath630 .
since @xmath603 , we have @xmath631 , and therefore @xmath632 . since @xmath633 , @xmath634 , and @xmath635
are nonsingular , we see that @xmath636 by corollary [ cor : cyclic ] , the module @xmath591 is cyclic , that is , @xmath637 for some @xmath638 as right @xmath544-modules . by ( [ eq : ex1 ] ) , we have @xmath639 and @xmath640 . since @xmath641 , for @xmath642 and @xmath643 , we have @xmath644 , and @xmath645 .
then , in @xmath544 there exists an element @xmath648 ( of degree 1 ) such that @xmath649 in @xmath594 and @xmath650 in @xmath595 .
therefore @xmath651 . since @xmath652 , we have @xmath653 .
since @xmath654 by ( [ eq : ex1 ] ) , one can see that @xmath655 .
therefore , we have @xmath656 .
therefore @xmath657 is a nontrivial proper submodule of @xmath594 . by lemma [ lem : cot knot property](3 )
, the module @xmath594 has a unique nontrivial proper submodule @xmath658 , and therefore in @xmath594 we have @xmath659 for some nonzero @xmath555 .
since @xmath660 , @xmath661 , and @xmath603 , it follows that @xmath662 and hence @xmath663 .
this implies that @xmath664 . by lemma [ lem : cot knot property](3 ) , the submodule @xmath665 is self - annihilating .
therefore , we have @xmath666 . since @xmath646 and @xmath667
, it is a contradiction .
since @xmath594 has a unique nontrivial proper submodule @xmath665 , we have @xmath670 .
therefore we have @xmath671 for some @xmath555 .
since @xmath672 , we have @xmath673 . since @xmath660 and @xmath674
, we have @xmath675 therefore @xmath676 , which is a contradiction . since @xmath678
, we have @xmath679 , and therefore there exists @xmath680 such that @xmath681 in @xmath595
. therefore , there exists @xmath682 such that @xmath683 . since @xmath684
, this implies that @xmath685 . by ( [ eq : ex1 ] )
, one sees that @xmath686
. therefore @xmath687 .
case ( 2 ) implies that @xmath688 , since @xmath689 where @xmath690 is considered as an @xmath544-submodule of @xmath595 .
therefore @xmath691 .
since @xmath594 has a unique nontrivial proper submodule @xmath658 , we may assume that in @xmath594 either @xmath692 for some nonzero @xmath555 , or @xmath693 for some @xmath694 with @xmath695 and @xmath696 .
suppose @xmath702 in @xmath594 .
then @xmath703 since @xmath594 has a unique nontrivial proper submodule @xmath658 .
therefore , in @xmath594 we have @xmath704 for some @xmath705 , and we obtain @xmath706 .
this shows @xmath707 , and hence @xmath708 .
this implies that @xmath709 , which is a contradiction .
therefore , @xmath710 in @xmath594 .
since @xmath711 for all @xmath712 , in @xmath594 we have @xmath713 note that since we are viewing @xmath594 as a right @xmath544-module and @xmath714 in general , we can not cancel @xmath78 in the above equation .
recall that @xmath537 is commutative . for @xmath715 ,
we write @xmath716 for @xmath717 for convenience .
now it follows that in @xmath718 we have @xmath719 therefore , in @xmath544 we have @xmath720 by ( * ? ? ?
* lemma 6.3 ) , it follows that in @xmath537 we have @xmath721 since @xmath695 and @xmath561 is an automorphism , it is a contradiction .
now suppose that @xmath722 for some @xmath723 with @xmath695 and @xmath696 .
since @xmath724 , as in the previous case of @xmath692 , in @xmath594 we obtain that @xmath725 this implies that in @xmath594 we have @xmath726 and @xmath727 since @xmath728 in @xmath594 , we have @xmath729 and @xmath730 and @xmath731 again by ( * ? ? ?
* lemma 6.3 ) , one can deduce that in @xmath537 we have @xmath732 and @xmath733 therefore , in @xmath537 we have @xmath734 and @xmath735 .
since @xmath736 , we have @xmath737 which is a contradiction . since @xmath738 , there exists @xmath555 such that @xmath739 in @xmath595 . since @xmath740 and @xmath741
, it follows that @xmath742 in @xmath594 and @xmath743 . by ( [ eq : ex1 ] )
, we have @xmath744 .
since @xmath594 has a unique nontrivial proper submodule @xmath658 , we obtain that @xmath745 for some @xmath705 .
therefore @xmath706 , and hence @xmath746 .
it follows that @xmath747 . by ( [ eq : ex1 ] )
, we have @xmath748 .
therefore , @xmath418 is a nontrivial proper submodule of @xmath594 , and hence @xmath749 . we wish to thank jae choon cha , tim cochran , chuck livingston , and kent orr for helpful conversations .
the first author was supported by the national research foundation of korea(nrf ) grant funded by the korea government(mest ) ( no .
20090074487 ) .
the second author was supported by basic science research program through the national research foundation of korea ( nrf ) grant funded by the ministry of education , science and technology ( no . 2012r1a1a001747 and no .
20120000341 ) .
this paper was written as part of konkuk university s research support program for its faculty on sabbatical leave in 2013 .
|
we give a sufficient condition under which vanishing property of cochran orr teichner knot concordance obstructions splits under connected sum .
the condition is described in terms of self - annihilating submodules with respect to higher - order blanchfield linking forms .
this extends results of levine and the authors on distinguishing knots with coprime alexander polynomials up to concordance . as an application
, we show that the knots constructed by cochran , orr and teichner as the first examples of nonslice knots with vanishing casson gordon invariants are not concordant to any knot of genus 1 .
|
0.2cm**harold : * * for quite sometime now , you have been talking about ` gravity being an emergent phenomenon ' and a ` thermodynamic perspective on gravity ' .
this is quite different from the conventional point of view in which gravity is a fundamental interaction and spacetime thermodynamics of , say , black holes is a particular result which can be derived in a specific context . honestly , while i find your papers fascinating i am not clear about the broad picture you are trying to convey .
maybe you could begin by clarifying what this is all about , before we plunge into the details ? what is the roadmap , so to speak ?
0.2cm**me : * * to begin with , i will show you that the equations motion describing gravity in _ any diffeomorphism invariant theory _ can be given @xcite a suggestive thermodynamic re - interpretation ( sections [ sec : lro ] , [ sec : reinterpret ] ) .
second , taking a cue from this , i can formulate a variational principle for a suitably defined entropy functional involving both gravity and matter which will lead to the field equations of gravity @xcite without varying the metric tensor as a dynamical variable ( section [ sec : eqnnewvar ] ) .
0.2cm**harold : * * suppose i have an action for gravity plus matter ( in @xmath0 dimensions ) @xmath1\ ] ] where @xmath2 is any scalar built from metric and curvature and @xmath3 is the matter lagrangian depending on the metric and some matter variables @xmath4 .
( i will assume @xmath5 does not involve derivatives of curvature tensor , to simplify the discussion . )
if i vary @xmath6 in the action i will get some equations of motion ( see e.g. refs .
@xcite ) , say , @xmath7 where @xmath8 is @xmath9 now , you are telling me that ( i ) you can give a thermodynamic interpretation to the equation @xmath7 just because it comes from a scalar lagrangian and ( ii ) you can also derive it from an entropy maximisation principle .
i admit it is fascinating . but why should i take this approach as more fundamental , conceptually , than the good old way of just varying the total lagrangian @xmath10 and getting @xmath7 ?
why is it more than a curiosity ?
0.2cm**me : * * that brings me to the third aspect of the formulation which i will discuss towards the end ( section [ sec : comaparison ] ) . in my approach , i can provide a natural explanation to several puzzling aspects of gravity and horizon thermodynamics all of which have to be thought of as mere algebraic accidents in the conventional approach you mentioned .
let me give an analogy . in newtonian gravity , the fact that inertial mass is equal to the gravitational mass is an algebraic accident without any fundamental explanation .
but in a geometrical theory of gravity based on principle of equivalence , this fact finds a natural explanation .
similarly , i think we can make progress by identifying key facts which have no explanation in the conventional approach and providing them a natural explanation from a different perspective .
you will also see that this approach connects up several pieces of conventional theory in an elegant manner .
0.2cm**harold : * * your ideas also seem to be quite different from other works which describe gravity as an emergent phenomena @xcite .
can you explain _ your _ motivation ?
0.2cm**me : * * yes . the original inspiration for my work , as for many others , comes from the old idea of sakharov @xcite which attempted to describe spacetime dynamics as akin to the theory of elasticity .
there are two crucial differences between my approach and many other ones .
to begin with , i realized that the thermodynamic description transcends einstein s general relativity and can incorporate a much wider class of theories this was first pointed out in ref .
@xcite and elaborated in several of my papers while many other approaches concentrated on just einstein s theory . in fact , many other approaches use techniques strongly linked to einstein s theory like for example , raychaudhuri equation to study rate of change of horizon area , which is difficult to generalize to theories in which the horizon entropy is _ not _ proportional to horizon area .
i use more general techniques .
second , i work at the level of action principle and its symmetries to a large extent so i have a handle on the off - shell structure of the theory ; in fact , much of the thermodynamic interpretation in my approach is closely linked to the structure of action functional ( like e.g. , the existence of surface term in action , holographic nature etc . ) for gravitational theories .
this link is central to me while it is not taken into account in any other approach .
0.2cm**harold : * * so essentially you are claiming that the thermodynamics of horizons is more central than the dynamics of the gravitational field while the conventional view is probably the other way around . why do you stress the thermal aspects of horizons so much ?
can you give a motivation ?
0.2cm**me : * * because thermal phenomenon is a window to microstructure ! let me explain .
we know that the continuum description of a fluid , say , in terms of a set of dynamical variables like density @xmath11 , velocity @xmath12 , etc . has a life of its own . at the same time , we also know that these dynamical variables and the description have no validity at a fundamental level where the matter is discrete .
but one can actually _ guess _ the existence of microstructure without using any experimental proof for the molecular nature of the fluid , just from the fact that the fluid or a gas exhibits _ thermal phenomena _ involving temperature and transfer of heat energy . if the fluid is treated as a continuum and
is described by @xmath13 , @xmath14 etc .
, all the way down , then it is _ not _ possible to explain the thermal phenomena in a natural manner .
as first stressed by boltzmann , the heat content of a fluid arises due to random motion of discrete microscopic structures which _ must _ exist in the fluid .
these new degrees of freedom which we now know are related to the actual molecules make the fluid capable of storing energy internally and exchanging it with surroundings .
so , given an apparently continuum phenomenon which exhibits temperature , boltzmann could infer the existence of underlying discrete degrees of freedom .
0.2cm**harold : * * i agree . but what does it lead to in the present context ?
0.2cm**me : * * the paradigm is : _ if you can heat it , it has microstructure ! _ and you can heat up spacetimes by collapsing matter or even by just accelerating @xcite . the horizons which arise in general relativity are endowed with temperatures @xcite which shows that , at least in this context , some microscopic degrees of freedom are coming into play .
so a thermodynamic description that links the standard description of gravity with the statistical mechanics of as yet unknown microscopic degrees of freedom must exist .
it is in this sense that i consider gravity to be emergent .
boltzmann s insight about the thermal behaviour has two other attractive features which are useful in our context .
first , while the existence of the discrete degrees of freedom is vital in such an approach , the exact nature of the degrees of freedom is largely irrelevant .
for example , whether we are dealing with argon molecules or helium molecules is largely irrelevant in the formulation of gas laws and such differences can be taken care of in terms of a few well - chosen numbers ( like , e.g. , the specific heat ) .
this suggests that such a description will have certain amount of robustness and independence as regards the precise nature of microscopic degrees of freedom .
second , the entropy of the system arises due to our ignoring the microscopic degrees of freedom . turning this around
, one can expect the form of entropy functional to encode the essential aspects of microscopic degrees of freedom , even if we do not know what they are .
if we can arrive at the appropriate form of entropy functional , in terms of some effective degrees of freedom , then we can expect it to provide the correct description
. will have universal applicability as long as effects of relativity or quantum theory are incorporated in the definition of @xmath15 appropriately . ]
0.2cm**harold : * * but most people working in quantum gravity will agree that there is some fundamental microstructure to spacetime ( `` atoms of spacetime '' ) and the description of spacetime by metric is an approximate long distance description .
so why are you making a big deal ?
i do nt see anything novel here .
0.2cm**me : * * i will go farther than just saying there is microstructure and show you how to actually use the thermodynamic concepts to provide an emergent description of gravity which no one else has attempted .
if you think of the full theory of quantum gravity as analogous to statistical mechanics then i will provide the thermodynamic description of the same system .
as you know , thermodynamics was developed and used effectively decades before we knew anything about the molecular structure of matter or its statistical mechanics .
similarly , even without knowing the microstructure of spacetime or the full quantum theory of gravity , we can make lot of progress with the thermodynamic description of spacetime .
the horizon thermodynamics , i will claim , provides @xcite valuable insights about the nature of gravity totally independent of what `` the atoms of spacetime '' may be .
it is somewhat like being able to describe or work with gases or steam engines without knowing anything about the molecular structure of the gas or steam .
0.2cm**harold : * * all right .
i hope all these will become clearer as we go along .
maybe i can suggest we plunge head - long into how _
you _ would like to describe gravity .
then i can raise the issues as we proceed .
0.2cm**me : * * the overall structure of my approach is shown in fig .
[ fig : synthesis ] . as you can see , i begin with the principle of equivalence which allows you to draw three key consequences .
first , it tells you that in the long wavelength limit gravity has @xcite a geometrical description in terms of the metric tensor @xmath16 and the effect of gravity on matter can be understood by using the laws of special relativity in the local inertial frames .
second , by writing maxwell s equations in curved spacetime using minimal coupling , say , i can convince myself that the light cone structure of the spacetime and hence the causal structure will , in general , be affected by the gravitational field .
0.2cm**harold : * * well , that is one possible way of interpreting principle of equivalence though people might have other views .
but once you have told me what you are assuming , viz .
, `` gravity = geometry '' and `` light cones are affected by gravity '' , we can proceed further .
0.2cm**me : * * yes . my aim here will be not to nitpick over definitions but develop the physics in a consistent manner . in that spirit
, i would draw one more conclusion from the fact that gravity can be described using a metric tensor .
in flat spacetime , we can choose a special coordinate system with the global metric being @xmath17 ; so if someone tells you that the metric is given by @xmath18 then you can always attribute the part @xmath19 to the choice of non - inertial coordinates .
we can not do this in a curved spacetime .
so it no longer makes sense to ask how much of @xmath16 " is due to our using non - inertial coordinates and how much " is due to genuine gravity .
different observers in different states of motion might use different coordinates leading to different sets of ten functions for @xmath18 . because we have no absolute reference
metric it follows that no coordinate system or observer is special .
the laws of physics should not select out any special class of observers .
0.2cm**harold : * * this smacks of principle of general covariance but essentially you are arguing @xcite for democracy of all observers , which i grant you . given all the philosophical controversies as to what `
general covariance ' means , i agree this is a safer procedure .
what next ?
0.2cm**me : * * given the fact that all observers are equal and that light cones are affected by gravity , it follows that there will exist observers who do not have access to part of the spacetime because of the existence of horizons they perceive .
this is a direct consequence of the fact that metric determines the paths of light rays and hence the causal structure .
the classic example is the rindler horizon in flat spacetime which is as effective in blocking information with respect to an accelerated observer as the schwarzschild horizon at @xmath20 is for an observer at @xmath21 .
0.2cm**harold : * * not so fast ; i have several problems here .
first , the conventional view is that black hole horizons are `` real horizons '' while rindler horizons are sort of fraudulent ; you seem to club them together .
second , you seem to link horizons to observers rather than treat them as well defined , geometrical , causal features of a spacetime .
0.2cm**me : * * you are quite right .
i treat all horizons at equal footing and claim that for _ my _ purpose all horizons are observer dependent .
this is because , the key property of horizons which i am concentrating here is that it can block information . in that sense ,
the rindler horizon does not block information for an inertial observer _ just as _ the schwarzschild horizon does _ not _ block information for someone plunging into the black hole .
so , for my purpose , there is no need to make artificial distinctions between a black hole horizon and a rindler horizon .
the state of motion of the observer is crucial in deciding the physical effects of a horizon _ in all cases_. it is , of course , true that one can give a geometric interpretation to , say , the black hole event horizon .
i am not denying that .
but that fact , as you will see , is quite irrelevant to the development of my approach .
0.2cm**harold : * * i see that you not only demand democracy of observers but also democracy of horizons !
you do nt think , for example , that black hole horizons are anything special .
0.2cm**me : * * yes . i do believe in the democracy of horizons , as you put it . the attempts to provide a quantum gravitational interpretation of black holes , their entropy etc . using _ very special _ approaches which are incapable of handling other horizons like the issues in de sitter @xcite ,
let alone rindler are interesting in a limited sort of way but may not get us anywhere ultimately . 0.2cm**harold : * * i have another problem .
you really havent characterized what exactly you mean by a horizon for an observer .
of course , you can not use any on - shell constructs since you are still developing your approach towards field equations .
there are horizons and horizons in the literature event , apparent , causal ....
0.2cm**me : * * i will try to make clear what i _ need _ without again going into all sorts of definitions @xcite .
choose any event @xmath22 and introduce a local inertial frame ( lif ) around it with riemann normal coordinates @xmath23 such that @xmath22 has the coordinates @xmath24 in the lif .
let @xmath25 be a future directed null vector at @xmath22 and we align the coordinates of lif such that it lies in the @xmath26 plane at @xmath22 .
next transform from the lif to a local rindler frame ( lrf ) coordinates @xmath27 by accelerating along the x - axis with an acceleration @xmath28 by the usual transformation .
the metric near the origin now reduces to the form @xmath29 where @xmath30 and ( @xmath31 ) or ( @xmath32 ) are the coordinates of lrf ( both these forms are useful in our discussion ) .
let @xmath33 be the approximate killing vector corresponding to translation in the rindler time such that the vanishing of @xmath34 characterizes the location of the local horizon @xmath35 in lrf . as usual
, we shall do all the computation on a timelike surface infinitesimally away from @xmath35 with @xmath36 constant , usually called a `` stretched horizon '' .
( it can be defined more formally using the orbits of @xmath33 and the plane orthogonal to the acceleration vector @xmath37 . )
let the timelike unit normal to the stretched horizon be @xmath38 .
this lrf ( with metric in ) and its local horizon @xmath35 will exist within a region of size @xmath39 ( where @xmath40 is a typical component of curvature tensor of the background spacetime ) as long as @xmath41 .
this condition can always be satisfied by taking a sufficiently large @xmath28 .
this procedure introduces a class of uniformly accelerated observers who will perceive the null surface @xmath42 as the local rindler horizon @xmath35 .
this is shown in fig .
[ fig : ray1 ] .
0.2cm**harold : * * i am with you so far
. essentially you are using the fact that you have two length scales in the problem at any event .
first is the length scale @xmath43 associated with the curvature components of the background metric over which you have no control ; second is the length scale @xmath44 associated with the accelerated trajectory which you can choose as you please .
so you can always ensure that @xmath41 .
in fact , i can see this clearly in the euclidean sector in which the horizon maps to the origin ( see fig.[fig : ray3 ] ) .
the locally flat frame in the euclidean sector will exist in a region of radius @xmath43 while the trajectory of a uniformly accelerated observer will be a circle of radius @xmath44 .
you can always keep the latter inside the former .
the metric in is just the metric of the locally flat region in polar coordinates .
0.2cm**me : * * yes . in fact , i can choose a trajectory @xmath45 such that its acceleration @xmath46 ( where @xmath47 is the time - like four velocity ) satisfies the condition @xmath48 . in a suitably chosen lif this trajectory will reduce to the standard hyperbola of a uniformly accelerated observer .
it is using these lrfs that i define my horizons around any event .
further the local temperature on the stretched horizon will be @xmath49 so that @xmath50 with @xmath51 .
0.2cm**harold : * * ha !
the classical gr is fine but your ` horizon ' is just a patch of null surface .
can you actually prove that local rindler observers will perceive a temperature proportional to acceleration ?
the usual proofs of unruh effect are @xcite horribly global .
0.2cm**me : * * recall that everything we do is in a local region with @xmath41 .
now if you have an accelerated detector with time - dependent , variable acceleration , say , then you will reproduce the standard unruh effect _ approximately _ to the necessary order of accuracy .
this should be intuitively obvious but can be demonstrated @xcite .
of course in the euclidean sector the rindler observer s trajectory is a circle of radius @xmath44 which can be made arbitrarily close to the origin .
suppose the observer s trajectory has the usual form @xmath52 which is maintained for a time interval @xmath53 .
then , the trajectory will complete a full circle _ in the euclidean sector _
irrespective of what happens later ! when we work in the limit of @xmath54 , this becomes arbitrarily local in both space _ and _ time @xcite .
i am sure all these can be made more rigorous but this is the essential reason behind the local ideas working .
i also want to stress that once i finally reach my goal ( of deriving the gravitational field equations from an entropy principle in section [ sec : eqnnewvar ] ) all these become irrelevant ; they are essentially part of ` motivation ' .
so your possible misgivings regarding some of these details will not affect the final result .
0.2cm**harold : * * ok , so you have local rindler observers crawling all over the spacetime with their local horizons .
what next ?
0.2cm**me : * * it is now easy to see that all horizons must have entropy vis - a - vis the observers who perceive the horizons .
if they do not , an observer can pour some hot tea with entropy across the horizon _ la wheeler _ @xcite , thereby violating the second law of thermodynamics in the region accessible to her and her friends who perceive the horizon @xmath35 .
such a violation of second law of thermodynamics can be avoided only if we _
demand _ that horizon should have an entropy which should increase when energy flows across it .
if energy @xmath55 flows across a hot horizon with temperature @xmath56 then @xmath57 should be the change in the entropy of the horizon .
we therefore conclude that all null surfaces which could locally act as one - way membranes should have an ( observer dependent ) entropy associated with them .
0.2cm**harold : * * hold on .
i understand from the reference you cite @xcite that such a thought experiment might have had something to do with the initial realization of a black hole entropy which is proportional to the area @xcite .
but i am not sure how to interpret it precisely . for one thing
, matter disappears into the horizon only after infinite time as perceived by the outside observer , even when you try to pour real tea in to real black hole .
so what is all this talk about `` loss '' of entropy ?
0.2cm**me : * * i do nt think this is a real objection though one often comes across this confusion .
note that , by the same argument , no black hole can ever form in finite time anywhere in the universe and we should not be talking about any black hole physics .
i believe this issue is well settled in chapter 33 of ref .
i recommend you read it !
if you really push me hard , i can wiggle out with the following argument .
it does not take much time ( certainly not infinite time ! ) for a cup of tea to reach a radial distance a few planck lengths away from the horizon @xmath20 .
we have considerable evidence of very different nature to suggest planck length acts as lower bound to the length scales that can be operationally defined and that no measurements can be ultra sharp at planck scales @xcite .
so one can not really talk about the location of the event horizon ignoring fluctuations of this order .
so , from the point of view of sensible physics , i only need to get the cup of tea up to @xmath58 to talk about entropy loss .
0.2cm**harold : * * you also seem to have quietly made entropy an observer dependent quantity .
_ this is pretty drastic _ and i need to understand it .
suppose , in a region around an event @xmath22 , there is some matter which is producing the curvature .
i would have normally thought that this matter say some amount of hot fluid has certain amount of entropy which is independent of who is measuring it .
but you are claiming that an inertial observer and a rindler observer will attribute different amounts of entropy to this matter .
is that correct ?
0.2cm**me : * * that s correct and may be i should write a paper explaining this @xcite but this really need not come as a surprise ( also see ref.@xcite ) . we know that an inertial observer will attribute zero temperature and zero entropy to the inertial vacuum . but a rindler observer will attribute a finite temperature and non - zero ( formally divergent ) entropy to the same vacuum state .
so entropy is indeed an observer dependent concept .
when you do quantum field theory in curved spacetime , it is not only that particles become an observer dependent notion so do the temperature and entropy .
this notion can be made more precise as follows : consider an excited state of a quantum field with energy @xmath59 above the ground state in an inertial spacetime .
when you integrate out the unobservable modes for the rindler observer , you will get a density matrix @xmath60 for this state and the corresponding entropy will be @xmath61 .
the inertial vacuum state has the density matrix @xmath62 and the entropy @xmath63 .
the difference @xmath64 is finite and represents the entropy attributed to this state by the rindler observer .
( this is finite though @xmath65 and @xmath66 can be divergent . ) in the limit of @xmath67 , in which we are working , we can actually compute it and show that @xmath68 to see this , note that if we write @xmath69 , then in the limit of @xmath67 we can concentrate on states for which @xmath70
. then we have @xmath71 where we have used the facts tr @xmath72 and @xmath73 where @xmath74 is the hamiltonian for the system in the rindler frame .
the last line defines the @xmath59 in terms of the difference in the expectation values of the hamiltonian in the two states .
( there are some subtleties in this derivation , especially regarding the assumption @xmath70 , but i will not get into it here @xcite . ) this is the amount of entropy a rindler observer would claim she has lost when the matter disappears into the horizon .
0.2cm**harold : * * that is very curious .
i would have thought that the expression for entropy of matter should consist of its energy @xmath59 and _ its own _
temperature @xmath75 rather than the horizon temperature .
it looks like that the matter somehow equilibrates to the horizon temperature so that @xmath76 gives the relevant entropy .
0.2cm**me : * * yes .
this can be explicitly proved , for example , for the one particle state @xcite and here is a possible interpretation .
you should think of horizon as a system with some internal degrees of freedom and temperature @xmath56 _ as far as rindler observer is concerned_. so when you add an energy @xmath59 to it , the entropy change is @xmath77 .
all these are not _ new _ mysteries but only the manifestation of the old mystery , viz .
, a rindler observer attributes a non - zero temperature to inertial vacuum .
this temperature influences every other thermodynamic variable .
i will come back to this point later on because it is quite important .
0.2cm**harold : * * ok . let us proceed .
i also see where your insistence of democracy of observers comes in .
you want to demand that the local rindler observer has a right to expect the standard laws of physics to hold as much as any other observer , horizons notwithstanding .
0.2cm**me : * * i am glad you brought this up .
this was first pointed out in ref .
@xcite in which we assert that _ all observers have a right to describe physics using an effective theory based only on the variables she can access .
_ in the study of particle physics models , this concept forms the cornerstone of the renormalization group theory .
to describe particle interactions at 10 gev in the lab , we usually do not need to know what happens at @xmath78 gev in theories which have predictive power . in the absence of such a principle ,
very high energy phenomena ( which are unknown from direct experiments in the lab ) will affect the low energy phenomena which we are attempting to study . in the context of a theory involving a nontrivial metric of spacetime , we need a similar principle to handle the fact that different observers will have access to different regions of a general spacetime .
if a class of observers perceive a horizon , they should still be able to do physics using only the variables accessible to them without having to know what happens on the other side of the horizon .
this , in turn , implies that there should exist a mechanism which will encode the information in the region @xmath79 which is inaccessible to a particular observer at the boundary @xmath80 of that region .
keep this in mind because i will show you later where this fits in with the holographic nature of action functionals .
0.2cm**harold : * * fine , we will get back to it . to get on with the story
, you need to formulate some kind of entropy balance when matter flows across a local horizon
. how do you propose to do it ?
0.2cm**me : * * around any event in any spacetime we now have a local inertial frame and by boosting along one of the axes with an acceleration @xmath28 we have introduced a local rindler observer who perceives a horizon with temperature proportional to @xmath28 .
she will attribute a loss of entropy @xmath81 when matter with an amount of energy @xmath59 gets close to the horizon ( within a few planck lengths , say ) . if @xmath33 is the approximate , killing vector corresponding to translations in rindler time ,
the appropriate energy - momentum density is @xmath82 .
( it is the integral of @xmath83 that gives the rindler hamiltonian @xmath74 , which leads to evolution in rindler time @xmath84 and appears in the thermal density matrix @xmath85 . )
the energy flux through a patch of stretched horizon with normal @xmath38 will be @xmath86 and the associated entropy flux will be @xmath87 where @xmath88 is the local temperature with @xmath89 being the standard lapse function giving the redshift factor .
( in conformity with , i am using the horizon temperature and not the matter temperature ) . this entropy flux manifests as the entropy change of the locally perceived horizon .
for all these to hold locally at every event there _ must exist a spacetime entropy current _ @xmath90 , built out of metric and its derivatives , such that @xmath91 gives the corresponding gravitational entropy flux .
so we expect the relation @xmath92 to hold at all events with some @xmath93 , once we introduce a local killing vector @xmath33 and a local temperature giving @xmath94 .
further @xmath93 must be conserved since we do not expect irreversible entropy production in the spacetime .
0.2cm**harold : * * this sounds strange!. why should there exists a conserved current @xmath93 , built from geometrical variables , at every event in some arbitrary spacetime , which will conveniently give you the entropy balance you require ?
0.2cm**me : * * it is actually not all that strange
! remember that we got into all these because of the democracy of the observers which , in turn , implies general covariance .
the mathematical content of general covariance is captured by the diffeomorphism invariance of whatever theory which is going to ultimately determine the dynamics of the spacetime . because the diffeomorphism invariance of the theory forced us to treat all observers on equal footing
, the diffeomorphism invariance must also provide us with the conserved current @xmath93 .
and indeed it does , in the form of the noether current @xcite .
let me explain .
consider a theory of gravity , obtained from a generally covariant action principle involving a gravitational lagrangian @xmath95 which is a scalar made from metric and curvature tensor .
the total lagrangian is the sum of @xmath5 and the matter lagrangian @xmath96 .
the variation of the gravitational lagrangian density generically leads to a surface term and hence can be expressed in the form , @xmath97 under suitable boundary conditions the theory will lead to the field equation @xmath98 where @xmath8 is given by and @xmath99 is defined through the usual relation @xmath100 .
we also know that , for any lagrangian @xmath5 , the functional derivative @xmath8 satisfies the generalized off - shell bianchi identity : @xmath101 consider now the variations in @xmath102 which arise through the diffeomorphism @xmath103 . in this case ,
@xmath104 , with @xmath105 . substituting these in and using @xmath106
, we obtain the conservation law @xmath107 , for the current , @xmath108 where @xmath109 represents the boundary term which arises for the specific variation of the metric in the form @xmath110 ) .
it is also convenient to introduce the antisymmetric tensor @xmath111 by @xmath112 . using the known expression for @xmath113 in ,
it is possible to write an explicit expression for the current @xmath93 for any diffeomorphism invariant theory . for the general class of theories we are considering , the @xmath111 and @xmath93
can be expressed @xcite in the form @xmath114 @xmath115 where @xmath116 .
these expressions simplify significantly at any event @xmath22 where @xmath33 behaves like an ( approximate ) killing vector and satisfies the conditions @xmath117 ( which a true killing vector will satisfy everywhere )
. then one can easily prove that @xmath118 at the event @xmath22 ; the expression for noether current simplifies considerably and is given by @xmath119 0.2cm**harold : * * ok . so you now have a conserved current @xmath93 and entropy current of matter .
what do you do now ?
0.2cm**me : * * recall that i argued , on very general grounds , that the relation in _ must _ hold at all events .
remarkably enough , _ the gravitational field equations of any diffeomorphism invariant theory implies that this relation does hold ! _ to see this , let us now consider the form of @xmath120 at any event @xmath121 around which we have introduced the notion of a local rindler horizon with @xmath33 being the approximate killing vector associated with the rindler time translation invariance that satisfies two conditions in at @xmath121 .
let @xmath38 be the spacelike unit normal to the stretched horizon @xmath122 , pointing in the direction of increasing @xmath123 .
we know that as @xmath124 and the stretched horizon approaches the local horizon and @xmath125 approaches @xmath126 . with this background ,
we compute @xmath93 for the @xmath33 introduced above in the neighborhood of @xmath22 . since it is an approximate killing vector , satisfying it follows that @xmath127 giving the current to be @xmath128 .
the product @xmath129 for the vector @xmath130 , which satisfies @xmath131 on the stretched horizon , becomes quite simple : @xmath132 .
this equation is valid around the local patch in which @xmath33 is the approximate killing vector .
the quantity @xmath133 ( in this limit ) is what we interpret as the local entropy flux density . on using the field equations @xmath7 , we immediately get @xmath134 which is exactly .
this tells you that the validity of field equations in any diffeomorphism invariant theory has a local , thermodynamic , interpret ion . in the limit of @xmath124 ,
this gives a _
finite _ result , @xmath135 as it should .
further , in this limit , @xmath126 goes to @xmath136 where @xmath137 is the affine parameter associated with the null vector @xmath25 we started with and all the reference to lrf goes away .
it is clear that the properties of lrf are relevant conceptually to define the intermediate notions ( local killing vector , horizon temperature .... ) but the essential result is independent of these notions . just as we introduce local inertial frame to decide how gravity couples to matter , we use local rindler frames to interpret the physical content of the field equations .
0.2cm**harold : * * that is cute !
i also see why you can afford to be a bit cavalier about the lrf etc ; ultimately , your interpretation is local at each event . the noether current you use , of course , is the same that appears in the definition of wald entropy @xcite .
but in the latter , it is used in an integral form while your approach seems to be completely local .
0.2cm**me : * * this is true and i think the local approach is crucial for proper interpretation . integrals over surfaces would require all sort of special assumptions for everything to work out in an arbitrary spacetime .
this is why i work in a local region around an arbitrary event with lif , lrf etc .
with l in everything . also note that the original definition of wald entropy is an on - shell construct and requires you to evaluate an integral on a solution .
the noether current itself is an off - shell construct and that is what i need .
incidentally , the noether current relation can also be used to provide an alternative interpretation of the entropy balance along the following lines . a local rindler observer , moving along the orbits of the killing vector field @xmath33 with four velocity @xmath138 , will associate an energy @xmath139 with a proper volume @xmath140 .
if this energy gets transfered across the horizon , the corresponding entropy transfer will be @xmath141 where @xmath142 is the local ( redshifted ) temperature of the horizon and @xmath89 is the lapse function . since @xmath143 , we find that @xmath144 as for gravitational entropy , since @xmath145 is the noether charge density , @xmath146 can be interpreted as the entropy associated with a volume @xmath140 as measured by an observer with four - velocity @xmath147 . for observers moving on the orbits of the killing vector @xmath33 with @xmath138 we get @xmath148\ , dv_{\rm prop}\ ] ] as one approaches the horizon , @xmath149
making the second term vanish and we get @xmath150 in the same limit @xmath151 will become proportional to the original null vector @xmath152 we started with .
so this equation can be again thought of as an entropy balance condition .
0.2cm**harold : * * so instead of thinking of field equations of gravity as @xmath7 , you want us to think of them as @xmath153k_ak_b = 0 \label{myeqn}\ ] ] for all null vectors @xmath25 .
this is equivalent to @xmath154 with some constant @xmath155 .
( i see that the constancy of @xmath155 follows from the conditions @xmath156 . ) interpreting @xmath157 as some kind of gravitational entropy density and @xmath158 as matter entropy in the local rindler frame , you are are providing a purely thermodynamical interpretation of the field equations of any diffeomorphism invariant theory of gravity .
0.2cm**me : * * yes . but note that is _ not _ quite the same as the standard equation @xmath159 because has an extra symmetry which standard gravitational field equations do not have : this equation is invariant under the shift @xmath160 with some constant @xmath161 .
( this symmetry has important implications for cosmological constant problem which we will discuss later . ) while the properties of lrf are relevant conceptually to define the intermediate notions ( local killing vector , horizon temperature .... ) , the essential result is independent of these notions .
0.2cm**harold : * * fine .
i like the fact that _ just as we introduce local inertial frames to decide how gravity couples to matter , we use local rindler frames to interpret the physical content of the field equations . _
but you only needed the part of @xmath93 given by @xmath162 for your analysis , right ?
the other two terms in are not needed at all .
so may be you do nt have to use all of noether current .
0.2cm**me : * * this is quite true .
in fact one can give @xmath162 an interesting interpretation .
suppose there are some microscopic degrees of freedom in spacetime , just as there are atoms in a solid .
if you make the solid undergo an elastic deformation @xmath163 , the physics can be formulated in terms of the displacement field @xmath164 and one can ask how thermodynamic potentials like entropy change under such displacement . similarly , in the case of spacetime , we should think of @xmath165 as the change in the gravitational entropy under the ` deformation ' of the spacetime @xmath166 as measured by the rindler observer with velocity @xmath147 .
one can show that this interpretation is consistent with all that we know about horizon thermodynamics .
so the left hand side of gravitational field equation @xmath167 actually gives the response of the spacetime entropy to the deformations .
0.2cm**harold : * * it certainly matches with the previous results .
since @xmath168 , you will get the entropy density to be proportional to @xmath157 on the horizon .
does it make sense ?
0.2cm**me : * * as i will show you soon , it makes lot of sense !
0.2cm**harold : * * but ca nt you now reverse the argument and claim that you can derive the field equations of the theory from the purely thermodynamic point of view of the entropy balance ?
0.2cm**me : * * that would be lovely and very tempting but i do nt think so .
such a ` reverse engineering ' faces some conceptual hurdles ; the mathematics will go through trivially but not the logic @xcite .
let me clarify the issues involved .
the key point is the following : if we have a justification for interpreting the expression @xmath91 as entropy current , _ independent of the field equations _ , then _ and only then _ can we invert the logic and obtain the field equations from the thermodynamic identity .
however , in the absence of field equations @xmath93 is just a noether current .
it can be interpreted as entropy current _ if and only if _
field equations are assumed to hold ; it is in this on - shell context that wald @xcite showed that it is entropy .
so we have no _ independent _ justification for demanding @xmath169 should be equal to matter entropy flux . until we come up with such a justification without using field equations we can prove that field equations imply local entropy balance at local horizons " but not local entropy balance at local horizons imply field equations " .
the issue at stake is not mathematics but logic . as a simple example , consider the noether current in einstein s theory for a killing vector @xmath33 , which is proportional to @xmath170 .
no one would have thought of this expression as entropy density independent of field equations .
it is only by studying physical processes involving black holes , say , _ and using field equations _ that one can give such a meaning .
0.2cm**harold : * * ok . i have one more worry . at this stage
, you have not chosen any specific theory of gravity at all , right ?
so this thermodynamic entropy balance seems to be very general and some people might even say it is _
too general_. what is your take on this ?
0.2cm**me : * * it is true that at this stage i have not specified what kind of theory of gravity we are dealing with . the field equation whatever the theory may be , as long as it obeys principle of equivalence and diffeomorphism invariance always has an interpretation in terms of local entropy balance ( the idea also works when @xmath171 depends on the derivatives of the curvature tensor but i will not discuss this case , for the sake of simplicity . ) _ different theories of gravity are characterized by different forms of entropy density just as different physical systems are characterized by different forms of entropy functionals .
_ i think this is completely in harmony with the thermodynamic spirit .
thermodynamics applies to any system ; if you want to describe a _
system , you need to specify its entropy functional or some other thermodynamic potential .
so what the development so far is telling us is that we need to put in some more extra physical input into the theory to find the field equations describing the theory .
0.2cm**harold : * * fine .
the above results imply that the field equations arising from any generally covariant action can be given a thermodynamic interpretation ; that is , you assumed the validity of the field equations and derived the local entropy balance .
your real aim , however , is to obtain the field equations from a dynamical principle rather than _ assume _ the field equations .
how do you propose to do that ?
0.2cm**me : * * to begin with , i want to paraphrase the above results in a slightly different manner which is probably more useful for the task we want to undertake .
note that , instead of dropping matter across the horizon , i could have equally well considered a virtual , infinitesimal ( planck scale ) , displacement of the @xmath35 normal to itself engulfing some matter .
we only need to consider infinitesimal displacements because the entropy of the matter is not ` lost ' until it crosses the horizon ; that is , until when the matter is at an infinitesimal distance ( a few planck lengths ) from the horizon .
all the relevant physical processes take place at a region very close to the horizon and hence an infinitesimal displacement of @xmath35 normal to itself will engulf some matter .
some entropy will be again lost to the outside observers unless displacing a piece of local rindler horizon costs some entropy .
so we expect the entropy balance condition derived earlier to ensure this and indeed it does . an infinitesimal displacement of a local patch of the stretched horizon in the direction of @xmath38 , by an infinitesimal proper distance @xmath172 , will change the proper volume by @xmath173 where @xmath174 is the metric in the transverse space
. the flux of energy through the surface will be @xmath175 and the corresponding entropy flux can be obtained by multiplying the energy flux by @xmath176 . hence the ` loss ' of matter entropy to the outside observer when the virtual displacement of the horizon swallows some hot tea is @xmath177 . to find the change in the gravitational entropy
, we again use the noether current @xmath93 corresponding to the local killing vector @xmath33 .
multiplying by @xmath130 and @xmath50 , we get @xmath178 as the stretched horizon approaches the true horizon , we know that @xmath179 and @xmath180 making the last term vanish .
so @xmath181 showing again the validity of local entropy balance .
0.2cm**harold : * * it appears to me that this is similar to switching from a passive point of view to an active point of view . instead of letting a cup of tea fall into the horizon
, you are making a virtual displacement of the horizon surface to engulf the tea which is infinitesimally close to the horizon .
but in the process , you have introduced the notion of virtual displacement of horizons and for the theory to be consistent , this displacement of these surface degrees of freedom should cost you some entropy . right ?
0.2cm**me : * * yes . if gravity is an emergent , long wavelength , phenomenon like elasticity then the diffeomorphism @xmath182 is analogous to the elastic deformations of the spacetime solid " @xcite .
it then makes sense to demand that the entropy density should be a functional of @xmath33 and their derivatives @xmath183 . by constraining the functional form of this entropy density
, we can choose the field equations of gravity .
recall that thermodynamics relies entirely on the form of the entropy functional to make predictions .
if we constrain the form of the entropy , we constrain the theory .
so the next step is to assume a suitable form of entropy functional for gravity @xmath184 in terms of the normal to the null surface .
then it seems natural to demand that the dynamics should follow from the extremum prescription @xmath185=0 $ ] for _ all null surfaces in the spacetime _ where @xmath186 is the matter entropy .
0.2cm**harold : * * what do we take for @xmath184 and @xmath186 ?
0.2cm**me : * * the form of @xmath186 is easy to ascertain from the previous discussion . if @xmath99 is the matter energy - momentum tensor in a general @xmath187 dimensional spacetime then an expression for matter entropy _
relevant for our purpose _ can be taken to be @xmath188 where @xmath189 is a null vector field . from our
we see that the entropy density associated with proper 3-volume is @xmath190 where on the horizon the vector @xmath33 becomes proportional to a null vector @xmath189 .
if we now use the rindler coordinates in in which @xmath191 and interpret the factor @xmath192 as arising from an integration of @xmath193 in the range @xmath194 we find that the entropy density associated with a proper four volume is @xmath195 .
this suggests treating as the matter entropy .
for example , if @xmath99 is due to an ideal fluid at rest in the lif then @xmath196 will contribute @xmath197 , which by gibbs - duhem relation is just @xmath198 where @xmath199 is the entropy density and @xmath200 is the properly redshifted temperature with @xmath201 being the periodicity of the euclidean time coordinate . then @xmath202 which matches with in the appropriate limit .
0.2cm**harold : * * well , that may be all right for an ideal fluid .
but for a general source , like say the electromagnetic field ( which will act as a source in metric )
i do nt even know how to define entropy .
but i am willing to accept as a definition .
0.2cm**me : * * actually , it is better than that .
we _ do _ have the notion of _ energy _ flux across a surface with normal @xmath130 being @xmath203 which holds for _ any _ source @xmath204 .
given some energy flux @xmath59 in the rindler frame , there is an associated entropy flux loss @xmath205 as given by .
you may think that an ordered field has no temperature or entropy but a rindler observer will say something different . for any state , she will have a corresponding density matrix @xmath11 and an entropy @xmath206 ; after all , she will attribute entropy even to vacuum state .
it is _ this _ entropy which is given by and .
0.2cm**harold : * * interesting .
there is also this time integration which you limit to the range @xmath194 in .
this is fine in euclidean sector and may be you can rotate back to lorentzian sector but it makes me a little uncomfortable .
0.2cm**me : * * well , i again have to invoke local nature of the argument which , as we discussed , is obvious in euclidean sector but the concept of causality , loss of information etc are obvious in the lorentzian sector in which i have light cones and null surfaces .
so i do have to switch back and forth .
may be there is a better way of formulating this which i have not yet figured out ; but for our purpose you can even think of all integrals being done in euclidean sector if you are happier with that .
0.2cm**harold : * * fine .
what about @xmath184 ?
0.2cm**me : * * for this , i will first describe the simplest possible choice and will then consider a more general expression .
the simplest choice is to postulate @xmath184 to be a quadratic expression @xcite in the derivatives of the normal : @xmath207 where the explicit form of @xmath208 is ascertained below .
the expression for the total entropy , now becomes : @xmath209=-\int_{\ensuremath{\mathcal{v}}}{d^dx\sqrt{-g } } \left(4p_{ab}^{{\phantom{a}}{\phantom{b}}cd } { \ensuremath{\nabla}}_cn^a{\ensuremath{\nabla}}_dn^b - t_{ab}n^an^b\right ) \
, , \label{ent - func-2}\ ] ] if you want , you can forget everything we said so far and start with this expression as defining our theory !
0.2cm**harold : * * i suppose this is your variational principle and you will now extremise @xmath210 with respect to @xmath211 .
0.2cm**me : * * yes , but i want to first explain the crucial conceptual difference between the extremum principle introduced here and the conventional one . usually , given a set of dynamical variables @xmath211 and a functional @xmath212 $ ] , the extremum principle will give a set of equations for the dynamical variable @xmath211 .
here the situation is completely different .
we expect the variational principle to hold for _ all _ null vectors @xmath189 thereby leading to a condition on the _ background metric . _
( of course , one can specify any null vector @xmath213 by giving its components @xmath214 with respect to fixed set of basis vectors @xmath215 with @xmath216 etc so that @xmath217 .
so the class of all null vectors can be mapped to the scalar functions @xmath218 with the condition @xmath219 . ) obviously , the functional in must be rather special to accomplish this and one needs to impose restrictions on @xmath208 ( and @xmath99 though that condition turns out to be trivial ) to achieve this .
it turns out as we shall see below that two conditions are sufficient to ensure this .
first , the tensor @xmath220 should have the same algebraic symmetries as the riemann tensor @xmath221 of the @xmath0-dimensional spacetime .
this condition can be ensured if we define @xmath222 as @xmath223 where @xmath224 is some scalar .
second , i will postulate the condition : @xmath225 as well as @xmath226 which is anyway satisfied by any matter energy - momentum tensor .
0.2cm**harold : * * can you give some motivation for these conditions ?
0.2cm**me : * * as regards , the motivation will become clearer later on .
basically , i will show that this approach leads to the same field equations as the one with @xmath5 as gravitational lagrangian in the conventional approach ( that is why i have used the symbol @xmath5 for this scalar ! ) .
one possible motivation for arises from fact that it will ensure the field equations do not contain any derivative higher than second order of the metric .
another possible interpretation arises from the analogy introduced earlier .
if you think of @xmath189 as analogous to deformation field in elasticity , then , in theory of elasticity @xcite one usually postulates the form of the thermodynamic potentials which are quadratic in first derivatives of @xmath211
. the coefficients of this term will be the elastic constants . here
the coefficients are @xmath227 and you may want to think of as saying the ` elastic constants of spacetime solid ' are actually ` constants ' .
but nothing depends on this picture .
in fact , i will show you later how this condition in can be relaxed .
0.2cm**harold : * * interesting .
you claim extremizing in this context with respect to all @xmath189 leads to an equation constraining the background metric .
if so , this is a peculiar variational principle .
0.2cm**me : * * let me show you how this arises . varying the normal vector field @xmath228 after adding a lagrange multiplier function @xmath229 for imposing the condition @xmath230 , we get @xmath231 where we have used the symmetries of @xmath208 and @xmath99 .
an integration by parts and the condition @xmath232 , leads to @xmath233\delta n^b}\nonumber\\ & & + 8\int_{{\ensuremath{\partial{\ensuremath{\mathcal{v}}}}}}{d^{d-1}x\sqrt{h}\left[k_d p_{ab}^{{\phantom{a}}{\phantom{b}}cd}\left({\ensuremath{\nabla}}_c n^a\right)\right]\delta n^b } \
, , \label{ent - func-4}\end{aligned}\ ] ] where @xmath25 is the @xmath0-vector field normal to the boundary @xmath234 and @xmath235 is the determinant of the intrinsic metric on @xmath234 . as usual , in order for the variational principle to be well defined , we require that the variation @xmath236 of the vector field should vanish on the boundary
. the second term in therefore vanishes , and the condition that @xmath237 $ ] be an extremum for arbitrary variations of @xmath228 then becomes @xmath238 where we used the antisymmetry of @xmath208 in its upper two indices to write the first term .
the definition of the riemann tensor in terms of the commutator of covariant derivatives reduces the above expression to @xmath239 and we see that the equations of motion _ do not contain _ derivatives with respect to @xmath189 which is , of course , the crucial point .
this peculiar feature arose because of the symmetry requirements we imposed on the tensor @xmath208 .
we require that the condition in holds for _ arbitrary _ vector fields @xmath228 .
one can easily show@xcite that this requires @xmath240= 8\pi t{}_b^a + \lambda\delta^a_b \label{ent - func-71}\ ] ] comparison with shows that these are precisely the field equations for gravity ( with a cosmological constant arising as an undetermined integration constant ; more about this later ) in a theory with lagrangian @xmath5 when is satisfied .
that is , we have @xmath241 with @xmath242 the crucial difference between and is that , the @xmath8 in contains no derivatives of the metric higher than second order thereby leading to field equations which are second order in the metric .
in contrast , can contain up to fourth order derivatives of the metric .
0.2cm**harold : * * let me get this straight .
suppose i start with a total lagrangian @xmath243 , define a @xmath227 by ensuring it satisfies .
then i get certain field equations by varying the metric .
you have just proved that i will get the _
same _ field equations ( but with a cosmological constant ) if i start with the expression in , maximize it with respect to @xmath189 and demand that it holds for all @xmath189 .
the maths is clear but i have several doubts . to begin with ,
why does the maths work out ? !
0.2cm**me : * * i will let you into the secret by doing it differently .
note that , using the constraints on @xmath227 i can prove the identity @xmath244 - 4n^ap_{ab}^{{\phantom{a}}{\phantom{b}}cd } { \ensuremath{\nabla}}_c{\ensuremath{\nabla}}_dn^b\nonumber\\ & = & 4{\ensuremath{\nabla}}_c[p_{ab}^{{\phantom{a}}{\phantom{b}}cd } n^a{\ensuremath{\nabla}}_dn^b]-2n^ap_{ab}^{{\phantom{a}}{\phantom{b}}cd } { \ensuremath{\nabla}}_{[c}{\ensuremath{\nabla}}_{d]}n^b\nonumber\\ & = & 4{\ensuremath{\nabla}}_c[p_{ab}^{{\phantom{a}}{\phantom{b}}cd } n^a{\ensuremath{\nabla}}_dn^b]-2n^ap_{ab}^{{\phantom{a}}{\phantom{b}}cd } r^b_{\phantom{b}icd}n^i\nonumber\\ & = & 4{\ensuremath{\nabla}}_c[p_{ab}^{{\phantom{a}}{\phantom{b}}cd } n^a{\ensuremath{\nabla}}_dn^b]+2n^ae_{ai}n^i\end{aligned}\ ] ] where the first line uses , the second line uses the antisymmetry of @xmath208 in c and d , the third line uses the standard identity for commutator of covariant derivatives and the last line is based on when @xmath245 and hold . using this in the expression for @xmath210 in and integrating the four - divergence term ,
i can write @xmath209=-\int_{\partial{\ensuremath{\mathcal{v}}}}{d^{d-1}x k_c\sqrt{h } } ( 4p_{ab}^{{\phantom{a}}{\phantom{b}}cd } n^a{\ensuremath{\nabla}}_dn^b ) -\int_{\ensuremath{\mathcal{v}}}{d^dx\sqrt{-g}}\left[(2e_{ab}-t_{ab})n^an^b\right ] \label{thetrick}\ ] ] so , when i consider variations ignoring the surface term i am effectively varying @xmath246 with respect to @xmath211 and demanding that it holds for all @xmath211 .
that should explain to you why it leads to @xmath247 except for a cosmological constant .
0.2cm**harold : * * ha !
of course , there is an ambiguity of adding a term of the form @xmath248 in the integrand of the second term in leading to the final equation @xmath249 but the bianchi identity @xmath250 along with @xmath251 will make @xmath229 actually a constant .
0.2cm**me : * * yes .
remember that .
we will discuss cosmological constant issue separately in the end .
0.2cm**harold : * * i see that it also connects up with your previous use of @xmath252 as some kind of gravitational entropy density .
your expression for gravitational entropy is actually @xmath253&=&-\int_{\ensuremath{\mathcal{v}}}{d^dx\sqrt{-g}}4p_{ab}^{{\phantom{a}}{\phantom{b}}cd } { \ensuremath{\nabla}}_cn^a{\ensuremath{\nabla}}_dn^b\\ & = & -\int_{\partial{\ensuremath{\mathcal{v}}}}{d^{d-1}x k_c\sqrt{h } } ( 4p_{ab}^{{\phantom{a}}{\phantom{b}}cd } n^a{\ensuremath{\nabla}}_dn^b ) -\int_{\ensuremath{\mathcal{v}}}{d^dx\sqrt{-g}}(2e_{ab}n^an^b)\nonumber\end{aligned}\ ] ] written in this form you have bulk contribution ( proportional to our old friend @xmath252 ) and a surface contribution .
when equations of motion hold , the bulk also get a contribution from matter which cancels it out leaving the entropy of a region @xmath254 to reside in its boundary @xmath234 . 0.2cm**me : * * yes . i need to think more about this . 0.2cm**harold : * * also , arising out of your letting me into the trick , i realize that i can now find an @xmath210 for any theory , even if does not hold .
you just have to reverse engineer it starting from @xmath246 as the entropy density and using the expression in for @xmath8 , right ?
so why do you insist on ?
0.2cm**me : * * you are right , of course . if you start with @xmath246 as the entropy density ( see eq .
14 of first paper in ref .
@xcite ) and work backwards you will get for @xmath255 the expression : @xmath256 varying this with respect to @xmath189 will then lead to the correct equations of motion and incidentally the same surface term . while one could indeed work with this more general expression , there are four reasons to prefer the imposition of the condition in .
first , it is clear from that when @xmath5 depends on the curvature tensor and the metric , @xmath8 can depend up to the fourth derivative of the metric if is not satisfied .
but when we impose then we are led to field equations which have , at most , second derivatives of the metric tensor which is a desirable feature .
second , as we shall see below , with that condition we can actually determine the form of @xmath5 ; it turns out that in @xmath257 , it uniquely selects einstein s theory , which is probably a nice feature . in higher dimensions
, it picks out a very geometrical extension of einstein s theory in the form of lanczos - lovelock theories .
third , it is difficult to imagine why the terms in should occur with very specific coefficients .
in fact , it is not clear why we can not have derivatives of @xmath221 in @xmath5 , if the derivatives of @xmath220 can occur in the expression for entropy . finally ,
if we take the idea of elastic constants being constants , then one is led to .
none of these rigorously exclude the possibility in and in fact this model has been explored recently @xcite .
0.2cm**harold : * * so far we have not fixed @xmath227 so we have not fixed the theory .
how does allow you to do this ?
0.2cm**me : * * in a complete theory , the explicit form of @xmath227 will be determined by the long wavelength limit of the microscopic theory just as the elastic constants can in principle be determined from the microscopic theory of the lattice . in the absence of such a theory
, we need to determine @xmath227 by general considerations .
essentially we need to determine scalar @xmath5 built from curvature tensor and the metric which satisfies the the constraint @xmath258 .
this problem can be completely solved the result is the lagrangian of a lanczos - lovelock theory .
such an @xmath5 can be written as a sum of terms , each involving products of curvature tensors with the @xmath259th term being a product of @xmath260 curvature tensors leading to @xmath261 where the @xmath262 are arbitrary constants and @xmath263 is the @xmath260-th order lanczos - lovelock lagrangian .
the @xmath264 term is proportional to @xmath265 and leads to einstein s theory .
it is conventional to take @xmath266 so that the @xmath267 , reduces to @xmath268 .
the normalizations for @xmath269 are somewhat arbitrary for individual @xmath263 since the @xmath262s are unspecified at this stage .
the @xmath270 term gives rise to what is known as gauss - bonnet theory . because of the determinant tensor , it is obvious that in any given dimension @xmath0 we can only have @xmath271 terms where @xmath272 .
it follows that , if @xmath257 , then only the @xmath273 are non - zero . of these
, the gauss - bonnet term ( corresponding to @xmath270 ) gives , on variation of the action , a vanishing bulk contribution in @xmath257 .
( in dimensions @xmath274 to 8 , one can have both the einstein - hilbert term and the gauss - bonnet term and so on . )
equivalently , the @xmath227 can be expressed as a series in the powers of derivatives of the metric as : @xmath275 where @xmath276 are coupling constants .
the lowest order term depends only on the metric with no derivatives .
the next term depends ( in addition to metric ) linearly on curvature tensor and the next one will be quadratic in curvature etc .
let us take a closer look at the structure which is emerging . the lowest order term in ( which leads to einstein s theory ) is @xmath277 to the lowest order , when we use for @xmath278 , the reduces to einstein s equations .
the corresponding gravitational entropy functional is : @xmath279=\int_{\ensuremath{\mathcal{v}}}\frac{d^dx}{8\pi } \left({\ensuremath{\nabla}}_an^b{\ensuremath{\nabla}}_bn^a - ( { \ensuremath{\nabla}}_cn^c)^2 \right)\ ] ] interestingly , the integrand in @xmath280 has the @xmath281 structure .
if we think of the @xmath257 spacetime being embedded in a sufficiently large k - dimensional _ flat _ spacetime we can obtain the same structure using the gauss - codazzi equations relating the ( zero ) curvature of k - dimensional space with the curvature of spacetime . as mentioned earlier , one can express any vector field @xmath189 in terms of a set of basis vector fields @xmath282 .
therefore , one can equivalently think of the functional @xmath283 as given by @xmath284=\int_{\ensuremath{\mathcal{v}}}\frac{d^dx}{8\pi } \left({\ensuremath{\nabla}}_an^b_i{\ensuremath{\nabla}}_bn^a_j - { \ensuremath{\nabla}}_cn^c_i{\ensuremath{\nabla}}_an^a_j \right)p^{ij}\ ] ] where @xmath285 is a suitable projection operator .
it is not clear whether the embedding approach leads to any better understanding of the formalism ; in particular , it does not seem to generalize in a natural fashion to lanczos - lovelock models .
the next order term ( which arises from the gauss - bonnet lagrangian ) is : @xmath286 and similarly for all the higher orders terms .
none of them can contribute in @xmath257 so we get einstein s theory as the unique choice if we assume @xmath257 .
if we assume that @xmath227 is to be built _ only _ from the metric , then this choice is unique in all @xmath0 .
0.2cm**harold : * * you originally gave a motivational argument as to why this @xmath210 should be thought of as entropy .
as far as the variational principle is concerned , this identification does not seem to play a crucial role .
0.2cm**me : * * it does rather indirectly . to see this
, you only need to consider the form of @xmath210 when the equations of motion are satisfies .
first of all , shows that when the equations of motion holds the total entropy of a bulk region is entirely on its boundary , which is nice .
further if you evaluate this boundary term @xmath287 ( where we have manipulated a few indices using the symmetries of @xmath227 ) in the case of a _ stationary _ horizon which can be locally approximated as rindler spacetime , one gets exactly the wald entropy of the horizon @xcite .
this is one clear reason as to why we can think of @xmath210 as entropy .
0.2cm**harold : * * is this entropy positive definite ?
do you worry about that ?
0.2cm**me : * * i do nt worry about that ( yet ! ) . in @xmath257
, i can prove that the on - shell entropy is positive definite .
but if one is dealing with a lanczos - lovelock model with horizons attributed with wald entropy , it is known that @xcite even on - shell entropy will not be positive definite for all range of parameters . may be this will put additional restrictions on the kind of gravitational theories which are physically reasonable .
( this approach has uncovered several other issues related to entropy , quasi - normal modes etc . and
even a possibility of entropy being quantized @xcite but all that will take us far afield . ) at present these questions are open .
0.2cm**harold : * * usually in an action principle one varies all the degrees of freedom in any order one chooses .
but in your extremum principle , we are expected to vary only @xmath189 .
how would i get equations of motion for matter in this approach ?
0.2cm**me : * * that is not a problem . at the classical level , the field equations are already contained in the condition @xmath288 which i impose ( with an intriguing interpretation , which you may not want to buy , that this is the constancy of elastic constants ! )
if you want to do quantum field theory in a curved spacetime , you can again use these field equations in the heisenberg picture .
the only question is when you insist that you need to do a path integral quantizations of the matter fields .
then , you have to , of course , vary @xmath189 first and get the classical equations for gravity because the expression in is designed as an entropy functional .
but after you have done that and written down the field equations for gravity , you can do the usual variation of matter lagrangian in a given curved spacetime and get the standard equations @xcite .
0.2cm**harold : * * i also notice that while the vector field @xmath211 in lif , to the lowest order , has no bulk dynamics , if you consider the integral over the lagrangian in a small region in lif , you will get a surface contribution . that seems strange , too .
0.2cm**me : * * not really .
i am not surprised by @xmath210 picking up just a surface contribution because even in the conventional approach einstein - hilbert action is holographic in a specific sense of the word @xcite and does exactly that .
0.2cm**harold : * * may be this is good time to sort this out .
you mentioned earlier that the democracy of observers and their right to do physics in spite of the existence of horizons has something to do with the holography of action .
i have no idea what you are talking about here !
0.2cm**me : * * let me elaborate .
we said that there should exist a mechanism which will encode the information in the region @xmath79 which is inaccessible to a particular observer at the boundary @xmath80 of that region@xcite . one possible way of ensuring this is to add a suitable boundary term to the action principle which will provide additional information content for observers who perceive a horizon .
such a procedure leads to three immediate consequences .
first , if the theory is generally covariant , so that observers with horizons ( like , for example , uniformly accelerated observers using a rindler metric ) need to be accommodated in the theory , such a theory _ must _ have an action functional that contains a surface term .
the generally covariant action in einstein s theory did contain a surface term .
the present approach explains the logical necessity for such a surface term in a generally covariant theory which was not evident in the standard approach .
0.2cm**harold : * * that s interesting .
you are now claiming that there is a connection between the following three facts .
( 1 ) the theory for gravity is built from a generally covariant lagrangian .
( 2 ) in a geometrical theory of gravity , horizons are inevitable but general covariance demands that all observers have an equal right to describe physics . ( 3 ) observers whose information is blocked by a horizon should still be able to somehow get around this fact with the information encoded on the boundary .
therefore , the lagrangian must have a boundary term .
viewed this way , it appears natural that the only generally covariant scalar lagrangian proportional to @xmath289 leads to a surface term in the action . but how does the surface term know what it is going on in the bulk ?
0.2cm**me : * * that is the second point .
if the surface term has to encode the information which is blocked by the horizon , then there must exist a simple relation between the bulk term and surface term in the action and hence you can not choose just any scalar .
this is indeed the case for the einstein - hilbert action ; there is a peculiar ( unexplained ) relationship between @xmath290 and @xmath291 : @xmath292 this shows that the einstein - hilbert gravitational action is ` holographic ' with the same information being coded in both the bulk and surface terms .
in fact , in any local region around an event , it is the surface term which contributes to the action at the lowest order . in the neighborhood of any event , the riemann
normal coordinates in which @xmath293 . in the gravitational lagrangian @xmath294 with @xmath295 and @xmath296 ,
the @xmath297 term vanishes in this neighborhood while @xmath298 leading to @xmath299 showing that in a small region around the event in the riemann normal coordinates , gravitational action can be reduced to a pure surface term .
0.2cm**harold : * * so in this perspective , you also expect the surface term to be related to the information content blocked by the horizon , right ?
0.2cm**me : * * indeed .
that is the third point .
if the surface term encodes information which is blocked by the horizon , then it should actually lead to the entropy of the horizon .
in other words , we should be able to compute the horizon entropy by evaluating the surface term .
this is indeed true and can be easily demonstrated @xcite .
the surface term does give the horizon entropy for any metric for which near - horizon geometry has the rindler form .
this explains another deep mystery in the conventional approach . in the usual approach ,
we _ ignore _ the surface term completely ( or cancel it with a counter - term ) and obtain the field equation from the bulk term in the action .
any solution to the field equation obtained by this procedure is logically independent of the nature of the surface term .
but we find that when the _ surface term _ ( which was ignored ) is evaluated at the horizon that arises in any given solution , it does correctly give the entropy of the horizon !
this is possible only because there is a relationship , given by , between the surface term and the bulk term which is again an unexplained feature in the conventional approach to gravitational dynamics . since the surface term has the thermodynamic interpretation as the entropy of horizons , and is related holographically to the bulk term , we are again led to an indirect connection between spacetime dynamics and horizon thermodynamics .
0.2cm**harold : * * i agree these results are extremely mysterious in the conventional approach , now that you brought it up .
i have not seen mentioned , let alone discussed in any work ( other than yours , of course . )
i presume this is one of what you call ` algebraic accidents ' .
but if your ideas about lanczos - lovelock theory being the natural candidate in d dimensions then the same ` algebraic accident ' should occur in lanczos - lovelock theories as well , right ?
0.2cm**me : * * yes . in fact it does which is gratifying and acts as a nontrivial consistency check on my alternative perspective .
one can show that the surface and bulk terms of all lanczos - lovelock theories satisfy an equation similar to . of course ,
since it was nt noticed for hilbert action , nobody bothered about lanczos - lovelock action till we @xcite unearthed it .
one can provide a simple , yet very general , proof of the connection between entropy and surface term in action in any static spacetime .
such a spacetime will have a killing vector @xmath33 and a corresponding noether current .
taking the @xmath145 component of and writing @xmath300 we obtain @xmath301 only spatial derivatives contribute in the first term on the right hand side when the spacetime is static .
this relation shows that the action obtained by integrating @xmath302 will generically have a surface term related to @xmath111 ( in einstein gravity will read as @xmath303 ; our result generalises the fact that @xmath304 can be expressed as a total divergence in static spacetimes .
) this again illustrates , in a very general manner , why the surface terms in the action functional lead to horizon entropy .
in fact can be integrated to show that in any static spacetime with a bifurcation horizon , the action can be interpreted as the free energy which generalises a result known in einstein gravity to lanczos - lovelock models .
0.2cm**harold : * * what are the other key algebraic accidents in the conventional approach which your perspective throws light on ?
0.2cm**me : * * there are several but let me describe one which is really striking ( and was first discussed in ref .
@xcite ) . consider a static , spherically symmetric horizon , in a spacetime described by a metric : @xmath305 let the location of the horizon be given by the simple zero of the function @xmath306 , say at @xmath307 .
the taylor series expansion of @xmath306 near the horizon @xmath308 shows that the metric reduces to the rindler metric near the horizon in the @xmath309 plane with the surface gravity @xmath310 .
then , an analytic continuation to imaginary time allows us to identify the temperature associated with the horizon to be @xmath311 where we have introduced the normal units .
the association of temperature in with the metric in only requires the conditions @xmath312 and @xmath313 .
the discussion so far did not assume anything about the dynamics of gravity or einstein s field equations .
we shall now take the next step and write down the einstein equation for the metric in , which is given by @xmath314 where @xmath315 is the radial pressure . when evaluated on the horizon @xmath307
we get the result : @xmath316 = 4\pi p a^2 \label{reqa}\ ] ] if we now consider two solutions to the einstein s equations differing infinitesimally in the parameters such that horizons occur at two different radii @xmath317 and @xmath318 , then multiplying the by @xmath319 , we get : @xmath320 the right hand side is just @xmath321 where @xmath322 is what is called the areal volume which is the relevant quantity when we consider the action of pressure on a surface area . in the first term , we note that @xmath323 is proportional to horizon temperature in . rearranging this term slightly and introducing a @xmath324 factor _ by hand _ into an otherwise classical equation to bring in the horizon temperature , we can rewrite as @xmath325 the labels below the equation indicate a natural and unique interpretation for each of the terms and the whole equation now becomes @xmath326 allowing us to read off the expressions for entropy and energy : @xmath327 where @xmath328 is the horizon area and @xmath329 .
the result shows that einstein s equations can be re - interpreted as a thermodynamic identity for a virtual displacement of the horizon by an amount @xmath319 .
0.2cm**harold : * * i suppose the uniqueness of the factor @xmath330 , where @xmath331 is the proper area of a surface of radius @xmath317 in spherically symmetric spacetimes , implies that we can not carry out the same exercise by multiplying by some other arbitrary factor @xmath332 instead of just @xmath319 in a natural fashion .
this , in turn , uniquely fixes both @xmath55 and the combination @xmath333 .
the product @xmath333 is classical and is independent of @xmath324 and hence we can determine @xmath56 and @xmath210 only within a multiplicative factor . the only place you introduced @xmath324 by hand is in using the euclidean extension of the metric to fix the form of @xmath56 and thus @xmath210 .
0.2cm**me : * * yes . with that i can remove the ambiguity in the overall multiplicative factor .
so , given the structure of the metric in and einstein s equations , we can determine @xmath334 and @xmath335 uniquely .
the fact that @xmath336 and @xmath337 is analogous to the situation in classical thermodynamics in contrast with statistical mechanics .
the @xmath333 in thermodynamics is independent of boltzmann s constant while statistical mechanics will lead to @xmath338 and @xmath339 .
0.2cm**harold : * * that is a bit mind - boggling .
usually , the rigorous way of obtaining the temperature of a horizon say , a black hole horizon is by studying a quantum field in the externally specified metric .
_ you never need to specify whether the metric is a solution to einstein s equations .
_ now you are telling me that the same result arises _ without any reference to an externally specified quantum field theory but on using einstein s equations on the horizons_. how does the einstein equations know that there is a temperature , entropy etc ?
0.2cm**me : * * yes . that is the algebraic coincidence .
more sharply stated , we have no explanation as to why an equation like should hold in classical gravity , if we take the conventional route .
this strongly suggests that the association of entropy and temperature with a horizon is quite fundamental and is actually connected with the dynamics ( encoded in einstein s equations ) of the gravitational field . the fact that quantum field theory in a spacetime with horizon exhibits thermal behaviour
should then be thought of as a _ consequence _ of a more fundamental principle .
0.2cm**harold : * * if so the idea should also have a more general validity .
does it ?
0.2cm**me : * * yes . one can again show that the field equations of more general theories of gravity ( like in lanczos - lovelock models ) also reduce to the same thermodynamic identity @xmath340 when evaluated on the horizon .
this has now been demonstrated @xcite for an impressively wide class of models like ( i ) the stationary axisymmetric horizons and ( ii ) evolving spherically symmetric horizons in einstein gravity , ( iii ) static spherically symmetric horizons and ( iv ) dynamical apparent horizons in lovelock gravity , and ( v ) three dimensional btz black hole horizons , ( vi ) frw cosmological models in various gravity theories and ( vii ) even @xcite in the case horava - lifshitz gravity .
it is not possible to understand , in the conventional approach , why the field equations should encode information about horizon thermodynamics .
0.2cm**harold : * * this is a fairly strong argument in favour of a thermodynamic underpinning for the dynamics of gravity .
but before i accept that in toto , i need to convince myself that there is no simpler explanation for this result .
i accept that none is given in the literature but how about the standard first law of black hole thermodynamics ?
your relation looks similar to it so i wonder whether there is a connection .
0.2cm**me : * * no . we are talking about very different things . in general , in spite of the superficial similarity , is _ different _ from the conventional first law of black hole thermodynamics due to the presence of @xmath321 term . the difference is easily seen , for example , in the case of reissner - nordstrom black hole for which @xmath341 is non - zero due to the presence of nonzero electromagnetic energy - momentum tensor in the right hand side of einstein s equations . if a _ chargeless _ particle of mass @xmath342 is dropped into a reissner - nordstrom black hole , then the standard first law of black hole thermodynamics will give @xmath343 .
but in , the energy term , defined as @xmath344 , changes by @xmath345dm\neq dm$ ] .
it is easy to see , however , that for the reissner - nordstrom black hole , the combination @xmath346 is precisely equal to @xmath342 making sure @xmath343 .
so we need the @xmath321 term to get @xmath343 from when a _ chargeless _ particle is dropped into a reissner - nordstrom black hole .
more generally , if @xmath319 arises due to changes @xmath342 and @xmath347 , it is easy to show that gives @xmath348 where the second term arises from the electrostatic contribution .
this ensures that is perfectly consistent with the standard first law of black hole dynamics in those contexts in which both are applicable but @xmath349 in general .
you would have also realized that the way was derived is completely local and quite different from the way one obtains first law of black hole thermodynamics .
0.2cm**harold : * * yes , i see that .
it appears that gravitational field equations and their solutions with horizons has a deeper connection with thermodynamics than is apparent .
in fact , i believe you would claim the thermodynamic perspective is more fundamental than the field equations describing gravity .
0.2cm**me : * * precisely .
that is why i spent lot of time explaining the thermodynamic motivation in sections [ sec : lro ] and [ sec : reinterpret ] while i could have derived the field equations just by extremizing the expression in .
but in a way everything else is just motivational if you are willing to accept the perspective based on as fundamental .
0.2cm**harold : * * that brings up the question you promised a discussion on . what about the cosmological constant @xcite ? in the conventional approach
, one introduces it as a term in the gravitational lagrangian .
you do nt have any such term in but nevertheless the cosmological constant appears in your final equations !
0.2cm**me : * * yes and i would claim that this is another very attractive feature of this new perspective . in the standard approach ,
one starts with an action @xmath350 and varies ( i ) the matter degrees of freedom to obtain the equations of motion for matter and ( ii ) the metric @xmath6 to obtain the field equations of gravity .
the equations of motion for matter remain invariant if one adds a constant , say , @xmath351 to the matter lagrangian , which is equivalent to adding a constant @xmath62 to the hamiltonian density of the matter sector . physically , this symmetry reflects the fact that the zero level of the energy is arbitrary in the matter sector and can be set to any value without leading to observable consequences
. however , gravity breaks this symmetry which the matter sector has .
a shift @xmath352 will change the energy - momentum tensor @xmath353 which acts as the source of gravity by a term proportional to @xmath354 .
therefore , having a nonzero baseline for energy density of matter is equivalent to a theory with cosmological constant which in turn will lead to observable consequences .
if we interpret the evidence for dark energy in the universe ( see ref .
@xcite ; for a critical look at data , see ref .
@xcite and references therein ) as due to the cosmological constant , then its value has to be fine - tuned to enormous instead of the usual @xmath355 . in standard model of particle physics
the ratio between planck scale to neutrino mass scale is @xmath356 for which we have no theoretical explanation .
so when we worry about the fine tuning of cosmological constant without expressing similar worries about standard model of particle physics , we are essentially assuming that @xmath357 is not a matter for concern but @xmath358 is .
this subjective view is defensible but needs to be clearly understood . ]
accuracy to satisfy the observational constraints .
it is not clear why a particular parameter in the low energy matter sector has to be fine - tuned in such a manner . in the alternative perspective described here ,
the functional in is clearly invariant under the shift @xmath352 or equivalently , @xmath359 , since it only introduces a term @xmath360 for any null vector @xmath211 .
in other words , one _ can not _ introduce the cosmological constant as a low energy parameter in the action in this approach .
we saw , however , that the cosmological constant reappears an _
an integration constant _ when the equations are solved .
the integration constants which appear in a particular solution have a completely different conceptual status compared to the parameters which appear in the action describing the theory .
it is much less troublesome to choose a fine - tuned value for a particular integration constant in the theory if observations require us to do so . from this point of view , the cosmological constant problem is considerably less severe when we view gravity from the alternative perspective .
0.2cm**harold : * * i suppose you succeed in having the extra symmetry under the shift @xmath359 because you are not treating metric as a dynamical variable , right ?
0.2cm**me : * * right .
in fact one can state a stronger result @xcite .
consider any model of gravity satisfying the following three conditions : ( 1 ) the metric is varied in a local action to obtain the equations of motion .
( 2 ) we demand full general covariance of the equations of motion . ( 3 ) the equations of motion for matter sector is invariant under the addition of a constant to the matter lagrangian .
then , we can prove ` no - go ' theorem that the cosmological constant problem can not be solved in such model @xcite .
the proof is elementary .
our demand ( 2 ) of general covariance requires the matter action to be an integral over @xmath361 .
the demand ( 3 ) now allows us to add a constant @xmath362 , say , to @xmath363 leading to a coupling @xmath364 between @xmath362 and the metric @xmath16 . by our demand ( 1 ) ,
when we vary @xmath16 the theory will couple to @xmath362 through a term proportional to @xmath365 thereby introducing an arbitrary cosmological constant into the theory .
the power of the above ` no - go theorem ' lies in its simplicity !
it clearly shows that we can not solve cosmological constant problem unless we drop one of the three demands listed in the above paragraph . of these
, we do not want to sacrifice general covariance encoded in ( 2 ) ; neither do we have a handle on low energy matter lagrangian so we can not avoid ( 3 ) .
so the only hope we have is to introduce an approach in which gravitational field equations are obtained from varying some degrees of freedom other than @xmath16 in a maximization principle .
this suggests that the so called cosmological constant problem has its roots in our misunderstanding of the nature of gravity .
0.2cm**harold : * * i thought that any spin-2 long range field @xmath366 ( arising , for example , in the linear perturbation around flat spacetime through @xmath367 ) obeying principle of equivalence has to generically couple to @xmath99 through a term in the action @xmath368 .
but in your model , this does not seem to happen .
0.2cm**me : * * that s correct .
it is sometimes claimed that a spin-2 graviton in the linear limit _ has to _ couple to @xmath99 in a universal manner , in which case , one will have the graviton coupling to the cosmological constant . in our approach , the linearized field equations for the spin-2 graviton field @xmath369 , in a suitable gauge , will be @xmath370 for all null vectors
this equation is still invariant under @xmath359 showing that the graviton does _ not _ couple to cosmological constant .
0.2cm**harold : * * but can you predict the observed value of the cosmological constant in your approach ?
0.2cm**me : * * alas , no .
but i claim providing a mechanism in which the _ bulk cosmological constant decouples from gravity _ is a major step forward .
if the cosmological constant was strictly zero , my perspective has a natural explanation for it which no one else had !
it was always thought that this should arise from some unknown symmetry and i have provided you with a model which has such symmetry .
i believe the small value of the observed cosmological constant arises from non - perturbative quantum gravitational effects at the next order , but i do nt have a fully satisfactory model .
( see , however , ref .
0.2cm**harold : * * we have covered a lot of ground some of which is purely technical while the rest are conceptual or interpretational . in your mind
the distinction may be unimportant but others will react differently to results which can be rigorously proved compared to interpretational aspects , however elegant the latter may be
. may be you would care to separate them out and provide a summary ?
0.2cm**me : * * fine . from a purely algebraic point of view , without bringing in any physical interpretation or motivation , we can prove the following mathematical results : * consider a functional of null vector fields @xmath213 in an arbitrary spacetime given by [ or , more generally , by ] . demanding that this functional is an extremum for all null vectors @xmath189 leads to the field equations for the background geometry given by @xmath371 where @xmath8 is given by [ or , more generally , by ] .
thus field equations in a wide class of theories of gravity can be obtained from an extremum principle without varying the metric as a dynamical variable .
* these field equations are invariant under the transformation @xmath359 , which relates to the freedom of introducing a cosmological constant as an integration constant in the theory .
further , this symmetry forbids the inclusion of a cosmological constant term in the variational principle by hand as a low energy parameter . _
that is , we have found a symmetry which makes the bulk cosmological constant decouple from the gravity .
_ when linearized around flat spacetime , the graviton inherits this symmetry and does not couple to the cosmological constant . * on - shell , the functional in [ or , more generally , by ] contributes only on the boundary of the region . when the boundary is a horizon , this terms gives precisely the wald entropy of the theory .
0.2cm**harold : * * it is remarkable that you can derive not only einstein s theory uniquely in @xmath257 but even lanczos - lovelock theory in @xmath372 from an extremum principle involving the null normals _ without varying @xmath16 in an action functional!_. i also see from that in the case of einstein s theory , you have an lagrangian @xmath373})n^b$ ] for a vector field @xmath189 which becomes vacuous in flat spacetime in which covariant derivatives become partial derivatives .
there is clearly no dynamics in @xmath189 but they do play a crucial role .
so i see that you can get away without ever telling me what the null vectors @xmath189 actually means because they disappear from the scene after serving their purpose .
while this may be mathematically clever , it is very unsatisfactory physically .
you may not have a rigorous model for these degrees of freedom but what is your picture ?
0.2cm**me : * * my picture is made of the following ingredients , each of which seems reasonable but far from having rigorous mathematical justification at this stage .
* assume that the spacetime is endowed with certain microscopic degrees of freedom capable of exhibiting thermal phenomena .
this is just the boltzmann paradigm : _ if one can heat it , it must have microstructure ! _ ; and one can heat up a spacetime .
* whenever a class of observers perceive a horizon , they are `` heating up the spacetime '' and the degrees of freedom close to a horizon participate in a very _ observer dependent _ thermodynamics .
matter which flows close to the horizon ( say , within a few planck lengths of the horizon ) transfers energy to these microscopic , near - horizon , degrees of freedom _ as far as the observer who sees the horizon is concerned_. just as entropy of a normal system at temperature @xmath56 will change by @xmath374 when we transfer to it an energy @xmath59 , here also an entropy change will occur .
( a freely falling observer in the same neighbourhood , of course , will deny all these ! ) * we proved that when the field equations of gravity hold , one can interpret this entropy change in a purely geometrical manner involving the noether current . from this point of view
, the normals @xmath189 to local patches of null surfaces are related to the ( unknown ) degrees of freedom that can participate in the thermal phenomena involving the horizon . * just as demanding the validity of special relativistic laws with respect to all freely falling observers leads to the kinematics of gravity , demanding the local entropy balance in terms of the thermodynamic variables as perceived by local rindler observers leads to the field equations of gravity in the form @xmath371 .
0.2cm**harold : * * it is an interesting picture but is _ totally observer dependent _ , right ?
a local rindler observer or an observer outside a black hole horizon might attribute all kinds of thermodynamics and entropy changes to the horizons she perceives .
but an inertial observer or an observer falling through the schwarzschild horizon will see none of these phenomena .
0.2cm**me : * * exactly .
i claim we need to accept the fact that a whole lot of thermodynamic phenomena needs to be now thought of as observer dependent . for example , if you throw some hot matter on to a schwarzschild black hole , then when it gets to a few planck lengths away from the horizon and hovers around it , i expect it to interact with the microscopic horizon degrees of freedom _ as far as an outside observer is concerned_. after all , such an observer would claim that all matter stays arbitrarily close but outside the horizon for all eternity
. a freely falling observer through the horizon will have a completely different picture but we have learnt to live with this dichotomy as far as elementary kinematics goes .
i think we need to do the same as regards thermodynamics and quantum processes .
0.2cm**harold : * * in a way , every key progress in physics involved realizing that something we thought as absolute is not absolute .
with special relativity it was the flow of time and with general relativity it was the concept of global inertial frames and when we brought in quantum fields in curved spacetime it was the notion of particles .
0.2cm**me : * * and the notion of temperature , do nt forget that .
we now know that the temperature attributed to even vacuum state depends on the observer . we need to go further and integrate the entire thermodynamic machinery involving highly excited semi - classical states , say , cups of tea with ( what we believe to be ) `` real '' temperature with this notion of lrfs having their own temperature .
i do nt think this has been done in a satisfactory manner yet @xcite .
0.2cm**harold : * * so , what next ? what is the `` to do list '' ?
0.2cm**me : * * to name a few which comes to ones mind , i can list them as technical ones and conceptual ones . on the technical side : \(i ) it would be nice to make the notion of lrf and the horizon a bit more rigorous .
for example , the idea of an approximate killing vector could be made more precise and one might like to establish the connection between locality , that is apparent in the euclidean sector and the causality , which is apparent in the lorentzian section that has the light cones . by and large
, one would like to make rigorous the use of lrfs by , say , computing the next order corrections .
\(ii ) a lot more can be done to clarify the observer dependence of the entropy .
( the study of horizon thermodynamics makes one realize that one does not really quite understand what entropy is ! ) .
it essentially involves _ exact _ computation of @xmath375 where @xmath376 and @xmath66 are the entropies attributed to the excited and ground state by a rindler observer .
this should throw more light on the expression for @xmath377 used in .
in particular , it would be nice to have a detailed model which shows why @xmath377 involves the combination of @xmath59 of matter and @xmath56 of the horizon . one
would then use these insights to understand why actually represents the relevant entropy functional for matter for arbitrary @xmath99 .
\(iii ) it will be nice to have a handle on the positivity or otherwise of the entropy functional used in .
these technical issues , i believe can be tackled in more or less straightforward manner , though the mathematics can be fairly involved . but
as i said , they are probably not crucial to the alternative perspective or its further progress .
the latter will depend on more serious conceptual issues , some of which are the following : \(i ) how come the microstructure of spacetime exhibits itself indirectly through the horizon temperature even at scales much larger than planck length ?
i believe this is because the event horizon works as some kind of magnifying glass allowing us to probe trans - planckian physics @xcite but this notion needs to be made more precise .
\(ii ) how does one obtain the expression for entropy in from some microscopic model ?
in particular , such an analysis even with a toy model should throw more light on why normals to local patches of null surfaces play such a crucial role as effective degrees of freedom in the long wavelength limit .
of course , such a model should also determine the expression for @xmath227 and get the metric tensor and spacetime as derived concepts - a fairly tall order!. ( this is somewhat like obtaining theory of elasticity starting from a microscopic model for a solid , which , incidentally , is not a simple task either . )
0.2cm**harold : * * but what about the `` deep questions '' like , for example , the physics near the singularities
? since you get the same field equations as anybody else does , you will have the same solutions , same singularities etc .
0.2cm**me : * * i told you that i am not doing statistical mechanics ( which would be the full quantum theory ) of spacetime but only thermodynamics . to answer issues related to singularities etc .
, one actually needs to discover the statistical mechanics underlying the thermodynamic description i have presented here .
we can have another chat , after i figure out the statistical mechanics of the spacetime microstructure !
the questions of harold mostly represent issues raised by several colleagues far too numerous to name individually in my lectures , discussions etc .
i thank all of them for helping me to sharpen the ideas .
i also thank a.d .
patel , k. subramanian , sudipta sarkar , aseem paranjape , d. kothawala and sunu engineer for several rounds of discussions over the past many years .
t. padmanabhan , _ entropy density of spacetime and thermodynamic interpretation of field equations of gravity in any diffeomorphism invariant theory _ , [ arxiv:0903.1254 ] ; t. padmanabhan , _ entropy density of spacetime and gravity : a conceptual synthesis _ , to appear in ijmpd , ( 2009 ) .
t. padmanabhan , _ advanced science letters _ , * 2 * , 174 ( 2009 ) [ arxiv:0807.2356 ] ; t. padmanabhan , _ gravity - the inside story _ , [ first prize essay , gravity research foundation essay contest , 2008 . ] gen .
* 40 * ( 2008 ) , 2031 .
nathalie deruelle , joseph katz and sachiko ogushi , class .
* 21 * , ( 2004 ) , 1971 .
[ gr - qc/0310098 ] ; g.l .
cardoso , b. de wit , t. mohaupt , arxiv : hep - th/9904005v2 ; t. padmanabhan , _ gravitation : foundations and frontiers _ , ( cambridge university press , 2009 ) , in press .
for a sample of other approaches , see e.g. , g. e. volovik , _ the universe in a helium droplet _ , ( oxford university press , 2003 ) ; t. jacobson , phys . rev . lett .
* 75 * ( 1995 ) , 1260 , [ gr - qc/9504004 ] ; l. sindoni et al .
, arxiv:0909.5391 ; b.l .
hu , arxiv:0903.0878 ; m. visser , _
mod.phys.lett_. * a17 * ( 2002 ) 977 [ gr - qc/0204062 ] ; c. barcelo et al . , _ int.j.mod.phys . _ * d10 * ( 2001 ) 799 [ gr - qc/0106002 ] ; chao - guang huang , jia - rui sun , gr - qc/0701078 ; j.makela , gr - qc/0701128 .
t. padmanabhan , apoorva patel , _ role of horizons in semiclassical gravity : entropy and the area spectrum _ [ gr - qc/0309053 ] ; t. padmanabhan , gen.rel.grav . , * 35 * , 2097 - 2103 ( 2003 ) [ fifth prize essay ; gravity research foundation essay contest , 2003 ] ; mod.phys.letts . a 17 , 923 ( 2002 ) [ gr - qc/0202078 ] .
this is related to the famous question first posed by wheeler to bekenstein : what happens if you mix cold and hot tea and pour it down a horizon , erasing all traces of crime " in increasing the entropy of the world ?
this is based on what wheeler told me in 1985 , from _ his _ recollection of events ; it is also mentioned in his book , j.a.wheeler , _ a journey into gravity and spacetime _ , [ scientific american library , ny , 1990 ] page 221 .
i have heard somewhat different versions from other sources .
h. s. snyder _ phys .
_ , * 71 * , 38 ( 1947 ) ; b. s. dewitt , _ phys . rev .
_ , * 13 * , 114 ( 1964 ) ; t. yoneya _ prog .
_ , * 56 * , 1310 ( 1976 ) ; t. padmanabhan _ ann .
( n.y . ) , * 165 * , 38 ( 1985 ) ; _ class . quantum grav . _ * 4 * , l107 ( 1987 ) ; t. padmanabhan _ phys .
lett . _ * 78 * , 1854 ( 1997 ) [ hep - th/9608182 ] ; _ phys .
_ * d 57 * , 6206 ( 1998 ) ; k.srinivasan et al .
_ * d 58 * 044009 ( 1998 ) [ gr - qc/9710104 ] ; x.calmet et al.,phys.rev.lett.*93*:211101,(2004);hep - th/0505144 ; m. fontanini et al .
_ phys.lett . _
* b 633 * , 627 ( 2006 ) hep - th/0509090 . for a review ,
garay , _ int .
phys . _ * a10 * , 145 ( 1995 ) .
m cvetic , et al . , _ nucl.phys . _
* b 628 * , 295 , 2002 , [ hep - th/0112045 ] ; s nojiri , s d odintsov , _ phys .
_ , * d66 * , 044012,2002 , [ hep - th/0204112 ] ; tim clunan , et al . , class .
* 21 * , 3447 ( 2004 ) .
[ gr - qc/0402044 ] dawood kothawala , t. padmanabhan , sudipta sarkar , phys .
, * d78 * , 104018 ( 2008 ) [ arxiv:0807.1481 ] ; t. padmanabhan , class .
grav . , * 21 * , l1 ( 2004 ) [ gr - qc/0310027 ] ; class.quan.grav . ,
* 21 * , 4485 ( 2004 ) [ gr - qc/0308070 ] ; t.roy choudhury , t. padmanabhan , phys .
, * d 69 * 064033 ( 2004 ) [ gr - qc/0311064 ] .
t. padmanabhan , gen.rel.grav . ,
* 34 * 2029 - 2035 ( 2002 ) [ gr - qc/0205090 ] [ second prize essay ; gravity research foundation essay contest , 2002 ] [ gr - qc/0209088 ] ; mod.phys.letts . * a 17 * , 1147 ( 2002)[hep - th/0205278 ] ; brazilian jour.phys .
( special issue ) 35 , 362 ( 2005 ) [ gr - qc/0412068 ] ; t. padmanabhan , _ gravity : a new holographic perspective _ ( lecture at the international conference on einstein s legacy in the new millennium , dec , 2005 ) int.j.mod.phys .
, * d 15 * , 1659 - 1675 ( 2006 ) [ gr - qc/0606061 ] . for a small sample ,
see e.g. , d. kothawala et al .
, _ phys .
* b 652 * , 338 ( 2007 ) [ arxiv : gr - qc/0701002 ] ; a. paranjape et al .
, _ phys .
rev . _ * d 74 * , 104015 ( 2006 ) [ arxiv : hep - th/0607240 ] ; dawood kothawala , t. padmanabhan , _ phys .
_ , * d 79 * , 104020 ( 2009)[arxiv:0904.0215 ] ; r. g. cai , et al . , _ phys . rev . _ * d 78 * , 124012 ( 2008 ) [ arxiv:0810.2610 ] ; _ phys .
* d 75 * , 084003 ( 2007 ) [ arxiv : hep - th/0609128 ] ; m. akbar and r. g. cai , _ phys .
_ , * b 635 * , 7 ( 2006 ) [ arxiv : hep - th/0602156 ] ; _ phys . lett . _
* b 648 * , 243 ( 2007 ) [ arxiv : gr - qc/0612089 ] ; y. gong and a. wang , phys .
lett . * 99 * , 211301 ( 2007 ) [ arxiv:0704.0793 [ hep - th ] ] ; s. f. wu , g. h. yang and p. m. zhang , [ arxiv:0710.5394 ] ; s. f. wu , b. wang and g. h. yang , nucl .
b * 799 * , 330 ( 2008 ) [ arxiv:0711.1209 ] ; s. f. wu , b. wang , g. h. yang and p. m. zhang , [ arxiv:0801.2688 ] ; j. zhou et al.,arxiv:0705.1264 ; r - g cai , l - m cao , [ gr - qc/0611071 ] ; m. akbar , [ hep - th/0702029 ] ; xian - hui ge , hep - th/0703253 ; a. sheykhi et al.,hep - th/0701198 ; g. allemandi et al.,gr - qc/0308019 .
perlmutter et al . , astrophys . j. ( 1999 ) * 517*,565 ; a.g .
reiss et al . , astron . j. ( 1998 ) , * 116*,1009 ; j. l. tonry et al . ,
apj , ( 2003 ) , * 594 * , 1 ; b. j. barris , astrophys.j . , * 602 * ( 2004 ) , 571 ; a. g.reiss et al .
, astrophys.j . * 607 * , ( 2004 ) , 665 .
h.k.jassal et al . , _ phys.rev .
_ * d 72 * , 103503 ( 2005 ) [ astro - ph/0506748 ] ; [ astro - ph/0601389 ] ; t. padmanabhan , t .
roy choudhury , _ mnras _ * 344 * , 823 ( 2003 ) [ astro - ph/0212573 ] ; t. roy choudhury , t. padmanabhan , _ astron.astrophys . _ * 429 * , 807 ( 2005 ) , [ astro - ph/0311622 ] ; s. nesseris , l.perivolaropoulos , jcap * 0702*,025 ( 2007 ) ; y. wang , p. mukherjee , phys . rev . *
d 76 * , 103533 ( 2007 ) .
t. padmanabhan , _
gravity s immunity from vacuum : the holographic structure of semiclassical action _ , [ third prize essay ; gravity research foundation essay contest , 2006 ] _
gen.rel.grav_. , * 38 * , 1547 - 1552 ( 2006 ) ; t. padmanabhan current science , * 88*,1057 , ( 2005 ) , [ astro - ph/0411044 ] ;
_ int.j.mod.phys._ , * d 15 * , 2029 ( 2006 ) [ gr - qc/0609012 ] .
t. padmanabhan class.quan.grav . , * 22 * , l107-l110 , ( 2005 ) [ hep - th/0406060 ]
. for earlier attempts in similar spirit , see t. padmanabhan , _
class.quan.grav . _
* 19 * , l167 ( 2002 ) , [ gr - qc/0204020 ] ; d. sorkin , _ int.j.theor.phys . _ * 36 * , 2759 ( 1997 ) ; for related work , see volovik , g. e. , gr - qc/0405012 ; j. v. lindesay et al . , astro - ph/0412477 ; y. s. myung , hep - th/0412224 ; j.d.barrow , gr - qc/0612128 ; e.elizalde et al . , hep - th/0502082 .
|
i describe the conceptual and mathematical basis of an approach which describes gravity as an emergent phenomenon . combining the principle of equivalence and the principle of general covariance with known properties of local rindler horizons , perceived by observers accelerated with respect to local inertial frames ,
one can provide a thermodynamic re - interpretation of the field equations describing gravity in any diffeomorphism invariant theory .
this fact , in turn , leads us to the possibility of deriving the field equations of gravity by maximising a suitably defined entropy functional , without using the metric tensor as a dynamical variable .
the approach synthesizes concepts from quantum theory , thermodynamics and gravity leading to a fresh perspective on the nature of gravity .
the description is presented here in the form of a dialogue , thereby addressing several frequently - asked - questions .
|
muons detected at ground level arise mainly as decay products of charged mesons .
these mesons are created abundantly in hadronic showers , most of which are caused by primary cosmic ray protons and nuclei interacting inelastically with the nuclei of the atmosphere . a small fraction of muons , however , have their first origin in photonuclear reactions induced by primary or secondary cosmic gamma rays .
several authors have performed analytical or monte carlo calculations of the muon flux produced in gamma showers ( see references [ 10 - 21 ] of @xcite ) .
most of these calculations were one dimensional and all of them referred to gamma energies much larger than 10 tev . despite their relative scarcity compared to the large background of muons from hadron - generated showers , muons originating from primary gamma ray interactions are important for ground - based high statistics cosmic ray experiments which are sensitive to energies @xmath0 tev such as milagro @xcite and grand @xcite .
cosmic gamma rays , unlike charged hadrons , are unaffected by the earth s and galactic magnetic fields and their direction points directly to the location of their source .
therefore , an accumulation of muon directions around a particular angle ( and also at a particular time in the case of pulsed sources or gamma ray bursts ) carries direct information about the location of the source and the neutrality of the primary causing the excess .
the ability to distinguish these muons from background muons depends upon statistics , the strength and energy spectrum of the emitting source , the angular resolution of the experiment , and the degradation of the angular resolution due to all the physical processes which occur between the primary and the muon at detection level .
this last point is the focus of this calculation .
knowledge of this angular resolution is important in order to determine a detection window which is large enough to include a good fraction of the gamma signal and , at the same time , small enough to minimize the large uniform background of hadronic origin .
the information about the primary gamma direction is degraded by multiple physical processes so the final measured direction of the muon is no longer collinear with the primary which originated the muon s ancestors . in particular ,
the direction resolution is affected by the angle of production of the mesons in the primary interaction ( and then in subsequent hadronic interactions of these mesons to create additional mesons ) , the decay angle of the muon relative to its parent meson , and the lengths of the charged particle paths including that of the detected muon at ground level . along these path lengths
there is coulomb scattering and deflection in the earth s magnetic field for the trajectories of the mesons and the muon ( usually most important for the muon which typically has the longest flight path and the lowest momentum . )
in addition , these effects depend upon the energy of the particles involved and on the varying properties of the atmosphere ( pressure , temperature . )
the muons are produced by mesons of second , third , or further generations .
the number of generations typically increases with increasing primary gamma energy and each new generation involves an additional production angle which tends to increase the angular disparity between muon and primary because of more steps in a random walk type process .
an opposing effect at higher energies is that all angular distributions are more forward - peaked thus decreasing the angular disparity ; the resulting effect is a combination of the two . the purpose of the present study is to estimate the influence of the various effects on the ultimate angular resolution which can be obtained utilizing ground level muons .
the complexity of the problem requires detailed analyses which can only be obtained by a full three - dimensional monte carlo ( mc ) calculation .
in addition , since only a small fraction of all primary photons gives rise to an event of interest , it is desirable to resort to statistical variance reduction techniques to bring the computing task within manageable limits .
the monte carlo program fluka @xcite provides an accurate and well - tested description of the hadronic and electromagnetic interactions with all the relevant physical effects , and a set of effective statistical tools to accelerate convergence of the results .
the program has been used to calculate the geometrical properties of secondary muons at sea level arising from primary gamma cosmic rays with energies from 1 gev to 10 tev interacting in the earth s atmosphere . in this energy range
the cross section for direct muon pair production is much smaller than that for hadroproduction and has been neglected .
this range has been shown in @xcite to be the interval which produces the maximum secondary muon flux at ground level for primary differential gamma spectra with spectral indices larger than 2.4 ( a smaller index would yield a harder spectrum and give a maximum secondary muon flux arising from primary energies beyond those considered here . ) in this paper , only incident cosmic gamma rays at perpendicular ( vertical ) incidence on top of the atmosphere are considered .
ground - based experiments often choose angles close to normal for reasons of simplicity and because the detection rate is a maximum , the rate dropping rather rapidly as the angle deviates markedly from normal . for small angles from normal
, the angular resolution results will probably not differ significantly from those presented in this paper .
in addition to the angular correlation of secondary muons at ground level , this study provides also additional information such as the distribution of heights at which the muons originate , the muon s kinetic energy distribution at the ground and production level , the radial distribution of the muons , and the number of generations which preceded that of the muon .
the 2000 version of the fluka monte carlo code was used to simulate the electromagnetic and hadronic particle cascades induced by primary gamma rays in the atmosphere .
the powerful biasing capabilities of the code allowed a simulation of the full three - dimensional shower in a single run from the top of the atmosphere down to ground level .
the physical models implemented in fluka which are relevant for the present study as well as various variance reduction techniques were already discussed in @xcite .
there it was emphasized that these models were validated by comprehensive comparisons of fluka results with experimental data obtained mostly at accelerators but also in recent cosmic ray studies @xcite .
nevertheless , the simulation of an atmospheric shower and , in particular , of its muon component is a relatively new area of application of fluka .
additional examples showing that the code allows remarkably accurate predictions in this area are given in section [ fluka-datcmp.subsec ] .
section [ fluka-simdet.subsec ] contains some details of the present gamma ray shower simulations which supplement those discussed in @xcite ; this reference contains a more complete description .
no data exist for the muon component of gamma ray - induced showers to which fluka results could be directly compared . on the other hand , data on muon flux and
energy spectra have become available recently for atmospheric showers induced by cosmic ray protons and nuclei .
these data were mostly obtained with spectrometers in balloon - borne experiments and cover the whole range of atmospheric depths from the top of the atmosphere down to sea level .
fortunately , the mechanisms and models for muon production and transport in gamma ray - induced showers are largely the same as those in pure hadronic air showers . in both cases ,
the production of the pions which eventually decay into muons is described by the dual parton model ( dpm ) @xcite .
the only difference is that in the former case the photon is assumed to fluctuate first into a quark - antiquark state which then interacts hadronically , identical to a meson .
this so - called vector meson dominance model ( vmd ) for the hadronic photon fluctuation is a well - established concept ( see @xcite and references therein ) and its application together with the dpm to hadronic interactions of photons has been proven to be very successful ( see , for example , @xcite ) .
a good description of muon production from proton- and nuclei - induced air showers should therefore give confidence that the corresponding results obtained with fluka for photon - induced air showers are reliable as well .
in the following a few examples are given .
a critical aspect of muon production in the atmosphere is the dependence of the total muon flux on the height above sea level .
this dependence is commonly expressed as a function of depth in the atmosphere ( in g/@xmath1 of air ) which is a measure of the amount of air penetrated by the cascade .
it has been measured by the caprice experiment which was able to distinguish between positively and negatively charged muons .
results obtained with fluka for negative muons are compared with caprice data @xcite in fig .
[ mu - capdep ] .
the different curves correspond to the muon flux calculated for different intervals of the muon momentum .
to correspond to the average geometrical response of the caprice apparatus , only muons with a polar angle of less than @xmath2 with respect to the vertical ( i.e. , vertical muons ) were scored .
the simulations are based on sophisticated models and methods for the sampling of the identity , energy , and angle of the hadronic primary from the distributions of galactic cosmic rays on top of the atmosphere at the time of the measurements .
more details can be found in @xcite . as can be seen from the figure
, fluka gives a good description of the negative muon flux for all depths .
similar agreement exists for positive muons ( not shown here ) .
energy spectra of muons were recorded by the same experiment at ground ( i.e. , at an atmospheric depth of 886 g/@xmath1 ) @xcite .
fluka results for positive and negative muons are compared to these data in fig .
[ mu - capspc25&ratio]a .
again , good agreement between simulation and data is found . finally , the ratio of the fluxes of positively and negatively charged muons is compared to caprice data @xcite in fig .
[ mu - capspc25&ratio]b . data and calculations are again for vertical muons at a depth of 886 g/@xmath1 .
the ratios are plotted as a function of the muon energy .
the above comparisons of fluka results and experimental data together with those published in earlier studies @xcite indicate that the application of fluka to cosmic gamma ray showers should give reliable predictions as well . in order to study the dependence of the shower properties on the primary photon energy , the showers were calculated for monoenergetic photons impinging vertically on top of the atmosphere ( taken to be 80 km above sea level ) . in total 9 sets of simulations
were performed for the following primary energies : 1 , 3 , 10 , 30 , 100 , 300 , 1000 , 3000 , and 10000 gev . as in the earlier study @xcite the atmosphere
was approximated by 50 layers with a constant density in each layer and with layer - densities decreasing exponentially with increasing altitude .
it has been verified @xcite that this approximation does not affect the conclusions drawn in this study .
the geometry is described in a right - handed orthogonal system with its origin at the intersection of the shower axis with sea - level elevation , @xmath3 points to the center of the earth and @xmath4 and @xmath5 point north and east , respectively .
most of the results discussed throughout the paper were obtained without the effect of the magnetic field of the earth included . however , for two primary energies ( 10 gev and 1 tev ) additional simulations were performed which included the magnetic field of three different geographical locations ( see sec .
[ angular-magfld.subsec ] ) .
the use of several variance reduction ( biasing ) techniques was essential to obtain results with reasonable statistical significance .
they included leading particle biasing at each electromagnetic interaction , biasing of the photon mean - free - path with respect to photonuclear interactions , biasing of the decay length of charged mesons , and particle splitting at the boundaries of different air layers .
each of these biasing techniques alters the statistical weight of a particle in the cascade .
a muon as produced and transported in the monte carlo simulation and which carries a certain statistical weight ( hereafter called `` mc muon '' ) contributes to distributions , yields , etc .
, of actual , i.e. , measurable muons with a probability equal to its weight .
the use of biasing techniques would not be appropriate to study fluctuations within the same shower which , however , is not the aim of this study .
more details on the biasing techniques can be found in @xcite . for each mc
muon reaching sea level ( detection level ) the following information was recorded in a file for later analysis : 1 . information on the muon at detection level : muon charge , lateral coordinates @xmath6 with respect to shower axis , direction cosines with respect to the @xmath4 and @xmath5 axes , kinetic energy , statistical weight of the muon , and number of the event ( i.e. , primary cosmic ray photon ) which produced this muon .
2 . information on the muon at its production vertex ( i.e.
, meson decay vertex ) : lateral coordinates @xmath6 with respect to shower axis , direction cosines with respect to the @xmath4 and @xmath5 axes , height @xmath3 , kinetic energy , statistical weight of the muon , and identity of the decaying meson .
3 . information on the grandparent at the parent production vertex : particle identity , lateral coordinates @xmath6 with respect to shower axis , direction cosines with respect to the @xmath4 and @xmath5 axes , height @xmath3 , kinetic energy , statistical weight of the particle , and generation number .
4 . information on the photonuclear interaction vertex preceding the hadronic cascade in which the muon at detection level has been created :
lateral coordinates @xmath6 with respect to shower axis , direction cosines with respect to the @xmath4 and @xmath5 axes , height @xmath3 , energy , statistical weight of the photon , and generation number of the photon the generation of a particle increases with each sampled discrete interaction ( electromagnetic or hadronic ) , i.e. , the primary photon is generation `` 1 , '' the generation of the electron and positron after the first pair production process would be `` 2 , '' etc .
delta ray production and coulomb scattering are not considered to increase the generation number .
a summary of the simulated particles cascades for each primary gamma ray energy is given in table [ no-of-muons.tab ] .
in the following , all results refer to the sum of positive and negative muons . the number of histories ( number of primary gamma rays ) , @xmath7 , calculated for each primary energy decreases with energy in order to obtain uniform statistical significance on the muons which reach ground level for the different primary energies . at the lowest energy ( 1 gev )
the muon flux per gamma at sea level is very small so that even with a significant amount of biasing in the simulations it is difficult to reach adequate statistics .
less emphasis was therefore put on the simulation at 1 gev .
the column @xmath8 gives the total number of muons reaching detection level for the given number of primary photons @xmath9 .
the predictions for the muon multiplicity at sea level per incident primary gamma ray @xmath10 are listed in the last column .
similar values were already reported in @xcite , however with smaller statistical significance than those in table [ no-of-muons.tab ] .
@xmath8 & @xmath10 + [ 0.5ex ] 1 & 4542000 & @xmath11 & @xmath12 + 3 & 26000000 & 117.7 & @xmath13 + 10 & 2250000 & 1191 & @xmath14 + 30 & 1260000 & 3843 & @xmath15 + 100 & 800000 & 11351 & @xmath16 + 300 & 324445 & 17259 & @xmath17 + 1000 & 143131 & 31709 & @xmath18 + 3000 & 27040 & 21954 & @xmath19 + 10000 & 18000 & 58874 & @xmath20 + [ 0.5ex ]
in this section the effect of the earth s magnetic field is not included ; section [ angular-magfld.subsec ] has a separate discussion of this added effect which depends on the geographical location of interest .
[ dndth ] summarizes the deviations of the angle of the muons which reach sea level from the primary gamma direction ( vertical angle ) .
it shows the distribution of the angle @xmath21 which is the muon s angle with respect to the @xmath3 axis projected onto the @xmath22 plane ( north - down plane ) .
this angle is defined as @xmath23 where @xmath24 and @xmath25 are the direction cosines of the detected muon .
negative projections have been reflected upon the positive values for @xmath21 since positive and negative angles would be symmetric in the absence of the magnetic field . also , the distributions in the @xmath26 plane ( east - down plane ) are similar to those shown in fig .
[ dndth ] . since the angle of the primary gamma with respect to the @xmath3 axis is zero ( vertically incident primaries )
, any deviation of the muon direction from zero degrees is due to the various interaction and transport processes occurring in the cascade between the primary gamma ray and the muon which reaches sea level .
thus the results shown in fig . [ dndth ] represent the correlation between the direction of the detected muon and the direction of the primary gamma ray .
all distributions in fig .
[ dndth ] are normalized to unit area to compare their shapes .
the shape of the distributions for primary energies above about 30 gev varies rather slowly .
therefore , only two histograms ( 100 gev and 10 tev ) are shown for this energy range . at lower energies the shape of the distributions changes significantly as can be seen from the histograms for 3 and 10 gev .
the distributions become wider with decreasing primary energy . at low primary gamma ray energies ,
the muons have , on average , lower energies which cause larger angular deviations from the primary due to kinematics and scattering effects which are larger for the lower energies .
this is illustrated in fig .
[ dndth-10gtev ] for primary energies of 10 gev and 10 tev .
an arbitrary cut of 2 gev was applied to the angular distributions so that contributions from muons with energies below and above this cut could be seen separately along with the total .
two measures of the widths of these distributions as a function of the energy of the incident primary cosmic gamma ray are given in table [ angwidths.tab ] .
the width parameters are : the half - width at [ 0.5ex ] @xmath27 ( gev ) & @xmath28 & @xmath29 gev & @xmath30 gev & @xmath31 gev + [ 0.5ex ] 1 & 31.0 & @xmath32 & @xmath32 & @xmath32 + 3 & 8.20 & @xmath32 & @xmath32 & @xmath32 + 10 & 3.23 & 2.95 & 2.57 & 1.87 + 30 & 1.87 & 1.73 & 1.59 & 1.35 + 100 & 1.12 & 1.04 & 0.95 & 0.81 + 300 & 0.92 & 0.85 & 0.77 & 0.66 + 1000 & 0.91 & 0.82 & 0.74 & 0.63 + 3000 & 0.98 & 0.89 & 0.78 & 0.66 + 10000 & 1.05 & 0.94 & 0.82 & 0.67 + [ 0.5ex ] & + [ 0.5ex ] 1 & 28.3 & @xmath32 & @xmath32 & @xmath32 + 3 & 7.36 & @xmath32 & @xmath32 & @xmath32 + 10 & 3.83 & 3.06 & 2.46 & 1.66 + 30 & 3.07 & 2.40 & 1.95 & 1.45 + 100 & 2.87 & 2.15 & 1.69 & 1.21 + 300 & 3.04 & 2.18 & 1.68 & 1.16 + 1000 & 3.32 & 2.31 & 1.75 & 1.19 + 3000 & 3.62 & 2.42 & 1.81 & 1.23 + 10000 & 4.05 & 2.57 & 1.89 & 1.27 + [ 0.5ex ] half - maximum - height ( hwhm ) and the angular half - width containing 68% of the muons , i.e. , the same number of events that a 1@xmath33 cut would include if the distributions had a gaussian shape ( which they do nt have , because of the long tails ) . both of these parameters provide a measure of the distribution widths which minimizes the distorting effect of a tail .
in addition , the values for various cuts on the kinetic energy of the detected muons ( 1 , 2 , and 4 gev ) are presented .
the cutoff energy results for primary energies of less than 10 gev were omitted due to lack of statistics . for the hwhm calculation , the maximum height is assumed to be the value in the first bin of the corresponding histogram , i.e. , the average value in the interval @xmath34 degrees .
the @xmath4 and @xmath5 rectangular coordinate system was chosen to represent the results of the muon s angular distribution as the corrections for the additional deflection due to the earth s magnetic field differ in these two directions .
in addition , @xmath6 is the natural coordinate system for grand .
for some experiments a space angle is more natural .
if we define @xmath35 , then the space angle resolution ( @xmath36(hwhm ) ) can be estimated directly from the numbers in table [ angwidths.tab ] by multiplying the @xmath37(hwhm ) values by @xmath38 due to the symmetry of @xmath4 and @xmath5 in the absence of a magnetic field . as can be seen in this table :
( i ) as the primary energy decreases below 10 gev , the width of the angular distribution increases dramatically , ( ii ) as the energy rises above 300 gev , there is only a small , gradual increase in the width of the angular distribution , and ( iii ) the width of the distribution narrows as the muon s cutoff energy is raised ; i.e. , the angular resolution becomes better by eliminating the lower energy muons which have , on average , poorer angular resolution . in the preceding section ,
angular distributions were given for all muons regardless of their lateral distance from the shower axis .
however , as the distance from the shower axis increases , the angle is systematically biased away from the shower axis .
[ dndthx-10gev1tevx ] shows the correlation between the projected angle @xmath21 and the @xmath4 coordinate ( i.e. , the northward distance from the shower core ) of the muon position at sea level for primary gamma ray energies of 10 gev and 1 tev .
the angular distributions are given for different intervals in @xmath4 and each distribution is normalized to unit area . here ,
negative values of @xmath21 have not been reflected onto positive values and are explicitly shown .
as expected , the peak of the @xmath21 distributions are shifted toward positive values as @xmath4 becomes more positive and the distributions become wider as @xmath4 increases .
both effects depend on the energy of the primary gamma ray and are more pronounced at higher energies .
hence , the angular distributions narrow significantly if they are limited to muons within a certain interval around the shower axis .
information on this correlation between the angle and the lateral distance of the muon from the shower axis allows experiments to improve the angular resolution by applying cuts to the data , e.g. , in @xmath4 .
if the available statistics demand using all available muons , their angle at large @xmath4 s ( @xmath5 s ) could be corrected by the most probable angle at that @xmath4 ( or @xmath5 ) to obtain a better estimate of the primary s direction .
this correlation can also be used to infer the height from which the muon originated , a parameter difficult to obtain experimentally .
if one extrapolates the muon direction backward ( upward ) , the intersection point with the centroid of the shower ( which can be obtained from the more numerous electrons in the same shower ) , is an estimator for the height where the muon was created .
however , deviations of the muon production vertex from the centroid and the physical processes which alter the muon track between the production vertex and sea level obscure this information somewhat . for this reason it is not possible to study this question on a muon - by - muon basis .
for example , tracks with projected angles close to zero which exist at almost all @xmath4-distances from the origin would yield , upon extrapolation , an infinite height - of - origin .
this problem can be circumvented by finding the mean @xmath4 within an interval of @xmath4 and dividing this mean by the corresponding mean of @xmath39 .
the averaging serves another purpose as well : the randomness of the various physical processes , such as scattering , tend to cancel in the averaging .
the results of the preceding sections can be combined to obtain the expected angular resolution for a spectrum of primary gamma rays .
various steps in the calculation are shown in fig .
[ flxgamma]a . here
, a differential energy spectrum @xmath40 with a spectral index of @xmath41 is assumed ; this @xmath42 corresponds to the average of the spectral indices reported in the third egret catalog @xcite .
folding @xmath43 with the number of muons reaching sea level per primary photon , @xmath44 , yields the number of muons at sea level as function of the primary gamma ray energy , @xmath45 , with @xmath46 .
note that the differential spectra in fig .
[ flxgamma]a are multiplied by @xmath47 .
the flatness of the differential muon flux per gamma at sea level ( @xmath48 ) above 10 gev signifies that the muons originate rather uniformly from a broad range of primary energies . below 10 gev ,
the muon flux decreases steeply as @xmath44 rapidly approaches zero .
harder spectral indices ( @xmath49 ) would enhance the muon flux from higher primary energies and , conversely , softer indices would enhance lower energies . furthermore , in fig .
[ flxgamma]a the half - widths at half - maximum - height @xmath37 ( hwhm ) ( given in table [ angwidths.tab ] in the `` @xmath28 '' column ) are plotted .
these widths are multiplied by the differential muon flux which is also shown in fig . [ flxgamma]a . as can be seen , the larger angular widths at the lower energies ( @xmath50 gev ) do not contribute due to the small sea level muon flux at these energies .
the average angular resolution ( i.e. , average angular width ) in the north - down plane , @xmath51 , for a differential gamma ray spectrum with an index @xmath42 is obtained from @xmath52 and shown in fig .
[ flxgamma]b . as can be seen
, the resolution is about constant ( @xmath53 ) up to a spectral index of 2 and then rises for larger indices due to the increasing importance of the wider angular widths at low energy .
the magnetic field of the earth deflects positively ( negatively ) charged particles primarily eastward ( westward ) causing the angular distributions in the projected angle @xmath54 ( i.e. , projected onto the east - down plane ) to become wider .
the size of the effect depends on the strength and direction of the field and , therefore , varies with geographic location . to obtain a measure of the magnitude of the added angular widths and how it depends on geographic location ,
three different locations are considered and summarized in table [ magfld.tab ] .
in order to estimate the maximum possible deflection , a location of maximum northward field component was chosen .
due to the offset of the magnetic field axis from the earth s center , this location is at about @xmath2n , @xmath55e ( below referred to as `` maximum '' ) .
in addition , two locations with intermediate magnetic fields were chosen : @xmath56n , @xmath57w ( grand ) and @xmath58n , @xmath59w ( milagro ) .
the magnetic field components for these locations were obtained from the latest revision of the international geomagnetic reference field ( igrf ) @xcite and are listed in table [ magfld.tab ] . here , the components of the magnetic field , b , are given in nanotesla ( nt ) and the components @xmath60 are directed north , east , and toward the center of the earth , respectively . for uniformity of comparison ,
all values are for sea level elevation .
( nt ) & @xmath61 ( nt ) + [ 0.5ex ] ( 1 ) grand , @xmath56n , @xmath57w & 18592 & -1361 & 52547 + [ 0.5ex ] ( 2 ) milagro , @xmath58n , @xmath59w & 22751 & 4141 & 46241 + [ 0.5ex ] ( 3 ) maximum , @xmath2n , @xmath55e & 41407 & -211 & 1326 + [ 0.5ex ] fluka is able to transport particles in arbitrary magnetic fields .
however , since the variation of the field in the atmosphere with height above sea level and with time is only minor ( about 4% ) , for simplicity the magnetic field used in the calculation for a certain location was assumed to be constant throughout the whole geometry .
in addition , although the milagro experiment is located at a height of 2170 m above sea level , all results reported for that location refer to sea level as this allows a direct comparison to the results at the other locations .
the angular distributions at sea level for the three magnetic field conditions and for primary gamma ray energies of 10 gev and 1 tev are shown in fig .
[ dndth-10gev1tevmag ] .
in addition , the distributions obtained without magnetic field are given .
as can be seen , for a 10 gev gamma primary , the effect of the earth s magnetic field is considerable and changes with geographic location ; as the northward component of the magnetic field increases , the width increases .
however , at 1 tev , there is only a slight widening of the distributions with increasing the north component of the field . to within the statistical uncertainties of the calculations , the total number of muons which reach sea level per primary gamma ray is independent of the effect of the earth s magnetic field for the two energies studied in this section ( 10 gev and 1 tev ) .
the half - widths of the angular distributions in the north - down ( @xmath21 ) and east - down planes ( @xmath54 ) containing 68% of the muons are presented for the different locations in table [ angwidthsmag.tab ] .
again , the [ 0.5ex ] & ( gev ) & n - d . &
e - d . & e - d.(b ) & n - d . &
e - d . + [ 0.5ex ] n.f .
& 10 & 3.83 & 3.84 & 3.41 & 3.06 & 3.06 & 2.46 & 2.45 & 1.66 & 1.64 + & 1000 & 3.32 & 3.32 & 3.03 & 2.31 & 2.30 & 1.75 & 1.74 & 1.19 & 1.19 + [ 0.5ex ] ( 1 ) & 10 & 3.85 & 4.41 & 3.38 & 3.06 & 3.53 & 2.47 & 2.88 & 1.67 & 2.03 + & 1000 & 3.31 & 3.48 & 3.0 & 2.30 & 2.42 & 1.74 & 1.84 & 1.19 & 1.26 + [ 0.5ex ] ( 2 ) & 10 & 3.89 & 4.73 & 3.40 & 3.09 & 3.83 & 2.48 & 3.15 & 1.67 & 2.25 + & 1000 & 3.33 & 3.61 & 3.02 & 2.32 & 2.52 & 1.75 & 1.91 & 1.19 & 1.30 + [ 0.5ex ] ( 3 ) & 10 & 3.86 & 6.44 & 3.38 & 3.05 & 5.28 & 2.44 & 4.44 & 1.64 & 3.34 + & 1000 & 3.32 & 4.19 & 3.02 & 2.31 & 2.93 & 1.74 & 2.24 & 1.20 & 1.56 + [ 0.5ex ] widths are given for distributions containing only muons above certain kinetic energy thresholds ( including zero threshold , or no cut ) .
in addition , the corresponding values from the calculations without the effect of the magnetic field are given .
the widths in the north - down plane are much less affected than in the east - down plane which contains the dominant effect of the magnetic deflection .
this effect increases with increasing values of @xmath62 or decreasing values of muon momentum .
in addition to the widths of the muon s sea level angular distributions , table [ angwidthsmag.tab ] also shows the values at their production vertex ( birth , e - d.(b ) ) .
interestingly , these values are almost constant for the different field conditions .
the effect of the magnetic field on the angular distributions is therefore mainly the deflection of the muon over its longer path and not the deflection of the parent pions or other charged particles preceding in the cascade due to their shorter path lengths and higher momenta as compared to the muons .
in addition , the dominant factor in the final width ( except for the highest magnetic field value ) is the muon s angular distribution at its production ( birth ) point .
air pressure and temperature changes cause small variations in the density profile of the atmosphere and thus affect the muon flux at sea level . in order to obtain a quantitative estimate of these two effects ,
the following cases were studied : ( i ) the air density in each layer was increased by 3% while all other variables ( geometry , cutoffs , etc . )
were kept constant and ( ii ) the height above sea level of each slab boundary was raised by 5% and the corresponding density decreased in the same proportion such that the total thickness of air ( in g/@xmath1 ) above the surface of the earth remained constant .
the former modification simulates an air pressure increase ( @xmath63 ) whereas the latter simulates an overall 5% increase in absolute temperature ( @xmath64 ) of the atmosphere .
since muons are produced at atmospheric heights far above those which determine the weather at the surface of the earth , the mean @xmath64 which is involved in this calculation has little correlation with the temperature at the earth s surface .
thus the temperature effect calculated here is not expected to correlate with the earth s surface temperature but would require temperature measurements of the earth s atmosphere averaged from 0 to 20 km above the earth s surface as obtained , for example , in weather balloon measurements .
the results presented in this section allow an investigation of , or corrections for , the muon flux due to small variations in the atmospheric pressure and height - averaged temperature .
the results for the total muon yield per primary gamma ray are summarized for two energies ( 10 gev and 1 tev ) in table [ temp-pres.tab ] . in all ( gev ) & ref .
atmosphere & 3% density increase & 5% temperature increase + [ 0.5ex ] 10 & @xmath14 & @xmath65 & @xmath66 + 1000 & @xmath67 & @xmath68 & @xmath69 + [ 0.5ex ] cases the muon yields decreased by 3% to 6% , the decrease being more pronounced at the lower energy . as for increased pressure ,
a higher density increases the interaction probability of the pions and thus decreases the probability of decay , i.e. , fewer muons are produced .
as well , the added thickness of air which the muons must traverse is a barrier which some will fail to overcome . on the other hand , raising the boundaries of the layers and reducing their density causes two counterbalancing effects : ( i ) the pions propagate in a less dense medium and thus have a greater probability to decay rather than interact thus producing more muons , and ( ii ) the muons must propagate longer distances resulting in a greater probability that they will decay before reaching ground level .
the calculated temperature dependence of the muon multiplicity ratio suggests that the enhanced muon decay probability dominates at these energies and is the more dominant at the lower energy .
information on the muons at sea level and on their ancestors ( i.e. , the parent , grandparent , etc .
, see sec . [ fluka-simdet.subsec ] ) allows a more detailed study of the properties of these ancestors .
table [ decaymesons.tab ] shows which parent contributes through its decay to the detected muons for the different primary gamma ray energies .
@xmath70 & k@xmath71 & k@xmath72 & neutral kaons + [ 0.5ex ] 1 & 0.106 & 0.894 & 0.0 & 0.0 & 0.0 + 3 & 0.495 & 0.485 & 0.020 & 0.0 & @xmath73 + 10 & 0.492 & 0.489 & 0.011 & 0.007 & @xmath74 + 30 & 0.482 & 0.482 & 0.019 & 0.014 & @xmath75 + 100 & 0.478 & 0.477 & 0.022 & 0.018 & @xmath76 + 300 & 0.477 & 0.476 & 0.023 & 0.019 & @xmath77 + 1000 & 0.475 & 0.476 & 0.024 & 0.019 & @xmath78 + 3000 & 0.476 & 0.475 & 0.025 & 0.020 & @xmath79 + 10000 & 0.474 & 0.477 & 0.024 & 0.020 & @xmath78 + [ 0.5ex ] at the lowest energy ( 1 gev ) about 89% of the muons are produced in decays of negative pions and 11% in decays of positive pions .
this asymmetry can be explained by the different interaction cross sections at low energy of pions of either charge and by the fact that these pions are mainly produced in secondary interactions of neutrons ( see below ) .
correspondingly , the probability for positive pions to interact instead of decay is larger than for negative pions . above 3
gev this asymmetry disappears and decaying kaons begin to contribute to the sea level muon flux .
the latter contribution increases with energy and amounts to about 5% at 10 tev .
it is interesting to note that at all energies fluka predicts the relative contribution of a positive kaon parent to be always larger than that of a negative kaon by about 20% .
this effect is due to the properties of the dpm describing inelastic hadronic interactions within fluka . in particular , it is a feature of the reggeon contribution which describes particle production by one quark - diquark string stretched between a valence quark of the fluctuating photon and a diquark of a target nucleon . in this picture ,
kaon production involves the creation of a @xmath80 quark - antiquark pair and , in case of negative kaons , also the creation of an @xmath81 pair . on the other hand ,
positive kaons can be readily formed also by a @xmath82-quark of the fluctuating photon leading to the observed asymmetry . the fractional contribution to the muon s grandparent
is given in table [ mother.tab ] .
the calculations for 1 gev show distinctly different features .
about 97% of the muons at sea level originate from mesons produced in interactions of neutrons in the close vicinity of the detector . at all other energies photoproduction
dominates the picture with a contribution decreasing with energy from 99% at 3 gev to 60% at the highest energy ; the remaining fraction is mainly from pions . and
@xmath83 & n and @xmath84 & @xmath85 & @xmath70 + [ 0.5ex ] 1 & @xmath86 & 0.023 & 0.97 & @xmath87 & @xmath88 + 3 & 0.99 & @xmath89 & @xmath90 & @xmath91 & @xmath92 + 10 & 0.99 & @xmath93 & @xmath94 & @xmath95 & @xmath96 + 30 & 0.97 & @xmath97 & @xmath98 & @xmath99 & @xmath100 + 100 & 0.91 & 0.014 & 0.015 & 0.026 & 0.027 + 300 & 0.83 & 0.025 & 0.026 & 0.056 & 0.056 + 1000 & 0.73 & 0.037 & 0.041 & 0.091 & 0.090 + 3000 & 0.65 & 0.049 & 0.049 & 0.12 & 0.12 + 10000 & 0.60 & 0.054 & 0.057 & 0.13 & 0.14 + [ 0.5ex ] the distributions of the generation number of the grandparent of the sea level muon is shown in fig .
the lower the energy of the primary gamma ray the smaller is the atmospheric shower resulting in a relatively narrow distribution .
this distribution is peaked at generation one for energies below a few hundred gev where the parent is mainly produced in a photoproduction process of the primary photon ( see also table [ mother.tab ] ) . at higher energies secondary hadron interactions
contributes significantly to the production of the parent .
these secondary hadrons are mainly of third or fourth generation and cause a shift of the peak of the total distribution at tev energies .
the large tail at these energies which extends up to more than 100 generations reflects photoproduction processes of secondary photons in the large electromagnetic shower .
the distributions of heights above sea level at which the detected muons were produced are shown for four primary energies in fig .
as expected , at high energy there is a decrease in height with increasing energy of the primary gamma ray because of the logarithmic increase of the shower length .
however , as the energy of the primary gamma ray decreases below 10 gev , the energy of the muon parent is close to the minimum energy required for the muon to penetrate the blanket of air between its production and sea level .
hence , the most probable height of muon production begins to decrease with energies decreasing below 10 gev .
the blip - up near zero height , most pronounced in the 3 and 10 gev distributions , represents the excess contribution from very low energy muons . the radial distributions of the muons at their production vertices and at sea level
are shown in fig .
[ dnda ] for five different gamma ray energies .
the radial distance , @xmath101 , is defined with respect to the shower axis ( @xmath3 axis ) .
the quantity @xmath102 denotes the number of muons per unit area and per primary gamma ray .
the radial distributions of the production vertices are labeled `` production . ''
all distributions extend to more than 10 km . whereas the sea level distributions have a relatively flat shape below @xmath103 km the production vertex distributions are increasing toward the shower axis and exhibit a change in slope or discontinuity at about 500 m which is most pronounced at low primary energy .
this discontinuity indicates that two components with different shapes contribute . in order to further investigate this feature
the muon production vertices were scored separately for those muons which originated from mesons produced in photoproduction processes and those from mesons produced by interactions of nucleons or other hadrons .
the contributions from these components are shown for a primary energy of 10 gev in fig .
[ dnda10gev ] . as can be clearly seen in the production vertex distributions ( fig .
[ dnda10gev]a ) the photoproduction component dominates at small radii and the component due to interacting nucleons constitutes the tail at large @xmath101 .
the difference is smaller at sea level ( fig .
[ dnda10gev]b ) due to multiple scattering of the muons in air .
the kinetic energy distributions of muons at sea level and at the production vertex of these muons are shown in fig .
[ dndekin ] .
note , that the distributions are multiplied by the muon energy @xmath104 in order to enhance possible spectral structures at the higher energies .
whereas the sea level spectra are a smooth function with energy , the distributions of the muon energies at their birth exhibit a two - component structure .
this two - component structure is again due to the superposition of muons from mesons generated in photoproduction processes and those from interacting hadrons .
the two contributions are plotted separately in fig .
[ dndekin10gev ] for 10 gev primary photon energy .
it is clear that muons from photoproduction processes have a higher average energy .
air showers caused by cosmic gamma rays with energies below 10 tev were simulated using the monte carlo code fluka .
the primary gamma rays were assumed to enter the atmosphere vertically .
the reliability of fluka predictions in this energy region was explored by comparing its predictions for proton- and nuclei - induced air showers with experimental data ; good agreement was found .
many general properties of muon production in atmospheric gamma ray showers were studied . as expected , most of the muons which reach sea level are decay products of charged pions ; the contribution from kaons increases as the primary energy increases .
below 100 gev primary energy , the decaying mesons are produced mainly in photoproduction processes . at higher energies , hadronic interactions of nucleons and pions
contribute significantly .
variations in atmospheric pressure and temperature were investigated : a 3.0% increase in atmospheric pressure decreases the muon yield at sea level by 4.0% for 10 gev primaries ( 3.2% at 1 tev ) .
similarly , a 5.0% increase in the height - averaged temperature decreases the muon yield by 6.0% ( 3.6% at 1 tev ) .
the main goal of the study was to determine the angular resolution for primary cosmic gamma rays from measurements of secondary muon angles at ground level .
the angular distribution of muons at sea level narrows as the muon energy increases since low energy muons constitute the tails of the angular distributions .
thus , the angular resolution can be improved by eliminating the lower energy muons .
the half - widths at half - maximum - height of the muon s projected angles are @xmath105 above 100 gev but rise to a value of @xmath106 as the primary gamma ray energy is lowered to 10 gev in the absence of the earth s magnetic field deflection .
primary energies below 10 gev essentially do not contribute muons at ground level .
the effects of magnetic bending were considered separately as they depend on several factors ; for example , they are more important in the east - down plane , for low primary energies , low muon energies , and larger north - components of the earth s magnetic field .
the width of the angular distribution can be narrowed if the correlation of @xmath21 ( angle projected onto the north - down plane ) with @xmath4 is taken into consideration ( assuming the location of the core of the shower is measured ) .
this narrowing can be accomplished either by eliminating large @xmath4 and @xmath5 values from the data or by compensating the measured angles knowing their expected mean values versus @xmath4 and @xmath5 as calculated from fluka .
finally , the average muon angular correlation was calculated for a spectrum of primary gamma rays characterized by a spectral index . for a soft spectrum , e.g. a differential spectral index @xmath107 , the average projected angular width ( hwhm ) is @xmath108 ; for a mean value , @xmath109 ,
the width is @xmath110 .
if the spectrum is harder , e.g. @xmath111 , the angular width improves to a constant value of @xmath112 .
corresponding values for a space angle resolution ( @xmath36(hwhm ) ) are @xmath38 larger . as mentioned above , if data with larger @xmath4 and @xmath5 values and/or the lower muon energies are removed , this angular correlation can be improved .
this precise calculation of the angular correlations provides experiments which study the angular location of primary gamma ray sources by measuring muon angles with information on the angular resolution which is difficult to obtain experimentally .
part of this work was supported by the department of energy under contract de - ac03 - 76sf00515 .
project grand is funded through grants from the university of notre dame and private donations . 999 a. fass and j. poirier , phys .
d * 63 * , 036002 ( 2001 ) .
the milagro collaboration , r. atkins _ et al .
_ , nucl .
meth . in phys .
res . a * 449 * , 478 ( 2000 ) .
j. poirier _ et al .
_ , in _ proceedings of the 26th international cosmic ray conference ( icrc ) _ , vol . 5 , p. 304
( 1999 ) , world wide web : http://www.nd.edu/@xmath113grand .
et al . _ , in _ proceedings of the 2nd workshop on simulating accelerator radiation environments , cern 1995 _ , edited by g.r .
stevenson , cern report tis - rp/97 - 05 , p. 158
a. fass _ et al .
_ , in _ proceedings of the 3rd workshop on simulating accelerator radiation environments , kek 1997 _ , edited by h. hirayama , kek proceedings 975 , p. 32
a. fass , a. ferrari , and p. r. sala , _ electron - photon transport in fluka : status _ , to appear in _ proceedings of the international conference on advanced monte carlo for radiation physics , particle transport simulation and applications , monte carlo 2000 _ , lisbon , portugal , 2000 .
a. fass _ et al .
_ , _ fluka : status and perspectives for hadronic applications _ , to appear in _ proceedings of the international conference on advanced monte carlo for radiation physics , particle transport simulation and applications , monte carlo 2000 _ , lisbon , portugal , 2000
. v. patera _ et al .
_ , nucl . instr
. meth . in phys .
a * 356 * , 514 ( 1995 ) .
a. ferrari , t. rancati , and p. r. sala , in _ proceedings of the 3rd workshop on simulating accelerator radiation environments , kek 1997 _ , edited by h. hirayama ( kek proceedings 975 , 1997 ) , p. 165 .
s. roesler , w. heinrich , and h. schraube , rad .
res . * 149 * , 87 ( 1998 ) .
g. battistoni _ et al .
_ , astropart .
* 12 * , 315 ( 2000 ) .
a. capella _ et al .
_ , phys . rep . * 236 * , 225 ( 1994 ) . t. h. bauer , r. d. spital , and d. r. yennie , rev . mod .
phys . * 50 * , 261 ( 1978 ) .
r. engel , z. phys .
c * 66 * , 203 ( 1995 ) .
s. roesler , r. engel , and j. ranft , phys .
d * 57 * , 2889 ( 1998 ) .
m. boezio _ et al .
_ , phys .
d * 62 * , 032007 ( 2000 ) .
s. roesler , w. heinrich , and h. schraube , _ monte carlo simulation of the radiation field at aircraft altitudes _ , in preparation . j. kremer _ et al .
_ , phys .
lett . * 83 * , 4241 ( 1999 ) .
r. c. hartman _ et al .
_ , astrophys . j. suppl
. ser .
* 123 * , 79 ( 1999 ) .
international association of geomagnetism and aeronomy ( iaga ) , division v , working group 8 , _ international geomagnetic reference field - epoch 2000 , revision of the igrf for 2000 - 2005 _ , world wide web : http://www.ngdc.noaa.gov/iaga/wg8/wg8.html
1 . dependence of the negative muon flux produced by hadronic primaries on the depth in the atmosphere shown for different intervals of the muon momentum .
fluka results ( solid lines ) are compared to data obtained by the caprice experiment @xcite ( points ) .
the comparison is shown for different intervals in muon momentum .
the curves and data points for the first five momentum intervals were shifted by constant factors as indicated .
\a ) energy spectra of positive and negative muons at 886 g/@xmath1 .
fluka results ( histograms ) are compared to data obtained by the caprice experiment @xcite .
@xmath114 is the detected muon momentum .
b ) muon charge ratio as a function of kinetic energy measured by the caprice experiment @xcite ( points ) and calculated with fluka ( solid line ) .
3 . angular distribution of muons at sea level .
@xmath21 is the angle of the detected muon projected onto the north - down or @xmath22 plane .
the distributions are given for different primary photon energies and are normalized to unit area .
effects of the earth s magnetic field are not included in this figure .
units of angle are in degrees .
\(a ) angular distribution of muons at sea level for 10 gev and ( b ) 10 tev primary gamma ray energy .
the distributions labeled `` total '' are the same as shown in fig .
[ dndth ] . in addition , the contributions from muons with kinetic energies below 2 gev and above 2 gev are presented .
units of angle are in degrees .
angular distributions of muons at sea level given for different intervals in the @xmath4 coordinate .
results are shown for primary gamma ray energies of ( a ) 10 gev and ( b ) 1 tev .
all distributions are normalized to unit area .
units of angle are in degrees .
\(a ) the figure shows the following quantities as a function of the primary gamma ray energy : the energy spectrum of primary gamma rays ( @xmath43 ) with a shape given by a differential spectral index @xmath109 and arbitrary normalization , the multiplicity of muons at sea level per primary gamma ray ( @xmath44 ) , the result of folding the multiplicity with the primary spectrum ( @xmath115 ; thin solid line with circles ) , the half - widths at half - maximum - height for all muon energies ( @xmath37(hwhm ) ) in degrees , and the result of folding these widths with @xmath45 ( dot - dashed lines with diamond points ) .
the calculated values are joined by lines to guide the eye .
note that all differential fluxes are multiplied by @xmath47 .
( b ) average half - width at half - maximum - height for the angle projected onto the north - down plane as function of the spectral index of the primary spectrum ; the ordinate is in units of degrees .
values for the space angle resolution ( @xmath36(hwhm ) ) would be @xmath38 times larger . 7 .
muon angular distribution at sea level including the effect of the earth s magnetic field for ( a ) 10 gev and ( b ) 1 tev primary gamma ray energy .
the distributions are shown as a function of the muon s angle projected onto the @xmath26 ( east - down ) plane , @xmath54 , and for three different geographic locations : ( 1 ) @xmath56n , @xmath57w ( grand ) , ( 2 ) @xmath58n , @xmath59w ( milagro ) , and ( 3 ) @xmath2n , @xmath55e ( maximum ) .
the magnetic deflection for @xmath21 ( north - down ) is minimal and not shown .
in addition , the distribution obtained without magnetic field ( labeled `` no field '' ) is given .
units of angle are in degrees .
distribution in the number of generations in the shower ( see text ) carried by the particles creating the parent - meson .
fluka results are shown for different energies of the primary photon
distribution of the muon production heights .
results are given for primary photon energies of @xmath116 , and @xmath117 gev , respectively , and are normalized per primary photon .
the top of the atmosphere is at 80 km in the calculations .
radial distribution of the muon production vertices ( i.e. meson decay vertices , dotted histograms ) and of the muons at sea level ( solid histograms ) .
results are given for primary photon energies of @xmath118 and @xmath119 gev , respectively ( from the bottom curve to the top ) .
values have been normalized per primary photon .
radial distribution of the muon production vertices for 10 gev primary photons at normal incidence .
the distributions are given for ( a ) the muon production vertex and ( b ) for sea level .
in addition to the total distribution , the contributions from different `` parents '' of the decaying meson are given .
for example , the histograms labeled `` photons '' show the radial distributions of muons from mesons which were produced by photonuclear interactions .
distributions are normalized per primary photon .
kinetic energy spectra of muons at their production vertices ( dotted histograms ) and at sea level ( solid histograms ) .
values are normalized per primary photon . 13
. kinetic energy spectra of muons from 10 gev primary photons at normal incidence .
the spectra are shown for ( a ) the production vertex and ( b ) for sea level .
in addition to the total spectra which are identical to those shown in fig .
[ dndekin ] , the contributions from different `` parents '' of the decaying meson are given ( similar to fig .
[ dnda10gev ] ) .
|
the fluka monte carlo program is used to predict the distributions of the muons which originate from primary cosmic gamma rays and reach sea level .
the main result is the angular distribution of muons produced by vertical gamma rays which is necessary to predict the inherent angular resolution of any instrument utilizing muons to infer properties of gamma ray primaries .
furthermore , various physical effects are discussed which affect these distributions in differing proportions .
air shower simulation , gamma rays , muons 95.75.pq , 98.70.sa , 98.70.rz , 13.60.le
|
it is widely believed that gamma - ray bursts ( grbs ) and their afterglows are caused by the dissipation of kinetic energy of an ultrarelativistic ejecta , with lorentz factor @xmath3 , releasing from the central engine ( so - called fireball " model , see piran 1999 and van paradijs , kouveliotou & wijers 2000 for detailed reviews ) .
grbs are the most energetic explosive phenomena in astronomy . in the fireball model , the collisions among different shells within the ejecta might be expected to produce a prompt burst ( paczyski & xu 1994 ; rees & mszros 1994 ) , and subsequently the ejecta interacts with its surrounding medium , producing a long - term and broad - band afterglow ( paczyski & rhoads 1993 ; mszros & rees 1997 ) . in the standard afterglow picture ,
the ejecta drives a relativistic blast wave expanding into the surrounding medium , which approaches a self - similar solution ( blandford & mckee 1976 ) after a short time once the swept - up medium attains an energy comparable to the initial energy of the burst .
electrons in the cold medium are heated in the shock front to highly relativistic energy and produce a broad - band afterglow via synchrotron / inverse - compton emission . the predicted emission spectrum and light curve ( mszros & rees 1997 ; sari et al .
1998 ) have met essential successes in describing the behavior of afterglows ( e.g. , vietri 1997 ; waxman 1997ab ; wijers et al .
the recent observations on grb afterglows have lead to numerous numerical calculations to model the observed afterglows ( e.g. , huang et al .
2000a , b ; gou et al .
2001 ; panaitescu & kumar 2001 ; panaitescu 2001 ) .
we re - investigate the dynamical evolution and temporal emission of afterglows by considering three aspects .
( i ) the electrons heated by the shock are always fast cooling when the cooling timescale for these electrons is shorter than the dynamical timescale ( see , e.g. , waxman 1997a ; mszros et al . 1997
; sari et al . 1998 ) , and the afterglow evolution should be in the radiative regime rather than the adiabatic one as long as a significant fraction ( @xmath4 ) of the fireball energy is transferred to newly heated electrons , @xmath5 . this shock decelerates faster than an adiabatic shock , as pointed out by bttcher & dermer ( 2000 ) .
the newly launched hete-2 satellite , due to its rapid and accurate location of grbs , should lead to much more follow - up observations on early x - ray and optical emission , allowing a detailed analysis of the afterglow feature .
( ii ) the inverse - compton emission is important in afterglows ( panaitescu & mszros 1998 ; wei & lu 1998 ; totani 1998 ; chiang & dermer 1999 ; dermer , bttcher & chiang 2000 ; panaitescu & kumar 2000 ; sari & esin 2001 ; zhang & mszros 2001 ) .
when the inverse - compton cooling of electrons dominates the synchrotron cooling , the temporal synchrotron emission is influenced by changing the scaling law of the cooling frequency @xmath1 ( defined later ) with time .
( iii ) jets are especially important in grbs since they are relevant to almost all aspects of grb phenomenon , e.g. the total energy released in a grb , the burst rate , the ejection mechanism of the central engine and the dynamical evolution of the ejecta .
it is of particular interest that the light curves of afterglows might be changed when the jet enters a spreading phase , as pointed out by rhoads ( 1999 ) , sari et al .
( 1999 ) and dai , huang & lu ( 2001 ) .
the discovery of polarization in the afterglow of grb 990510 has shown evidence for a jet - like outflow ( covino et al .
1999 ; wijers et al .
1999 ) .
we consider grbs as jets .
the jet evolves as spherical - like expansion before entering the spreading phase ( rhoads 1999 ; sari et al .
1999 ) , and the heated electrons are always fast cooling at first .
this radiative and spherical - like phase ends once either ( 1 ) the electrons become slow cooling or ( 2 ) the relativistic jet transits to the spreading phase .
so a `` radiative break '' in the light curve may appear for the case where the jet transits to the spreading phase when the electrons are still fast cooling and the jet is in the radiative regime .
this break is different from the widely discussed `` adiabatic break '' for the case where the transition occurs in the adiabatic regime in which the electrons have been slow cooling . in this work
we derive the overall analytical evolution of synchrotron emission from a relativistic jet .
we first introduce the dynamical evolution of a relativistic jet in section 2
. then the temporal evolution of synchrotron emission from the jet is calculated in section 3 . in section 3.1
we analyze the radiative regime when the jet evolution is spherical - like .
section 3.2 discusses the adiabatic break in which the electrons become slow cooling before the jet evolves into the spreading phase while section 3.3 explores the radiative break in which the transition to the spreading phase occurs in the radiative regime .
we give some discussions and conclusions in section 4 .
let s consider a relativistic jet expanding into its surrounding medium with an initial half opening angle of @xmath6 , a laterally - spreading velocity of @xmath7 and a bulk lorentz factor of @xmath8 .
though the shock is beamed , the jet evolution is spherical - like as long as @xmath9 . because the electrons accelerated by a shock are always fast cooling at first
, the jet should be radiative if the energy density of the electrons is a significant fraction @xmath4 of the total energy density of the shocked medium .
the jet hydrodynamics in this radiative regime will be different from the well known self - similar solution for adiabatic blast waves derived by blandford & mckee ( 1976 ) , @xmath10 , where @xmath11 is the shock radius .
thus , the jet behaves as a radiative and spherical - like expansion at the initial stage .
the jet evolution later is divided into two cases .
( _ a _ ) in the case of an adiabatic break , the accelerated electrons become slow cooling before the jet evolves into a spreading phase , so the jet exhibits an adiabatic and spherical - like evolution as @xmath10 sequentially and finally a spreading phase of @xmath8 dropping exponentially with increasing @xmath11 .
( _ b _ ) in the case of a radiative break , the jet transition into the spreading phase occurs in the radiative regime where the shocked electrons are still fast cooling , and therefore the jet evolves from the initial radiative spherical - like phase to the spreading phase directly , without the second phase of adiabatic and spherical - like evolution of case ( _ a _ ) .
1 depicts these two cases of the dynamical evolutions of relativistic jets .
since the evolution of an adiabatic spherical blast wave is well known , we introduce here only the dynamics of so - called `` semiradiative '' blast waves for the initial stage of radiative and spherical - like evolution and of the radiative jet for the final stage in the case of a radiative break . recently , for semiradiative blast waves ,
in which the energy radiated is considerable to affect the evolution of the blast waves , cohen et al .
( 1998 ) have also derived a self - similar solution under the assumption that a fixed fraction of the energy generated by the shock is radiated away .
this assumption is applicable at early times , since the heated electrons are in the fast - cooling regime and a fixed fraction of the total energy that has been transferred to electrons is radiated away while the protons do not radiate their energy and remain hot .
following cohen et al .
( 1998 ) , the radius @xmath11 and the lorentz factor @xmath8 of a semiradiative blast wave are in relation of @xmath12 , and they evolve with the observer time as @xmath13 @xmath14 @xmath15 here @xmath16 is the initial coasting lorentz factor of the jet , @xmath17 is the deceleration radius where the energy in the shocked medium equals that in the original explosion , @xmath18 is the observer time at which the deceleration radius is reached , @xmath19 is a fraction of energy that is radiated away and should be understood as @xmath20 in the fast cooling regime , and @xmath21 parameterizes the proton number density @xmath22 of the surrounding medium with @xmath23 . for @xmath24 and @xmath19=(01 ) , @xmath25 varies from 3 to 12 and the lorentz factor of the blast wave evolves based on @xmath26 to @xmath27 .
it is the adiabatic solution case with @xmath28 . denoting the energy in the original explosion as @xmath29 , and the redshift of this explosion as @xmath30
, we have @xmath31 @xmath32 in the fully radiative ( @xmath33 ) case , the scaling @xmath34 differs from the widely used blandford - mckee solution , @xmath35 ( see , e.g. , sari , piran & narayan 1998 , bttcher & dermer 2000 ) , which assumes a thin shell with a cooled interior and the kinetic energy of the shell can be written as @xmath36 , where @xmath37 is the shell mass .
in fact , the quantities , such as velocity , density and energy density , should be a function of distance from the shock front .
thus , we prefer to use the self - similar solution of cohen et al .
( 1998 ) which assumes a shell with finite width within which the interior exerting a pressure onto the outer . here
after we use the equations above to describe the evolution of blast waves in the following calculations . for the dynamics of a relativistic jet , rhoads ( 1999 ) has derived an adiabatic solution , in which the jet evolution in the spreading phase is an exponential slowing down of @xmath8 with increasing @xmath11 .
we can expect the jet slows down even faster if it is in the radiative regime and loses energy more quickly than an adiabatic jet does .
thus , for a radiative jet , @xmath11 is practically a constant during the spreading phase .
therefore , from @xmath38 , we have the same scaling relation of @xmath39 as that of an adiabatic jet .
as shown later , the previous studies on jetted afterglows are only suitable for the adiabatic break case with weak compton - cooling effect . at first , we concentrate on deriving the spectra and light curves in the initial radiative and spherical - like stage , and then we discuss the light - curve change around the break time in both the adiabatic and radiative break cases with strong or weak compton - cooling effect .
we only consider the emission from the accelerated electrons , neglecting the emission from the hot protons .
the electron distribution is assumed to be a power - law of lorentz factor , @xmath40 , with the minimum random lorentz factor , @xmath41 where @xmath42 is the usual hydrogen mass fraction of the surrounding medium .
we assume that @xmath43 here after so that the total electron energy is dominated by electrons with @xmath44 [ recently dai & cheng ( 2001 ) have studied the case of @xmath45 .
the strength of the magnetic field in the co - moving frame is given by @xmath46 where @xmath47 is a fraction of the total internal energy densty that is carried by the magnetic field .
we concentrate here on the radiative regime when the cooling timescale for electron with @xmath44 is shorter than the dynamical timescale , @xmath48 for @xmath49 . in this case
, the distribution of electrons extends to a lower lorentz factor @xmath50 as a power - law @xmath51 for @xmath52 and @xmath53 for @xmath54 , where @xmath50 is the lorentz factor of electrons that cool on the dynamical timescale @xmath55 , @xmath56 here we have considered the inverse - compton cooling of electrons by the compton parameter ( sari & esin 2000 ) @xmath57 where @xmath58 is a fraction of the electron energy that is radiated away .
we should note that @xmath59 for fast cooling and @xmath60 for slow cooling .
the simple model of synchrotron spectra of afterglows is a broken power - law with three break frequencies .
one is the absorption frequency @xmath61 below which the synchrotron photons are self - absorbed by electrons .
the other two are the characteristic synchrotron frequencies according to electrons with lorentz factor @xmath44 and @xmath50 , denoted @xmath62 and @xmath1 , respectively . following the improved treatment of synchrotron emission by wijers & galama ( 1999 )
, we have @xmath63 @xmath64 where the dimensionless factor @xmath65 , for @xmath66 .
the absorption frequency is defined to be the point where @xmath67 .
note that in the co - moving frame of the shocked shell ( denoted with a prime ) , the absorption coefficient @xmath68 scales as @xmath69 for @xmath70 and @xmath71 for @xmath72 ( waxman 1997b ) , where @xmath73 is the index of the distribution of those electrons relevant to synchrotron self - absorption around @xmath74 .
for the absorption coefficient at co - moving cooling frequency @xmath75 we should take @xmath76 and then ( rybicki & lightman 1979 ) @xmath77 where @xmath78 ( coming from @xmath79 for @xmath80 ) , @xmath81 and @xmath82 is the gamma function .
the optical depth at @xmath83 is then @xmath84 , where @xmath85 is the co - moving width of the shocked shell , thus we can get the absorption frequency @xmath86 for @xmath87 in the radiative regime .
we have neglected the case of @xmath88 since @xmath89 lasts for only few seconds in general parameter space .
the peak flux in the synchrotron spectrum is @xmath90 where @xmath91 is a factor defined by wijers & galama ( 1999 ) , which is @xmath92 for @xmath66 here , @xmath93 is the luminosity distance , and @xmath94 is the number of electrons in the shocked shell and given by @xmath95 substituting @xmath8 and @xmath11 with aid of equations ( 1)-(3 ) , one obtains expressions of @xmath62 , @xmath1 , @xmath61 and @xmath96 ( shown in appendix ) , and their scaling relations with the observer s time are @xmath97 @xmath98 @xmath99 @xmath100 given these three break frequency and peak flux , we can calculate the synchrotron spectrum and light curve .
note that @xmath101 in the radiative regime . for
a homogeneous medium , @xmath24 ( note that hereafter in this paper the discussions are all concerned about this homogeneous medium case ) , the fluxes in the four frequency ranges divided by three break frequencies evolve as @xmath102 the light curve decay indexes are relevant to @xmath103 , and shown in fig .
2 . therefore , observations of early - time afterglows following a prompt burst can help to determine the parameter @xmath4 .
we have neglected the effect of inverse - compton process on the shape of synchrotron spectrum , because the thomson optical depth of synchrotron photons is generally very small and only a negligibly small fraction of synchrotron photons is scattered .
besides , the typical energy of inverse - compton emission , @xmath104 , usually far exceeds that of the synchrotron component which we are concerned about , since the accelerated electrons are always relativistic because @xmath105 for relativistic blast waves
. the end of this radiative and spherical - like stage is determined by the minimum of two time scales .
one is the time @xmath106 when @xmath107 . after this time ,
the accelerated electrons around @xmath108 become slow cooling so that the shock is practically not radiative any more and transits from the radiative to the adiabatic phase ( the transition is gradual due to the gradual change of @xmath19 from @xmath4 to 0 , since @xmath109 . thus , in this transition phase , the blast wave evolution will not be self - similar ) .
another is the break time @xmath110 when @xmath111 and the jet enters the spreading phase . having the expressions of @xmath62 and @xmath1 we can derive @xmath106 ( a detailed expression given in appendix ) . with the aid of the evolution of @xmath8
, we have the break time @xmath112 ( see a detail in appendix ) . here
we have taken @xmath113 .
3 shows two times , @xmath106 and @xmath110 , varying with @xmath4 .
it is the adiabatic break case in the small @xmath4 end since @xmath114 , while the radiative break case tends to appear in the right end where @xmath115 .
we first discuss the case of @xmath114 , which corresponds to fig .
1a . when the time @xmath106 is reached , the jet enters an adiabatic and spherical - like stage which have been studied by many authors .
the blast wave evolves as @xmath116 .
we have the well - known scaling relations @xmath117 in some @xmath118 cases , i.e. a magnetic field extremely far below equipartition like the cases of grb 971214 ( wijers & galama 1999 ) and 990123 ( galama et al .
1999 ) , the adiabatic blast wave will be strong compton cooling with @xmath119 .
the evolution of the cooling frequency @xmath1 becomes somewhat complicated if the @xmath0 case is considered .
we combine @xmath120 and @xmath121 for @xmath0 with @xmath122 to obtain the general scaling @xmath123 for an adiabatic and spherical - like jet , this becomes @xmath124 we can see that @xmath1 might increase or decrease with time , depending on whether @xmath125 or @xmath126 , respectively . note that @xmath127 is also a function of time due to decreasing @xmath58 .
so if @xmath128 , then @xmath0 in the beginning fast cooling phase , and @xmath129 generally in the slow cooling phase .
the slowing - down leads to @xmath130 for @xmath131 , or @xmath132 for @xmath133 until @xmath134 ( we assume @xmath135 here after , which is consistent with the observations which show that all afterglows appear to have the electron energy distribution of @xmath135 ) .
thus there might be a transition from the strong compton cooling ( @xmath0 ) to the weak compton cooling ( @xmath136 ) phase , as pointed out by sari et al .
( 2000 ) . however , if @xmath137 , then @xmath136 throughout .
it is now easy to derive the light curves in four frequency ranges @xmath138 note that the transition of a jet from the radiative phase to the adiabatic phase at @xmath106 results in a flattening of the light curve in the energy range of @xmath139 , and even the light curve is more flattening in the case of @xmath0 than in the case of @xmath136 .
a break time is reached when @xmath111 . in this adiabatic break case , eq .
( 21 ) is invalid to calculate this break time since an adiabatic and spherical - like phase has appeared before the final spreading phase .
we need to derive the break time again as follows . from the integral @xmath140 where @xmath141 , the break time is given by @xmath142 ( see a detail in appendix ) for @xmath143 or @xmath144 . this value is larger than that of eq .
( 21 ) .
having entered the spreading phase , the jet evolves as @xmath39 and the shock radius @xmath11 can be regarded as a constant .
we rewrite the scalings of the break frequencies and the peak flux as @xmath145 it is a rather remarkable result that @xmath1 is actually increasing with time in the @xmath146 case .
therefore the flux turns to evolve as @xmath147 note that if @xmath136 , the transition of a relativistic jet into the spreading phase results in a steepening of the light curve in @xmath148 , with decay index @xmath149 independent of whether @xmath139 or @xmath150 .
but in the case of strong compton cooling , @xmath0 , something different happens : the transition yields the same steepening of the light curve into @xmath151 in @xmath152 , but due to the cooling frequency @xmath1 increasing in time , a flattening from @xmath153 to @xmath154 appears in @xmath139 .
define the light curve index @xmath155 and the spectral index @xmath156 as @xmath157 .
table 1 summarizes the relations between @xmath155 and @xmath156 above @xmath62 for different cases .
4 presents the effect of strong compton cooling on the @xmath158 relation above @xmath1 . [ cols="^,^,^,^ " , ] [ t : afterglow ] finally , when @xmath107 , the electrons become slow cooling , the temporal scalings of the break frequencies and the fluxes are the same as those in the later spreading phase of an adiabatic jet in section 3.2.2 .
so eq . ( 28 ) are valid here .
but the @xmath106 calculated in section 3.2.2 is not valid here because the jet has been in the spreading phase rather than the spherical - like phase .
this time becomes @xmath159 ( see a detail in appendix ) , where the subscript jet " denotes the quantities in time @xmath110 .
we have analyzed the whole process of a relativistic jet expanding into a homogeneous environment , from the radiative to adiabatic regime and from the spherical - like to spreading phase .
then we have found two different kinds of light - curve break for the jet effect .
one is an adiabatic jet break which appears if the transition of the jet into the spreading phase happens in the adiabatic regime when the accelerated electrons are already slow cooling .
this case is widely discussed already .
another is a radiative jet break which corresponds to the jet spreading obviously in the radiative regime when the accelerated electrons are still fast cooling .
this case leads to a weaker break in the light curve if @xmath160 .
based on the dynamics , we have derived the light curves of synchrotron emission from jetted afterglows , and summarized our main results as follows .
first , in the earliest radiative and spherical - like phase , the light curve decay index is steep and relevant to @xmath4 shown in eq .
( 20 ) , e.g. , the flux at high energy @xmath148 decays as fast as @xmath161 for @xmath162 and @xmath163 .
thus , as pointed out by bttcher & dermer ( 2000 ) , the rapid detections of afterglows at early times will provide a way to determine the electron energy fraction @xmath4 of total energy density .
second , if the compton cooling of electrons dominates the synchrotron cooling , the afterglows will exhibit quite different behaviors from the previous predictions . in the case of an adiabatic jet break , the sharp steepening of light curve in both @xmath150 and @xmath139
is only expected for the case of @xmath164 , i.e. @xmath136 in the beginning . while for the case of @xmath165 , the jet begins with strong compton scattering ( @xmath0 ) , and it might dominate total cooling over the whole relativistic stage ( sari & esin 2000 ) . in this case , the light curve steepening in @xmath150 is accompanied with a flattening ( or even a climb ) in @xmath139 , due to the increasing cooling frequency @xmath1 for @xmath0 . finally , in the case of a radiative jet break , the late - time change of the dynamical evolution due to jet effect is weaker than in the case of an adiabatic jet break , which results in a weaker light curve break . if @xmath160 , the change is practically smooth , yielding a smooth `` break '' .
thus , this kind of jet will show a singly steep light curve without obvious steepening break in @xmath139 , even though the jet is highly collimated with a small @xmath6 .
a few grb afterglows are observed to have an achromatic light curve break , e.g. , grb 990123 ( kulkarni et al .
1999 ; castro - tirado et al .
1999 ; fruchter et al .
1999 ) , grb 990510 ( harrison et al . 1999 ; stanek et al .
1999 ) , grb 991216 ( halpern et al .
2000 ) , grb 000301c ( rhoads & fruchter 2001 ; masetti et al . 2000 ; jensen et al . 2001 ; berger et al .
2000 ; sagar et al .
2000 ) , grb 000418 ( berger et al .
2001 ) , and grb 000926 ( price et al . 2001
; harrison et al . 2001 ; sagar et al .
2001a ; piro et al .
2001 ) , grb 010222 ( masetti et al .
2001 ; stanek et al . 2001 ; sagar et al . 2001b ; cowsik et al . 2001 ; in t zand et al .
jets in grbs are usually proposed to account for these breaks .
however , in view of our work , two cases of relativistic jets are ruled out to explain these breaks : ( i ) an adiabatic jet ( @xmath114 ) with strong compton cooling ( @xmath0 ) and the cooling frequency @xmath1 locating in the observed energy range ; ( ii ) a radiative jet ( @xmath115 ) with a significant fraction of the total energy occupied by electrons ( @xmath5 ) .
it should be noted that all the discussions here are concerned about the relativistic stage .
if a relativistic jet expands into a medium as dense as @xmath166 , such as a circumstellar cloud ( galama & wijers 2001 ) , the blast wave must enter the non - relativistic stage within a few days , leading to a steepening of the afterglow light curve ( dai & lu 1999 , 2000 ) .
this transition from the relativistic to non - relativistic stage produces another promising explanation for the broken afterglow light curves .
z. li would like to thank d. m. wei , x. y. wang and y. f. huang for valuable discussions .
the authors thank the anonymous referee for valuable and detailed comments .
this work was supported by the national natural science foundation of china under grants 19973003 and 19825109 , and the national 973 project ( nkbrsf g19990754 ) .
the time - evolution of the three break frequencies and peak flux in the earliest radiative and spherical - like phase are given below : @xmath167 where @xmath168 @xmath169 where @xmath170 @xmath171 where @xmath172 @xmath173 where @xmath174^{\frac{6+k(m-2)-2m}{2(k-3)(1+m ) } } 10^{\frac{32.33 - 16.17k-10.78m}{1+m}}.\ ] ] the time when @xmath107 is given by @xmath175^{\frac{1+m}{k+m-1 } } ( 1+z)x_p^{\frac{1+m}{2(k+m-1)}}a^{\frac{k - m-3}{(k-3)(k+m-1 ) } } \nonumber \\ & \times & \ e_0^{\frac{(k-2)m}{(k-3)(k+m-1)}}\gamma_0^{-\frac{2(k-2)(k+m-3)}{(k-3)(k+m-1 ) } } ( \epsilon_b\epsilon_e)^{\frac{1+m}{k+m-1}},\end{aligned}\ ] ] where @xmath176^{\frac{(k-2)m}{(k-3)(k+m-1)}}.\ ] ] for the case of an adiabatic break @xmath177 , and by @xmath178 where @xmath179 for the case of a radiative break @xmath180 .
the break time when a relativistic jet begins to spread exponentially is given by @xmath181 where @xmath182 for the case of a radiative break , and by @xmath183^{-{m-3\over3(m-1 ) } } ( 1+z)^{3-m\over3(1+m)}x_p^{-\frac{m-3}{6(m-1 ) } } \nonumber \\ & \times & a^{-{5m-9\over 9(m-1)}}e_0^{2m\over 9(m-1 ) } \gamma_0^{-\frac { 4(m-3)}{9(m-1)}}(\epsilon_b\epsilon_e)^{-{m-3\over3(m-1 ) } } \theta_0^{8/3}{\rm s},\end{aligned}\ ] ] where @xmath184 for the case of an adiabatic break .
berger , e. et al .
2000 , apj , 545 , 56 berger , e. et al .
2001 , apj , 556 , 556 blandford , r. d. & mckee , c. f. 1976 , phys .
fluids , 19 , 1130 bttcher , m. & dermer , c. d. 2000 , apj , 532 , 281 castro - tirado , a. j. et al .
1999 , science , 283 , 2069 chiang , j. & dermer , c. d. 1999 , apj , 512 , 699 .
cohen , e. , piran , t. & sari , r. 1998 , apj , 509 , 717 covino , s. et al . 1999 ,
a&a , 348 , l1 cowsik , r. et al .
2001 , basi , in press ( astro - ph/0104363 ) dai , z. g. & cheng , k. s. 2001 , apj , 558 , l109 dai , z. g. , huang , y. f. & lu , t. 2001 , mnras , 324 , l11 dai , z. g. & lu , t. 1999 , apj , 519 , l155 dai , z. g. & lu , t. 2000 , apj , 537 , 803 dermer , c. d. , bttcher , m. & chiang , j. 2000 , apj , 537 , 255 fruchter , a. s. et al . 1999 ,
apj , 519 , l13 galama , t. j. et al .
1999 , nature , 398 , 394 galama , t. j. & wijers , r. a. m. j. 2001 , apj , 549 , l209 gou , l. j. , dai , z. g. , huang , y. f. & lu , t. 2001 , a&a , 368 , 464 granot , j. , pran , t. & sari , r. 1999 , apj , 527 , 236 halpern , j. p. et al .
2000 , apj , 543 , 697 harrison , f. a. et al .
1999 , apj , 523 , l121 harrison , f. a. et al .
2001 , apj , 559 , 123 huang , y. f. , dai , z. g. & lu , t. 2000b , mnras , 316 , 943 huang , y. f. , gou , l. j. , dai , z. g. & lu , t. 2000a , apj , 543 , 90 in t zand , j. j. m. et al .
2001 , apj , 559 , 710 jensen , b. l. et al .
2001 , a&a , 370 , 909 kulkarni , s. r. et al .
1999 , nature , 398 , 389 masetti , n. et al .
2000 , a&a , 359 , l23 masetti , n. et al .
2001 , a&a , 374 , 382 mszros , p. & rees , m. j. 1997 , apj , 476 , 232 mszros , p. , rees , m. j. & wijers , r. a. m. j. 1997 , apj , 499 , 301 moderski , r. , sikora , m. & bulik t. 2000 , apj , 529 , 151 paczyski , b. & rhoads , j. 1993 , apj , 418 , l5 paczyski , b. & xu , g. , 1994 , apj , 424 , 708 panaitescu , a. & mszros , p. 1998 , apj , 501 , 772 panaitescu , a. & kumar , p. 2000 , apj , 543 , 66 panaitescu , a. 2001 , apj , 556 , 1002 piran , t. 1999 , physics reports , 314 , 575 piro , l. et al .
2001 , apj , 558 , 442 price , p. a. et al .
2001 , apj , 549 , l7 rees , m. j. & mszros , p. 1994
, apj , 430 , l93 rhoads , j. 1999 , apj , 525 , 737 rhoads , j. & fruchter , a. s. 2001 , apj , 546 , 117 rybicki , g. b. & lightman , a. p. 1979
, radiative processes in astrophysics ( new york : wiley interscience ) , p. 189 sagar , r. et al . 2000 , basi , 28 , 499 sagar , r. et al .
2001a , basi , 29 , 1 sagar , r. et al . 2001b , basi , 29 , 91 sari , r. & esin , a. a. 2001 , apj , 548 , 787 sari , r. , piran , t. & halpern , j. 1999 , apj , 524 , l43 sari , r. , piran , t. & narayan , r. 1998 , apj , 497 , l17 stanek , k. z. et al .
1999 , apj , 522 , l39 stanek , k. z. et al .
2001 , apj , in press ( astro - ph/0104329 ) totani , t. 1998 , apj , 502 , l13 vietri , m. 1997 , apj , 478 , l9 van paradijs , j. , kouveliotou , c. , & wijers , r. a. m. j. 2000 , ara&a , 38 , 379 waxman , e. 1997a , apj , 485 , l9 waxman , e. 1997b , apj , 489 , l33 wei , d. m. & lu , t. 1998 , apj , 505 , 252 wei , d. m. & lu , t. 2000 , apj , 541 , 203 wijers , r. a. m. j. et al .
1999 , apj , 523 , l33 wijers , r. a. m. j. & galama , t. j. 1999 , apj , 523 , 177 wijers , r. a. m. j. , rees , m. j. , & mszros , p. 1997 ,
mnras , 288 , l51 zhang , b. & mszros , p. 2001
, apj , 559 , 110
|
we discuss the afterglow emission from a relativistic jet that is initially in the radiative regime in which the accelerated electrons are fast cooling .
we note that such a `` semiradiative '' jet decelerates faster than an adiabatic jet does .
we also take into account the effect of strong inverse - compton scattering on the cooling frequency in the synchrotron component and therefore on the light curve decay index .
we find that there are two kinds of light - curve break for the jet effect .
the first is an `` adiabatic break '' if the electrons become slow cooling before the jet enters a spreading phase , and the second is a `` radiative break '' which appears on the contrary case .
we then show how a relativistic jet evolves dynamically and derive the overall temporal synchrotron emission in both cases , focusing on the change in light curve decay index around the break time .
finally , in view of our results , we rule out two cases for relativistic jets to account for the observed light curve breaks in a few afterglows : ( i ) an adiabatic jet with strong compton cooling ( @xmath0 ) and with the cooling frequency @xmath1 locating in the observed energy range ; ( ii ) a radiative jet with a significant fraction of total energy occupied by electrons ( @xmath2 ) .
-0.5 cm gamma rays : bursts radiation mechanisms : non - thermal
|
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * figure 1 .
* ground state energy for @xmath53 .
dotted curve : homogeneous phase .
dotted - dashed curve : ab phase .
solid curve : incommensurate phase ( @xmath8 optimised ) .
divergence of @xmath72 at @xmath55 .
divergence of @xmath14 for optimised @xmath8 at @xmath58 .
maxwell constructions between @xmath73 and @xmath74 or @xmath56 .
incommensurate phase between @xmath56 and @xmath58 . to facilitate visibility a linear function of @xmath75 has been added : @xmath76 , @xmath77
. * figure 2 .
* phase diagram .
hom : homogeous phase , ab : ab phase , ip : incommensurate phase , ps - ab : phase separated region between homogeneous and ab phase , ps - ip : phase separated region between incommensurate and ab phase .
@xmath78 , @xmath79 : lower and upper interaction bound of the ip .
* figure 3 . * the value of @xmath8 along the phase borders of the incommensurate phase .
solid curve : right border ; dashed curve : left border .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
|
we focus on the ground state phase diagram of a system of spinless fermions with repulsion on a hypercubic lattice in the limit of infinite dimensions .
it occurs spontaneous symmetry breaking into a charge density wave ( cdw ) . using an ansatz for the order parameter which includes the homogeneous , the ab- and a large class of incommensurate phases we are able to calculate the phase diagram .
+ july 20 , 1994 univ .
zu kln + uva / itf 16/94 * existence of an incommensurate ground state of interacting spinless fermions in infinite dimensions * + g.s .
uhrig@xmath0 and r. vlaming@xmath1 + @xmath0 _ inst .
f. theor .
physik , univ .
zu kln , 50937 kln , germany _ + @xmath1 _ inst .
v. theor .
fysica , uva , 1018 xe , amsterdam , the netherlands _
+ it is an important object to disclose the structure of the phase diagram of models with strong electron correlation .
the problem is twofold .
first , one has to have a clear picture which phases are likely to be realised and second , for these phases it must be possible to calculate the free energy . in case of the hubbard model for example
both problems can hardly be surmounted .
even if one restricts oneself to a small number of phases and to the dynamic mean field theory resulting from the limit of high dimensions @xcite , the problem still requires massive computational effort ( e.g. @xcite ) .
the situation is different for the model of interacting spinless fermions @xcite on an infinite dimensional hypercubic lattice .
although the diagrammatic theory is simple , the physics is far from trivial .
the hamiltonian of the spinless fermion model in second quantisation is @xmath2 scaling with the inverse dimension @xmath3 is performed to ensure the continuity of the limit of infinite dimensions @xcite . in a previous work
the stability of the homogeneous phase was investigated in the limit @xmath4 @xcite . in this
limit hartree- and random phase approximation become exact @xcite .
besides the occurrence of an ab - charge density wave ( cdw ) characterised by the wave vector @xmath5 it was found that the density - density susceptibility @xmath6 diverges at incommensurate values of the wave vector @xmath7 .
the parameter @xmath8 is defined as @xmath9 and @xmath10 is the inverse bare susceptibility .
commensurate values of @xmath7 which belong to a value of @xmath11 are of measure zero .
the complete stability analysis is possible in @xmath12 because @xmath13 depends on @xmath7 only via @xmath8 @xcite .
thus it is sufficient to examine @xmath14 @xcite .
the divergence of @xmath14 indicates a second order transition from the homogeneous phase at high doping and/or low interaction into an incommensurate cdw . the structure of the new incommensurate phase has not yet been investigated . in finite dimensions and for related models such as
the hubbard model and the @xmath15-@xmath16 model there exist many works which treat different variational or approximative approaches to incommensurate phases @xcite .
since the perturbation theory of the model ( 1 ) becomes exactly tractable in @xmath4 it is the natural candidate to examine the energetic effects of different spatial structures of the order parameter .
the resulting equations are self - consistent and include infinitely many parameters in the thermodynamic limit so that no systematic approach to their solution exists @xcite . here
we give results for the following ansatz for the spatial structure of the order parameter @xmath17 where @xmath18 is the average particle density : @xmath19 .
the index @xmath20 counts the directions ; @xmath21 is the @xmath20th component of @xmath22 , @xmath23 is the overall amplitude and the new defined functions @xmath24 take the values @xmath25
. the product form of this ansatz allows to profit best from the simplifications of @xmath4 .
yet it is still very general since it allows any sequence of @xmath26 and @xmath27 independently for the @xmath28 directions without assuming translational invariance . a close inspection of the free energy shows that it depends on @xmath29 only via the amplitude @xmath23 and the relative frequency @xmath30 that @xmath29 has the same sign on adjacent sites @xcite .
these are the only parameters which are relevant for the influence of the spatial structure on the energy . in order to find an explicit functional for the energy we exploit the freedom in choosing the @xmath24 .
for @xmath31 we set @xmath32 and for @xmath33 we set @xmath34 .
this choice is special since the order parameter can be characterised by the wave vector @xmath35 but does not restrict the freedom contained in the ansatz for @xmath29 .
this gives a general relationship @xmath36 , although only for special realisations @xmath8 can be defined .
the easy structure of the cdw with wave vector @xmath37 helps to resum the perturbation series for the grand canonical potential since the green function is a @xmath38 matrix in @xmath39-space . in infinite dimensions
the diagram for the free energy is simple .
compared to the free fermion system the grand potential gains a term @xmath40 , where @xmath41 is the green function and @xmath42 the self - energy . at @xmath43
this yields the ground state energy @xmath44 where @xmath45 and @xmath46 .
the doping @xmath47 is given by @xmath48 and the equations which define the energetic minimum are @xmath49 and @xmath50 .
for a particular value of the interaction @xmath51 it is possible to determine for each doping the parameters @xmath52 and @xmath8 which describe the physical state .
additionally , one has to be on the watch for first order phase transitions . as example we look at the case @xmath53 .
we plotted ground state energy as function of doping in fig . 1 .
for large doping and at half - filling the situation is clear . for the former
the system is in the homogeneous phase whereas for the latter the system is in the commensurate cdw characterised by @xmath54 ( ab - phase ) .
inbetween four special points at @xmath55 , @xmath56 , @xmath57 , and @xmath58 can be identified .
let us focus on the situation with fixed @xmath54 .
the system is forced to choose between the homogeneous phase and the ab - phase .
the response of the system towards fluctuations with wave vector @xmath59 becomes infinite at @xmath55 .
the ground state energy in indicated by a dotted and a dotted - dashed curve , respectively in fig . 1 .
for smaller doping the ground state energy is concave so that a first order phase transition will occur .
the maxwell construction gives rise to the value @xmath57 .
thus , there is phase separation between the homogeneous phase and the ab phase for @xmath60 .
this situation turns out to be generic for @xmath61 and @xmath62 .
inbetween and thus for the case @xmath53 a different scenario is realised .
the divergence of @xmath14 is at @xmath58 for general values of @xmath8 .
for @xmath63 holds @xmath64 . by minimizing @xmath65
we obtain the solid curve in fig . 1 .
again , the curve is concave for low dopings so that a maxwell construction must be made .
the doping where the phase separation sets in is @xmath56 .
it is important to realise that @xmath66 . for dopings @xmath67
the pure incommensurate phase is the absolute energy minimum .
the overall result is depicted in fig . 2 .
at half - filling the system is in the ab - phase for all @xmath68 . for large doping
the system is in the homogeneous phase . for @xmath69 and @xmath70
there is phase separation between the ab - phase and a hole - enriched homogeneous phase .
for @xmath63 there exists a region where an incommensurate phase is present , and a region of phase separation between the ab - phase and the hole - enriched incommensurate phase .
the parameter @xmath8 is close to @xmath27 in the incommensurate phase . to give an impression we plotted those values of @xmath8 in fig . 3 which arise at the borders of the incommensurate phase in fig . 2 .
the function @xmath71 at constant @xmath51 is monotonic so that the figure gives a good impression of the deviation of @xmath8 from @xmath27 .
note that the values for @xmath8 belonging to the left border ( fig .
2 ) , given by the solid curve in fig . 3 , are important for the whole region of phase separation . in summary , we demonstrated the complexity of the ground state phase diagram of the model of interacting spinless fermions in infinite dimensions .
most important is the existence of an incommensurate phase and of the phase separation between the ab - phase and the incommensurate or homogeneous phase .
this work was supported by the deutsche forschungsgemeinschaft ( sfb 341 ) and the stichting voor fundamenteel onderzoek der materie ( fom ) . 10 d. vollhardt . in _ correlated electron systems _
, ed .
v. j. emery , p. 57
, world scientific , singapore , 1993 .
m. jarrell .
phys .
rev .
lett .
* 69 * , 168 ( 1992 )
. for exact results for this model in finite dimensions see c. n. yang and c. p. yang .
phys . rev . *
150 * , 321 and 327 ( 1966 ) ; j. des cloizeaux and m. gaudin . j. math
. phys .
* 7 * , 1384 ( 1966 ) ; j. des cloizeaux . j. math
. phys .
* 7 * , 2136 ( 1966 ) .
p. lemberger and n. macris .
lett .
math .
phys .
* 28 * , 295 ( 1993 ) ; r. shankar .
rev .
mod .
phys . * 66 * , 129 ( 1994 ) .
w. metzner and d. vollhardt .
phys .
rev .
lett . * 62 * , 324 ( 1989 ) .
e. mller - hartmann .
z. phys .
b - condens .
matter * 74 * , 507 ( 1989 ) .
g. s. uhrig and r. vlaming .
phys .
rev .
lett . * 71 * , 271 ( 1993 ) ; g. s. uhrig , r. strack and d. vollhardt .
phys .
rev .
b * 46 * , 9940 ( 1992 ) ; such an examination was also performed for the falicov - kimball model by j. k. freericks .
phys .
rev .
b * 47 * , 9263 ( 1993 ) .
r. eder .
phys .
rev .
b * 43 * , 10706 ( 1991 ) ; b. i. shraiman and e. d. siggia .
phys .
rev .
lett . * 62 * , 1564 ( 1989 ) ; c. l. kane , p. a. lee , t. k. ng , b. chakraborty and n. read .
phys .
rev .
b * 41 * , 2653 ( 1990 ) ; c. jayaprakash , h. r. krishnamurthy and s. sarker .
phys .
rev .
b * 40 * , 2610 ( 1989 ) ; m. dzierzawa .
z. phys .
b - condens .
matter * 86 * , 49 ( 1992 ) ; m. dzierzawa and r. frsard .
z. phys .
b - condens .
matter * 91 * , 245 ( 1993 ) .
g. s. uhrig and r. vlaming . to be published , ( 1994 ) .
|
the goal of visual texture synthesis is to infer a generating process from an example texture , which then allows to produce arbitrarily many new samples of that texture .
the evaluation criterion for the quality of the synthesised texture is usually human inspection and textures are successfully synthesised if a human observer can not tell the original texture from a synthesised one . in general , there are two main approaches to find a texture generating process .
the first approach is to generate a new texture by resampling either pixels @xcite or whole patches @xcite of the original texture .
these non - parametric resampling techniques and their numerous extensions and improvements ( see @xcite for review ) are capable of producing high quality natural textures very efficiently .
however , they do not define an actual model for natural textures but rather give a mechanistic procedure for how one can randomise a source texture without changing its perceptual properties .
in contrast , the second approach to texture synthesis is to explicitly define a parametric texture model .
the model usually consists of a set of statistical measurements that are taken over the spatial extent of the image . in the model a texture is uniquely defined by the outcome of those measurements and every image that produces the same outcome should be perceived as the same texture
. therefore new samples of a texture can be generated by finding an image that produces the same measurement outcomes as the original texture .
conceptually this idea was first proposed by julesz @xcite who conjectured that a visual texture can be uniquely described by the nth - order joint histograms of its pixels .
later on , texture models were inspired by the linear response properties of the mammalian early visual system , which resemble those of oriented band - pass ( gabor ) filters @xcite .
these texture models are based on statistical measurements taken on the filter responses rather than directly on the image pixels .
so far the best parametric model for texture synthesis is probably that proposed by portilla and simoncelli @xcite , which is based on a set of carefully handcrafted summary statistics computed on the responses of a linear filter bank called _ steerable pyramid _
however , although their model shows very good performance in synthesising a wide range of textures , it still fails to capture the full scope of natural textures .
synthesis method .
texture analysis ( left ) .
the original texture is passed through the cnn and the gram matrices @xmath0 on the feature responses of a number of layers are computed .
texture synthesis ( right ) .
a white noise image @xmath1 is passed through the cnn and a loss function @xmath2 is computed on every layer included in the texture model .
the total loss function @xmath3 is a weighted sum of the contributions @xmath2 from each layer .
using gradient descent on the total loss with respect to the pixel values , a new image is found that produces the same gram matrices @xmath4 as the original texture . ] in this work , we propose a new parametric texture model to tackle this problem ( fig .
[ method ] ) . instead of describing textures on the basis of a model for the early visual system @xcite
, we use a convolutional neural network a functional model for the entire ventral stream as the foundation for our texture model .
we combine the conceptual framework of spatial summary statistics on feature responses with the powerful feature space of a convolutional neural network that has been trained on object recognition . in that way we obtain a texture model that is parameterised by spatially invariant representations built on the hierarchical processing architecture of the convolutional neural network .
we use the vgg-19 network , a convolutional neural network trained on object recognition that was introduced and extensively described previously @xcite . here
we give only a brief summary of its architecture .
we used the feature space provided by the 16 convolutional and 5 pooling layers of the vgg-19 network .
we did not use any of the fully connected layers .
the network s architecture is based on two fundamental computations : 1 .
linearly rectified convolution with filters of size @xmath5 where @xmath6 is the number of input feature maps . stride and padding of
the convolution is equal to one such that the output feature map has the same spatial dimensions as the input feature maps . 2 .
maximum pooling in non - overlapping @xmath7 regions , which down - samples the feature maps by a factor of two .
these two computations are applied in an alternating manner ( see fig . [ method ] ) .
a number of convolutional layers is followed by a max - pooling layer .
after each of the first three pooling layers the number of feature maps is doubled . together with the spatial down - sampling , this transformation results in a reduction of the total number of feature responses by a factor of two .
[ method ] provides a schematic overview over the network architecture and the number of feature maps in each layer .
since we use only the convolutional layers , the input images can be arbitrarily large .
the first convolutional layer has the same size as the image and for the following layers the ratio between the feature map sizes remains fixed .
generally each layer in the network defines a non - linear filter bank , whose complexity increases with the position of the layer in the network .
the trained convolutional network is publicly available and its usability for new applications is supported by the caffe - framework @xcite . for texture generation
we found that replacing the max - pooling operation by average pooling improved the gradient flow and one obtains slightly cleaner results , which is why the images shown below were generated with average pooling .
finally , for practical reasons , we rescaled the weights in the network such that the mean activation of each filter over images and positions is equal to one .
such re - scaling can always be done without changing the output of a neural network if the non - linearities in the network are rectifying linear .
the texture model we describe in the following is much in the spirit of that proposed by portilla and simoncelli @xcite . to generate a texture from a given source image
, we first extract features of different sizes homogeneously from this image .
next we compute a spatial summary statistic on the feature responses to obtain a stationary description of the source image ( fig .
[ method]a ) .
finally we find a new image with the same stationary description by performing gradient descent on a random image that has been initialised with white noise ( fig .
[ method]b ) .
the main difference to portilla and simoncelli s work is that instead of using a linear filter bank and a set of carefully chosen summary statistics , we use the feature space provided by a high - performing deep neural network and only one spatial summary statistic : the correlations between feature responses in each layer of the network . to characterise a given vectorised texture @xmath8 in our model
, we first pass @xmath8 through the convolutional neural network and compute the activations for each layer @xmath9 in the network . since each layer in the network can be understood as a non - linear filter bank , its activations in response to an image form a set of filtered images ( so - called _ feature maps _ ) .
a layer with @xmath10 distinct filters has @xmath10 feature maps each of size @xmath11 when vectorised .
these feature maps can be stored in a matrix @xmath12 , where @xmath13 is the activation of the @xmath14 filter at position @xmath6 in layer @xmath9 .
textures are per definition stationary , so a texture model needs to be agnostic to spatial information . a summary statistic that discards the spatial information in the feature maps
is given by the correlations between the responses of different features .
these feature correlations are , up to a constant of proportionality , given by the gram matrix @xmath15 , where @xmath16 is the inner product between feature map @xmath17 and @xmath18 in layer @xmath9 : @xmath19 a set of gram matrices @xmath20 from some layers @xmath21 in the network in response to a given texture provides a stationary description of the texture , which fully specifies a texture in our model ( fig .
[ method]a ) .
to generate a new texture on the basis of a given image , we use gradient descent from a white noise image to find another image that matches the gram - matrix representation of the original image .
this optimisation is done by minimising the mean - squared distance between the entries of the gram matrix of the original image and the gram matrix of the image being generated ( fig .
[ method]b ) .
let @xmath8 and @xmath1 be the original image and the image that is generated , and @xmath22 and @xmath23 their respective gram - matrix representations in layer @xmath9 ( eq . [ gram ] ) .
the contribution of layer @xmath9 to the total loss is then @xmath24 and the total loss is @xmath25 where @xmath26 are weighting factors of the contribution of each layer to the total loss .
the derivative of @xmath2 with respect to the activations in layer @xmath9 can be computed analytically : @xmath27 the gradients of @xmath2 , and thus the gradient of @xmath28 , with respect to the pixels @xmath1 can be readily computed using standard error back - propagation @xcite .
the gradient @xmath29 can be used as input for some numerical optimisation strategy . in our work
we use l - bfgs @xcite , which seemed a reasonable choice for the high - dimensional optimisation problem at hand .
the entire procedure relies mainly on the standard forward - backward pass that is used to train the convolutional network .
therefore , in spite of the large complexity of the model , texture generation can be done in reasonable time using gpus and performance - optimised toolboxes for training deep neural networks @xcite .
generated stimuli .
each row corresponds to a different processing stage in the network . when only constraining the texture representation on the lowest layer , the synthesised textures have little structure , similarly to spectrally matched noise ( first row ) . with increasing number of layers on which we match the texture representation we find that we generate images with increasing degree of naturalness ( rows 25 ; labels on the left indicate the top - most layer included ) .
the source textures in the first three columns were previously used by portilla and simoncelli @xcite . for better comparison
we also show their results ( last row ) .
the last column shows textures generated from a non - texture image to give a better intuition about how the texture model represents image information.,width=434 ]
we show textures generated by our model from four different source images ( fig . [
generated ] ) .
each row of images was generated using an increasing number of layers in the texture model to constrain the gradient descent ( the labels in the figure indicate the top - most layer included ) .
in other words , for the loss terms above a certain layer we set the weights @xmath30 , while for the loss terms below and including that layer , we set @xmath31 . for example the images in the first row ( ` conv1_1 ' ) were generated only from the texture representation of the first layer ( ` conv1_1 ' ) of the vgg network .
the images in the second row ( ` pool1 ' ) where generated by jointly matching the texture representations on top of layer ` conv1_1 ' , ` conv1_2 ' and ` pool1 ' . in this way we obtain textures that show what structure of natural textures
are captured by certain computational processing stages of the texture model .
the first three columns show images generated from natural textures .
we find that constraining all layers up to layer ` pool4 ' generates complex natural textures that are almost indistinguishable from the original texture ( fig .
[ generated ] , fifth row ) .
in contrast , when constraining only the feature correlations on the lowest layer , the textures contain little structure and are not far from spectrally matched noise ( fig . [
generated ] , first row ) .
we can interpolate between these two extremes by using only the constraints from all layers up to some intermediate layer .
we find that the statistical structure of natural images is matched on an increasing scale as the number of layers we use for texture generation increases .
we did not include any layers above layer ` pool4 ' since this did not improve the quality of the synthesised textures . for comparability we used source textures that were previously used by portilla and simoncelli @xcite and
also show the results of their texture model ( fig .
[ generated ] , last row ) . to give a better intuition for how the texture synthesis works ,
we also show textures generated from a non - texture image taken from the imagenet validation set @xcite ( fig . [ generated ] , last column ) .
our algorithm produces a texturised version of the image that preserves local spatial information but discards the global spatial arrangement of the image .
the size of the regions in which spatial information is preserved increases with the number of layers used for texture generation .
this property can be explained by the increasing receptive field sizes of the units over the layers of the deep convolutional neural network . * a * , number of parameters in the texture model .
we explore several ways to reduce the number of parameters in the texture model ( see main text ) and compare the results . *
b * , textures generated from the different layers of the caffe reference network @xcite .
the textures are of lesser quality than those generated with the vgg network .
* c * , textures generated with the vgg architecture but random weights .
texture synthesis fails in this case , indicating that learned filters are crucial for texture generation . ]
when using summary statistics from all layers of the convolutional neural network , the number of parameters of the model is very large . for each layer with @xmath10 feature maps ,
we match @xmath32 parameters , so if we use all layers up to and including ` pool4 ' , our model has @xmath33k parameters ( fig .
[ alexnet_and_random]a , fourth column ) . however , we find that this texture model is heavily over - parameterised .
in fact , when using only one layer on each scale in the network ( i.e. ` conv1_1 ' , and ` pool1 - 4 ' ) , the model contains @xmath34k parameters while hardly loosing any quality ( fig .
[ alexnet_and_random]a , third column ) .
we can further reduce the number of parameters by doing pca of the feature vector in the different layers of the network and then constructing the gram matrix only for the first @xmath6 principal components . by using the first 64 principal components for layers ` conv1_1 ' , and ` pool1 - 4 ' we can further reduce the model to @xmath35k parameters ( fig .
[ alexnet_and_random]a , second column ) .
interestingly , constraining only the feature map averages in layers ` conv1_1 ' , and ` pool1 - 4 ' , ( @xmath36 parameters ) , already produces interesting textures ( fig .
[ alexnet_and_random]a , first column ) .
these _ ad hoc _ methods for parameter reduction show that the texture representation can be compressed greatly with little effect on the perceptual quality of the synthesised textures .
finding minimal set of parameters that reproduces the quality of the full model is an interesting topic of ongoing research and beyond the scope of the present paper
. a larger number of natural textures synthesised with the @xmath37k parameter model can be found in the supplementary material as well as on our website .
there one can also observe some failures of the model in case of very regular , man - made structures ( e.g. brick walls ) . in general , we find that the very deep architecture of the vgg network with small convolutional filters seems to be particularly well suited for texture generation purposes . when performing the same experiment with the caffe reference network @xcite , which is very similar to the alexnet @xcite , the quality of the generated textures decreases in two ways .
first , the statistical structure of the source texture is not fully matched even when using all constraints ( fig [ alexnet_and_random]b , ` conv5 ' ) .
second , we observe an artifactual grid that overlays the generated textures ( fig [ alexnet_and_random]b ) .
we believe that the artifactual grid originates from the larger receptive field sizes and strides in the caffe reference network . while the results from the caffe reference network show that the architecture of the network is important , the learned feature spaces are equally crucial for texture generation . when synthesising a texture with a network with the vgg architecture but random weights , texture generation fails ( fig .
[ alexnet_and_random]c ) , underscoring the importance of using a trained network .
performance of a linear classifier on top of the texture representations in different layers in classifying objects from the imagenet dataset .
high - level information is made increasingly explicit along the hierarchy of our texture model . ] to understand our texture features better in the context of the original object recognition task of the network , we evaluated how well object identity can be linearly decoded from the texture features in different layers of the network .
for each layer we computed the gram - matrix representation of each image in the imagenet training set @xcite and trained a linear soft - max classifier to predict object identity . as we were not interested in optimising prediction performance
, we did not use any data augmentation and trained and tested only on the @xmath38 centre crop of the images .
we computed the accuracy of these linear classifiers on the imagenet validation set and compared them to the performance of the original vgg-19 network also evaluated on the @xmath38 centre crops of the validation images .
the analysis suggests that our texture representation continuously disentangles object identity information ( fig .
[ gram_results ] ) .
object identity can be decoded increasingly well over the layers .
in fact , linear decoding from the final pooling layer performs almost as well as the original network , suggesting that our texture representation preserves almost all high - level information . at first sight
this might appear surprising since the texture representation does not necessarily preserve the global structure of objects in non - texture images ( fig .
[ generated ] , last column ) .
however , we believe that this inconsistency " is in fact to be expected and might provide an insight into how cnns encode object identity .
the convolutional representations in the network are shift - equivariant and the network s task ( object recognition ) is agnostic to spatial information , thus we expect that object information can be read out independently from the spatial information in the feature maps .
we show that this is indeed the case : a linear classifier on the gram matrix of layer ` pool5 ' comes close to the performance of the full network ( 87.7% vs. 88.6% top 5 accuracy , fig .
[ gram_results ] ) .
we introduced a new parametric texture model based on a high - performing convolutional neural network .
our texture model exceeds previous work as the quality of the textures synthesised using our model shows a substantial improvement compared to the current state of the art in parametric texture synthesis ( fig . [
generated ] , fourth row compared to last row ) .
while our model is capable of producing natural textures of comparable quality to non - parametric texture synthesis methods , our synthesis procedure is computationally more expensive . nevertheless , both in industry and academia , there is currently much effort taken in order to make the evaluation of deep neural networks more efficient @xcite .
since our texture synthesis procedure builds exactly on the same operations , any progress made in the general field of deep convolutional networks is likely to be transferable to our texture synthesis method .
thus we expect considerable improvements in the practical applicability of our texture model in the near future . by computing the gram matrices on feature maps ,
our texture model transforms the representations from the convolutional neural network into a stationary feature space .
this general strategy has recently been employed to improve performance in object recognition and detection @xcite or texture recognition and segmentation @xcite . in particular
cimpoi et al .
report impressive performance in material recognition and scene segmentation by using a stationary fisher - vector representation built on the highest convolutional layer of readily trained neural networks @xcite . in agreement with our results , they show that performance in natural texture recognition continuously improves when using higher convolutional layers as the input to their fisher - vector representation . as
our main aim is to synthesise textures , we have not evaluated the gram matrix representation on texture recognition benchmarks , but would expect that it also provides a good feature space for those tasks . in recent years , texture models inspired by biological vision have provided a fruitful new analysis tool for studying visual perception .
in particular the parametric texture model proposed by portilla and simoncelli @xcite has sparked a great number of studies in neuroscience and psychophysics @xcite .
our texture model is based on deep convolutional neural networks that are the first artificial systems that rival biology in terms of difficult perceptual inference tasks such as object recognition @xcite . at the same time , their hierarchical architecture and basic computational properties admit a fundamental similarity to real neural systems . together with the increasing amount of evidence for the similarity of the representations in convolutional networks and those in the ventral visual pathway @xcite , these properties make them compelling candidate models for studying visual information processing in the brain .
in fact , it was recently suggested that textures generated from the representations of performance - optimised convolutional networks may therefore prove useful as stimuli in perceptual or physiological investigations " @xcite .
we feel that our texture model is the first step in that direction and envision it to provide an exciting new tool in the study of visual information processing in biological systems .
this work was funded by the german national academic foundation ( l.a.g . ) , the bernstein center for computational neuroscience ( fkz 01gq1002 ) and the german excellency initiative through the centre for integrative neuroscience tbingen ( exc307)(m.b . ,
a.s.e , l.a.g . )
|
here we introduce a new model of natural textures based on the feature spaces of convolutional neural networks optimised for object recognition .
samples from the model are of high perceptual quality demonstrating the generative power of neural networks trained in a purely discriminative fashion . within the model , textures
are represented by the correlations between feature maps in several layers of the network .
we show that across layers the texture representations increasingly capture the statistical properties of natural images while making object information more and more explicit .
the model provides a new tool to generate stimuli for neuroscience and might offer insights into the deep representations learned by convolutional neural networks .
|
in the analysis of lifetime data we can use the gompertz , exponential and generalized exponential distributions .
it is known that the exponential distribution have only constant hazard rate function where as gompertz , and generalized exponential distributions can have only monotone ( increasing in case of gompertz and increasing or decreasing in case of generalized exponential distribution ) hazard rate .
these distributions are used for modelling the lifetimes of components of physical systems and the organisms of biological populations .
the gompertz distribution has received considerable attention from demographers and actuaries .
pollard and valkovics @xcite were the first to study the gompertz distribution , they both defined the moment generating function of the gompertz distribution in terms of the incomplete or complete gamma function and their results are either approximate or left in an integral form . later , marshall and olkin @xcite described the negative gompertz distribution ; a gompertz distribution with a negative rate of aging parameter .
recently , a generalization of the gompertz distribution based on the idea given in @xcite was proposed by @xcite this new distribution is known as generalized gompertz ( gg ) distribution which includes the exponential ( e ) , generalized exponential ( ge ) , and gompertz ( g ) distributions . a new generalization of th gompertz ( g ) distribution which results of the application of the gompertz distribution to the beta generator proposed by @xcite , called the beta - gompertz ( bg ) distribution which introduced by @xcite .
on the other hand the two - parameter exponentiated exponential or generalized exponential distribution ( ge ) introduced by @xcite .
this distribution is a particular member of the exponentiated weibull ( ew ) distribution introduced by @xcite .
the ge distribution is a right skewed unimodal distribution , the density function and hazard function of the exponentiated exponential distribution are quite similar to the density function and hazard function of the gamma distribution .
its applications have been wide - spread as model to power system equipment , rainfall data , software reliability and analysis of animal behavior .
recently @xcite proposed a new class of univariate distributions called the odd generalized exponential ( oge ) family and studied each of the oge - weibull ( oge - w ) distribution , the oge - frchet ( oge - fr ) distribution and the oge - normal ( oge - n ) distribution .
this method is flexible because of the hazard rate shapes could be increasing , decreasing , bathtub and upside down bathtub . in this article
we present a new distribution from the exponentiated exponential distribution and gompertz distribution called the odd generalized exponential - gompertz ( oge - g ) distribution using new family of univariate distributions proposed by @xcite .
a random variable @xmath0 is said to have generalized exponential ( ge ) distribution with parameters @xmath1 if the cumulative distribution function ( cdf ) is given by @xmath2 the odd generalized exponential family by @xcite is defined as follows .
let @xmath3 is the cdf of any distribution depends on parameter @xmath4 and thus the survival function is @xmath5 , then the cdf of oge - family is defined by replacing @xmath6 in cdf of ge in equation ( [ 1 ] ) by @xmath7 to get @xmath8 ^{\beta } , \;x>0 , \alpha > 0,\epsilon > 0 , \beta>0.\ ] ] this paper is outlined as follows . in section 2 , we define the cumulative distribution function , density function , reliability function and hazard function of the odd generalized exponential - gompertz ( oge - g ) distribution . in section 3
, we introduce the statistical properties include , the quantile function , the mode , the median and the moments .
section 4 discusses the distribution of the order statistics for ( oge - g ) distribution .
moreover , maximum likelihood estimation of the parameters is determined in section 5 .
finally , an application of oge - g using a real data set is presented in section 6 .
in this section we define new four parameters distribution called odd generalized exponential - gompertz distribution with parameters @xmath9 written as oge - g(@xmath10 ) , where the vector @xmath10 is defined by @xmath11 .
a random variable @xmath0 is said to have oge - g with parameters @xmath9 if its cumulative distribution function given as follows @xmath12 } \right\}^{\beta } , \;x>0,\alpha , \lambda , c , \beta>0,\ ] ] where @xmath13 are scale parameters and @xmath14 is shape parameter .
if a random variable @xmath0 has cdf in ( [ 3 ] ) then the corresponding probability density function is @xmath15 } \left\ { 1-e^{-\alpha \left[e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^{\beta -1 } , \;\ ] ] where @xmath16 a random variable @xmath17 has survival function in the form @xmath18 } \right\}^\beta.\ ] ] the failure rate function of @xmath19 ) is given by @xmath20 } \left\ { 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^{\beta -1}}{1-\left\ { 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right]}\right\}^\beta}.\ ] ] + figure 1 : the cdf of various oge - g distributions for some values of the parameters .
+ figure 2 : the pdf s of various oge - g distributions for some values of the parameters . + figure 3 : the hazard of various oge - g distributions for some values of the parameters .
in this section , we study some statistical properties of oge - g , especially quantile , median , mode and moments .
the quantile of oge - g(@xmath10 ) distribution is given by using @xmath21 substituting from ( [ 3 ] ) into ( [ quan ] ) , @xmath22 can be obtained as @xmath23 \right\ } , \;0<q<1.\ ] ] setting @xmath24 in ( [ 7 ] ) , we obtain the median of oge - g(@xmath25 ) distribution as follows @xmath26 \right\}.\ ] ] in this subsection , we will derive the mode of the oge - g(@xmath10 ) distribution by deriving its pdf with respect to @xmath6 and equal it to zero thus the mode of the oge - g(@xmath10 ) distribution can be obtained as a nonnegative solution of the following nonlinear equation @xmath27 \left\ { 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\ } + \nonumber\\ & & \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \alpha \lambda ( \beta -1)e^{cx}e^{\frac{\lambda } { c}(e^{cx}-1)}e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } = 0.\end{aligned}\ ] ] it is not possible to get an explicit solution of ( [ 9 ] ) in the general case .
numerical methods should be used such as bisection or fixed - point method to solve it .
moments are necessary and important in any statistical analysis , especially in applications .
it can be used to study the most important features and characteristics of a distribution ( e.g. , tendency , dispersion , skewness and kurtosis ) . in this subsection
, we will derive the rth moments of the oge - g(@xmath10 ) distribution as infinite series expansion .
[ theorem1 ] if @xmath28 , where @xmath29 , then the rth moment of @xmath0 is given by @xmath30 the rth moment of the random variable @xmath0 with pdf @xmath31 is defined by @xmath32 substituting from ( [ 4 ] ) into ( [ 10 ] ) , we get @xmath33 } \left\ { 1-e^{\mathbf{-\alpha } \left [ \mathbf{e}^{\frac{\lambda } { c}(e^{cx}-1)}\mathbf{-1}\right ] } \right\ } ^{\beta -1}dx.\ ] ] since @xmath34 } < 1 $ ] for @xmath35 , we have @xmath36 } \right\}^{\beta -1}=\sum_{i=0}^\infty \begin{pmatrix } \beta -1 \\i \end{pmatrix } ( -1)^{i}e^{-\alpha i\left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } .\ ] ] substituting from ( [ 12 ] ) into ( [ 11 ] ) , we obtain @xmath37 } dx.\ ] ] using the series expansion of @xmath38 , we get @xmath39 using binomial expansion of @xmath40 , we get @xmath41 using series expansion of @xmath42 , we get @xmath43 by using the definition of gamma function in the form @xmath44 thus we obtain the moment of oge - g(@xmath45 ) as follows @xmath46 this completes the proof .
let @xmath47 be a simple random sample of size @xmath48 from oge - g(@xmath45 ) distribution with cumulative distribution function @xmath49 and probability density function @xmath50 given by ( [ 3 ] ) and ( [ 4 ] ) respectively .
let @xmath51 denote the order statistics obtained from this sample .
the probability density function of @xmath52 is given by @xmath53^{r-1 } \left [ 1-f(x,\theta ) \right]^{n - r } f(x;\theta ) , \ ] ] where @xmath54 and @xmath55 are the pdf and cdf of oge - g(@xmath10 ) distribution given by ( [ 3 ] ) and ( [ 4 ] ) respectively and @xmath56 is the beta function . since @xmath57 for @xmath58 , we can use the binomial expansion of @xmath59 ^{n - r}$ ] given as follows @xmath60^{n - r}=\sum_{i=0}^{n - r } \begin{pmatrix }
i \end{pmatrix } ( -1)^i \left [ f(x,\theta ) \right]^i.\ ] ] substituting from ( [ 14 ] ) into ( [ 13 ] ) , we have @xmath61^{i+r-1}.\ ] ] substituting from ( [ 3 ] ) and ( [ 4 ] ) into ( [ 15 ] ) , we obtain @xmath62 thus @xmath63 defined in ( [ 16 ] ) is the weighted average of the oge - g distribution with different shape parameters .
now , we determine the maximum - likelihood estimators ( mle s ) of the oge - g parameters .
let @xmath47 be a random sample of size @xmath48 from oge - g(@xmath64 ) , where @xmath65 , then the likelihood function @xmath66 of this sample is defined as @xmath67 substituting from ( [ 4 ] ) into ( [ 15 ] ) , we get @xmath68 } \left\ { 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^{\beta -1}.\ ] ] the log - likelihood function is given as follow @xmath69}\right\}.\end{aligned}\ ] ] the log - likelihood can be maximized either directly or by solving the nonlinear likelihood equations obtained by differentiating equation ( [ 18 ] ) with respect to @xmath70 and @xmath14 .
the components of the score vector @xmath71 are given by @xmath72 } \right\ } , \\
\label{20 } \frac{\partial l}{\partial \alpha } & = & \frac{n}{\alpha } -\sum_{i=1}^n(e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1)-(\beta -1)\sum_{i=1}^n \frac { 1-e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}}{e^{\alpha \left [ e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1\right]}-1 } , \\
\label{21 } \frac{\partial l}{\partial \lambda } & = & \frac{n}{\lambda } + \frac{1}{c}\sum_{i=1}^n(e^{cx_{i}}-1)-\frac{\alpha } { c } \sum_{i=1}^n ( e^{cx_{i}}-1)e^{\frac{\lambda } { c}(e^{cx_{i}}-1 ) } + \frac{\alpha ( \beta -1)}{c } \sum_{i=1}^n \frac{\left ( e^{cx_{i}}-1\right ) e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}}{e^{\alpha\left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right]}-1 } , \nonumber\\ & & \\ \label{22 } \frac{\partial l}{\partial c } & = & \sum_{i=1}^n x_{i}-\sum_{i=1}^n \tau(x_i,\lambda , c ) - \alpha \sum_{i=1}^n \tau(x_i,\lambda , c ) e^{\frac{\lambda } { c } ( e^{cx_{i}}-1 ) } + ( \beta -1)\alpha \sum_{i=1}^n \frac{\tau(x_i,\lambda , c ) e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}}{e^{\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } -1 } , \nonumber\\ & & \end{aligned}\ ] ] where @xmath73 the normal equations can be obtained by setting the above non - linear equations ( [ 19])-([22 ] ) to zero .
that is , the normal equations take the following form @xmath74 } \right\ } = 0 , \\ \label{24 } & & \frac{n}{\alpha } -\sum_{i=1}^n(e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1)-(\beta -1)\sum_{i=1}^n \frac { 1-e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}}{e^{\alpha \left [ e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1\right]}-1 } = 0 , \\ \label{25 } & & \frac{n}{\lambda } + \frac{1}{c}\sum_{i=1}^n(e^{cx_{i}}-1)-\frac{\alpha } { c } \sum_{i=1}^n ( e^{cx_{i}}-1)e^{\frac{\lambda } { c}(e^{cx_{i}}-1 ) } + \frac{\alpha ( \beta -1)}{c } \sum_{i=1}^n \frac{\left ( e^{cx_{i}}-1\right ) e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}}{e^{\alpha\left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right]}-1 } = 0 , \nonumber\\ & &
\\ \label{26 } & & \sum_{i=1}^n x_{i}-\sum_{i=1}^n \tau(x_i,\lambda , c ) - \alpha \sum_{i=1}^n \tau(x_i,\lambda , c ) e^{\frac{\lambda } { c } ( e^{cx_{i}}-1 ) } + ( \beta -1)\alpha \sum_{i=1}^n \frac{\tau(x_i,\lambda , c ) e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}}{e^{\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } -1 } = 0 .
\nonumber\\ & & \end{aligned}\ ] ] the normal equations do not have explicit solutions and they have to be obtained numerically . from equation ( [ 23 ] ) the mles of @xmath75 can be obtained as follows @xmath76 } \right\ } } .\ ] ] substituting from ( [ 27 ] ) into ( [ 24 ] ) , ( [ 25 ] ) , and ( [ 26 ] ) , we get the mles of @xmath70 by solving the following system of non - linear equations @xmath77 } -1}=0 , \\ \label{29 } & & \frac{n}{\hat{\lambda}}+\frac{1}{\hat{c } } \sum_{i=1}^n(e^{\frac{\hat{\lambda}}{\hat{c}}(e^{\hat{c}x_{i}}-1)}-1)- \frac{\hat{\alpha}}{\hat{c } } \sum_{i=1}^n ( e^{\hat{c } x_{i}}-1)e^{\frac{\hat{\lambda}}{\hat{c}}(e^{\hat{c}x_{i}}-1 ) } + \frac{\hat{\alpha}(\hat{\beta}-1)}{\hat{c } } \sum_{i=1}^n \frac { ( e^{\hat{c}x_{i}}-1 ) e^{\frac{\hat{\lambda}}{\hat{c}}(e^{\hat{c}x_{i}}-1 ) } } { e^{\hat{\alpha}\left [ e^{\frac{\hat{\lambda } } { \hat{c}}(e^{\hat{c}x_{i}}-1)}-1\right ] } -1}=0 , \nonumber\\ & & \\
\label{30 } & & \sum_{i=1}^n x_{i}- \sum_{i=1}^n \tau(x_i,\hat{\lambda } , \hat{c } ) -\hat{\alpha } \sum_{i=1}^n\tau(x_i,\hat{\lambda } , \hat{c } ) e^{\frac{\hat{\lambda}}{\hat{c}}(e^{\hat{c } x_{i}}-1 ) } + ( \hat{\beta}-1)\hat{\alpha}\sum_{i=1}^n \frac{\tau(x_i , \hat{\lambda } , \hat{c } ) e^{\frac{\hat{\lambda}}{\hat{c}}(e^{\hat{c}x_{i}}-1)}}{e^{\hat{\alpha}\left [ e^{\frac{\hat{\lambda}}{\hat{c}}(e^{\hat{c}x_{i}}-1)}-1\right ] } -1}=0 , \nonumber\\ & & \end{aligned}\ ] ] where @xmath78 these equations can not be solved analytically and statistical software can be used to solve the equations numerically .
we can use iterative techniques such as newton raphson type algorithm to obtain the estimate @xmath79 . in this subsection , we derive the asymptotic confidence intervals of the unknown parameters @xmath80 when @xmath81 and @xmath82 .
the simplest large sample approach is to assume that the mles(@xmath80 ) are approximately multivariate normal with mean ( @xmath80 ) and covariance matrix @xmath83 , where @xmath83 is the inverse of the observed information matrix which defined as follows @xmath84 ^{-1 } \nonumber\\ & = & \left [ \begin{array}{cccc } var(\hat{\alpha } ) & cov(\hat{\lambda } , \hat{\alpha } ) & cov(\hat{c } , \hat{\alpha } ) & cov(\hat{\beta } , \hat{\alpha } ) \\
cov(\hat{\alpha } , \hat{\lambda } ) & var(\hat{\lambda } ) & cov(\hat{c } , \hat{\lambda } ) & cov(\hat{\beta } , \hat{\lambda } ) \\
cov(\hat{\alpha } , \hat{c } ) & cov(\hat{\lambda } , \hat{c } ) & var(\hat{c } ) & cov(\hat{\beta } , \hat{c } ) \\
cov(\hat{\alpha } , \hat{\beta } ) & cov(\hat{\lambda } , \hat{\beta } ) & cov(\hat{c } , \hat{\beta } ) & var(\hat{\beta } ) \end{array}\right ] .\end{aligned}\ ] ] the second partial derivatives included in @xmath83 are given as follows @xmath85 } } { 1-e^{-\alpha \left[e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1\right ] } } , \\
\frac{\partial^2 l } { \partial \lambda \partial \beta } & = & \frac{\alpha } { c } \sum_{i=1}^n \frac{\left ( e^{cx_{i}}-1\right ) a_{i}}{1-e^{-\alpha\left [ e^{\frac{\lambda}{c}(e^{cx_{i}}-1)}-1\right ] } } , \\
l } { \partial c \partial \beta } & = & \alpha \sum_{i=1}^n \frac{a_{i}b_{i}}{1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1\right ] } } , \\
\frac{\partial^2 l}{\partial \alpha^2 } & = & \frac{-n } { \alpha ^{2}}-(\beta -1 ) \sum_{i=1}^n\frac{\left ( e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1\right ) ^{2}e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-1\right ] } } { \left\ { 1-e^{-\alpha\left [ e^{\frac{\lambda } { c}(e^{cx}-1 ) } -1\right ] } \right\}^2 } , \\
\frac{\partial^2 l } { \partial \lambda \partial \alpha } & = & \frac{-1}{c}\sum_{i=1}^n \left ( e^{cx_{i}}-1\right ) e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}+\frac{(\beta-1)}{c } \sum_{i=1}^n \frac{\left ( e^{cx_{i}}-1\right ) a_{i}c_{i}}{\left\ { 1-e^{-\alpha\left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^2 } , \\
\frac{\partial^2 l } { \partial c \partial \alpha } & = & - \sum_{i=1}^n b_{i}e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}+(\beta -1)\sum_{i=1}^n \frac{a_{i}b_{i}c_{i}}{\left\ { 1-e^{-\alpha\left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^2 } , \\
\frac{\partial^2 l}{\partial \lambda^2 } & = & \frac{-n}{\lambda ^{2}}-\frac{\alpha } { c^{2 } } \sum_{i=1}^n \left ( e^{cx_{i}}-1\right ) ^{2}e^{\frac{\lambda } { c}(e^{cx_{i}}-1 ) } + \frac{\alpha { \small ( \beta -1)}}{c^{2 } } \sum_{i=1}^n \frac{\left ( e^{cx_{i}}-1\right ) ^{2}a_{i}d_{i}}{\left\ { 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right]}\right\}^2 } , \\
\frac{\partial^2 l } { \partial c \partial \lambda } & = & \frac{1}{\lambda } \sum_{i=1}^n b_i- \frac{\alpha}{c } \sum_{i=1}^n \left [ \frac{-1}{c}(e^{cx_i}-1)+ x_i e^{cx_i}+ b_i \left ( e^{cx_i}-1 \right ) \right ] e^{\frac{\lambda } { c}(e^{cx_i}-1 ) } \nonumber\\ & & + \frac{(\beta -1)\alpha } { \lambda } \sum_{i=1}^n \frac{\left [ 1-e^{-\alpha \left [ e^{\frac{\lambda}{c}(e^{cx_{i}}-1 ) } -1\right ] } + \frac{\lambda } { c}(e^{cx_{i}}-1)d_{i}\right ] a_i b_i}{\left\ { 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^2},\end{aligned}\ ] ] @xmath86 e^{\frac{\lambda } { c}(e^{cx_{i}}-1 ) } + \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \\ & & ( \beta -1)\alpha \sum_{i=1}^n \frac{\left [ b_i^2 d_i + e_i \left ( 1-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx_{i}}-1 ) } -1 \right ] } \right ) \right ] a_i}{\left\{1-e^{-\alpha\left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } \right\}^2},\end{aligned}\ ] ] where @xmath87 } , \qquad b_i = -\frac{\lambda } { c^{2}}(e^{cx_{i}}-1)+\frac{\lambda } { c } ( x_{i}e^{cx_{i}}),$ ] @xmath88 } , \qquad d_i = 1-\alpha e^{\frac{\lambda } { c}(e^{cx_{i}}-1)}-e^{-\alpha \left [ e^{\frac{\lambda } { c}(e^{cx}-1)}-1\right ] } , \ ] ] @xmath89 the asymptotic @xmath90 confidence intervals of @xmath91 and @xmath14 are @xmath92 and @xmath93 respectively , where @xmath94 is the upper ( @xmath95)th percentile of the standard normal distribution .
in this section we perform an application to real data to illustrate that the oge - g can be a good lifetime model , comparing with many known distributions such as the exponential , generalized exponential , gompertz , generalized gompertz and beta gompertz distributions ( ed , ge , g , gg , bg ) , see @xcite .
consider the data have been obtained from aarset @xcite , and widely reported in some literatures , see @xcite .
it represents the lifetimes of 50 devices , and also , possess a bathtub - shaped failure rate property , table 1 .
table 1 : the data from aarset @xcite . [ cols= "
> , > , > , > , > , > , > , > , > , > , > , > , > , > , > , > , > , > " , ] based on tables 2 and 3 , it is shown that oge - g@xmath96 model is the best among of those distributions because it has the smallest value of ( k - s ) , aic , caic and bic test . substituting the mle s of the unknown parameters @xmath80 into ( [ 31 ] ) , we get estimation of the variance covariance matrix as the following @xmath97\ ] ] the approximate 95% two sided confidence intervals of the unknown parameters @xmath70 and @xmath75 are @xmath98 , \ ; \left [ 0,0.001\right ] , \ ; \left [ 0.060,0.096\right ] , \ ; \left [ 0.116,0.270\right]$ ] , respectively .
+ to show that the likelihood equation have unique solution , we plot the profiles of the log - likelihood function of @xmath99 and @xmath100 in figures 4 - 5 . + figure 4 : the profile of the log - likelihood function of @xmath101 .
+ figure 5 : the profile of the log - likelihood function of @xmath102 .
the nonparametric estimate of the survival function using the kaplan - meier method and its fitted parametric estimations when the distribution is assumed to be @xmath103 and @xmath104 are computed and plotted in figure 6 .
+ figure 6 : the kaplan - meier estimate of the survival function .
figure 7 , gives the form of the hazard rate for the @xmath105 and @xmath104 which are used to fit the data after replacing the unknown parameters included in each distribution by their mle .
+ figure 7 : the failure rate function for the data .
in this paper , we propose a new model , called the odd generalized exponential - gompertz ( oge - g ) distribution and studied its different properties .
some statistical properties of this distribution have been derived and discussed .
the quantile , median , mode and moments of oge - g are derived in closed forms . the distribution of the order statistics are discussed .
both point and asymptotic confidence interval estimates of the parameters are derived using the maximum likelihood method and we obtained the observed fisher information matrix .
we use application on set of real data to compare the oge - g with other known distributions such as exponential ( e ) , generalized exponential ( ge ) , gompertz ( g ) , generalized gompertz ( gg ) and beta - gompertz ( bg ) .
applications on set of real data showed that the oge - g is the best distribution for fitting these data sets compared with other distributions considered in this article .
tahir , m.h . ,
cordeiro , g.m . ,
alizadeh , m. , mansoor , m. , zubair , m. and hamedani , g. g. ( 2015 ) . the odd generalized exponential family of distributions with applications .
_ journal of statistical distributions and applications _ , 2(1 ) , 128 .
|
in this paper we propose a new lifetime model , called the odd generalized exponential gompertz distribution , we obtained some of its mathematical properties . some structural properties of the new distribution
are studied .
the method of maximum likelihood is used for estimating the model parameters and the observed fisher s information matrix is derived .
we illustrate the usefulness of the proposed model by applications to real data . ' '' '' * keywords : * _ gompertz distribution ; hazard function ; moments ; maximum likelihood estimation . _ * 2010 msc : * _ 60e05 , 62f10 , 62f12 , 62n02 , 62n05 . _
|
recent years have witnessed many breakthroughs in research on two - dimensional ( 2d ) materials due to their potential applications in next - generation electronic and energy conversion devices @xcite .
recently , a new type of 2d material , borophene ( 2d boron sheet ) @xcite , has been successfully grown on single crystal ag(111 ) substrates by two parallel experiments @xcite .
although various proposals of stable 2d boron sheets and quasiplanar boron clusters have been made @xcite , three main structures ( @xmath0 , @xmath1 and striped borophene ) have been observed by scanning tunneling microscopy in these two experiments : @xmath0 and @xmath1 borophene has planar structure with periodic holes @xcite , while striped borophene has buckled structure with anisotropic corrugation @xcite .
the following first principles calculations have predicted that striped borophene possesses remarkable mechanical properties @xcite , which may rival graphene @xcite .
however , phonon instability in striped borophene is observed @xcite , which may challenge previous results demonstrating that borophene is stiffer than graphene along @xmath2 direction @xcite . theoretical investigation of the formation of boron sheet on ag(111 ) surface has demonstrated that stable boron sheet should contain 1/6 vacancies in a striped pattern @xcite .
this is consistent with previous theoretical studies indicating that planar boron sheets with vacancies are more stable @xcite .
the ground state of 2d boron is in debate @xcite .
these debates raise several questions : ( i ) are these three structures stable ?
( ii ) what is the relative stability of @xmath0 , @xmath1 and striped borophene ? ( iii ) could these structures possess high hardness in thermodynamic aspect ?
in fact , due to the structural complexity of boron , even the relative stability of @xmath3 and @xmath4rhombohedral boron has been discussed for over 30 years @xcite . thus a systematic investigation is needed to examine the stability and strength of these three structures . when discussing the stability of crystal structures , it is important to distinguish dynamical , thermodynamical and mechanical stability @xcite . a material is dynamically stable when no imaginary phonon frequencies exist .
thermodynamical stability can be described by the helmholtz free energy @xcite , which demonstrates how phonons determine the relative stability at finite temperatures . regarding mechanical stability , the born - huang criteria for elastic constants
must be fullfilled @xcite .
the born - huang mechanical stability criteria provide a _ necessary _ condition for the dynamical stability , but not a _ sufficient _ one @xcite . therefore , although the mechanical properties of striped borophene have been studied intensively @xcite , the questions of stability still remain unanswered . in this work , the dynamical , thermodynamical and mechanical stabilities of @xmath0 , @xmath1 and striped borophene
are evaluated using first principles calculations .
we further study for the first time the mechanical properties of @xmath0 and @xmath1 borophene , and compare the stiffness of three distinguished structures .
the bonding characteristics are also examined to understand the stability and strength of striped borophene .
three structures of borophene are shown in fig .
[ structure ] . for @xmath0 ( @xmath1 ) borophene , five ( four ) b atoms in the unit cell are arranged in the same plane . for striped borophene , there is no corrugation along @xmath2 direction , while a vertical buckling along the @xmath5 direction is observed .
the major difference between striped borophene and the other two structures is the absence of vacancies .
the introduction of vacancies leads to lower cohesive energy , as shown in table [ lattice ] , which is in good agreement with previous results @xcite .
the optimized geometries of three structures are listed in table [ lattice ] .
the predicted @xmath2 of striped borophene corresponds to 1/3 the @xmath2 observed in the experiment , while the lattice constants of @xmath0 and @xmath1 borophene is in excellent agreement with experimental results @xcite .
thus the reliability of the present calculations is confirmed .
[ band ] presents the calculated band structures of all three phases , which agree well with previous theoretical results @xcite .
it should be noticed that due to highly anisotropic crystal structure , striped borophene shows anisotropic metallic behaviour . [ cols="^,^,^,^,^,^,^ " , ] [ stiff ] using the elastic tensor , the mechanical properties such as 2d young s modulus @xmath6 and the corresponding poisson s ratio @xmath7 can be calculated @xcite , as shown in table [ stiff ] .
the calculated @xmath6 of striped borophene are in good agreement with other theoretical data @xcite .
2d young s modulus is defined as the ratio between stress and strain , and provides a measure of in - plane stiffness of the solid materials .
the mechanical properties of @xmath0 and @xmath1 borophene are similar due to similar in - plane bonding and atomic mass density .
as shown in table [ stiff ] , the stiffness of @xmath0 and @xmath1 borophene is lower than that of graphene and monolayer bn , but higher than that of silicene @xcite .
negative possion s ratio is observed in striped borophene , which is due to highly buckled structure @xcite .
in addition , the stiffness for striped borophene along @xmath2 direction is much higher than that along @xmath5 direction , and even rivals graphene ( 342.2 gpa@xmath8 nm ) @xcite .
in fact , although striped borophene is much stiffer than @xmath0 and @xmath1 borophene , thermodynamically , high stiffness means unstability due to increasing helmholtz free energy via the increase in vibration frequency @xcite .
both the high stiffness and high instability of striped borophene along @xmath2 direction can be attributed to strong directional bonding . to understand the chemical bonding of striped borophene
, we calculate its electron localization function ( elf ) @xcite in comparison with @xmath0 borophene , which has similar 2d orthorhombic structure . a higher value of elf corresponds to higher electron localization . as shown in fig .
[ elf ] , elf profiles of striped borophene is more anisotropic than that of @xmath0 borophene . strong directional bonding may prevent dislocations from forming to accommodate strains and thereby cause the material to be brittle @xcite .
this is consistent with previous theoretical studies showing that striped borophene is dynamically unstable even under high tensile stress @xcite . as for @xmath0 borophene ,
strong b - b bonds along different directions in the 2d plane of @xmath0 borophene stabilize the structure . thus @xmath0 borophene is more stable than striped borophene considering the bonding characteristics . , ( b ) @xmath1 and ( c ) striped borophene . ] we further consider the nature of the electronic bonding by calculating the projected density of states ( dos ) for three structures of borophene with separated in - plane ( @xmath9 , @xmath10 and @xmath11 ) and out - of - plane ( @xmath12 ) projections in fig . [ dos ] .
generally , in - plane @xmath13 bonds ( @xmath14 bonds ) are stronger than @xmath15 bonds ( @xmath16 bonds ) derived from @xmath12 orbitals . for striped borophene ,
the @xmath12 projected dos vanishes from -3.5 to 2.5 ev . as a result ,
localized @xmath14 bonds along @xmath2 direction are observed in elf profiles .
however , these strong @xmath14 bonds will be destabilized by any flattening of the boron sheet , leading to highly metastable structure @xcite .
thus striped borophene becomes instable under tension along @xmath2 direction @xcite .
in addition , some of the strong in - plane @xmath13 bonding states are unoccupied in fig .
[ dos](c ) , and subsequently striped borophene tends to accepting electrons to increase its stability @xcite .
in conclusion , first - principles calculations are performed on 2d borophene sheet to evaluate the dynamical , thermodynamical and mechanical stability of two distinguished structures .
our results show that the free - standing @xmath0 and @xmath1 borophene is thermodynamically , mechanically and dynamically stable , while striped borophene is thermodynamically and dynamically unstable due to high stiffness along @xmath2 direction .
our calculated elf shows that the bonding characteristic of striped borophene leads to high stiffness and high instability at the same time .
this work is supported by the national natural science foundation of china under grants no .
11374063 and 11404348 , and the national basic research program of china ( 973 program ) under grant no . 2013cba01505 .
71ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1038/nature04235 [ * * , ( ) ] link:\doibase 10.1021/nl0731872 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.107.076802 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.111.136804 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.226801 [ * * , ( ) ] in @noop _ _ ( , ) pp . in @noop
_ _ ( , ) pp . @noop * * , ( ) link:\doibase 10.1039/c4nr01600a [ * * , ( ) ] link:\doibase 10.1038/srep20225 [ * * , ( ) ] link:\doibase 10.1039/c5ra19747c [ * * , ( ) ] link:\doibase 10.1002/andp.201500354 [ * * , ( ) ] link:\doibase 10.1038/ncomms4113 [ * * , ( ) ] link:\doibase 10.1126/science.aad1080 [ * * , ( ) ] link:\doibase 10.1038/nchem.2491 [ * * , ( ) ] link:\doibase 10.1103/physrevb.55.16426 [ * * , ( ) ] link:\doibase 10.1002/anie.200351874 [ * * , ( ) ] link:\doibase 10.1038/nmat1012 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.115501 [ * * , ( ) ] link:\doibase 10.1021/jp066719w [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.041402 [ * * , ( ) ] link:\doibase 10.1021/nl3004754 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1002/anie.201207972 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.085502 [ * * , ( ) ] link:\doibase 10.1038/nchem.1999 [ * * , ( ) ] link:\doibase 10.1039/c5nr04359j [ * * , ( ) ] link:\doibase 10.1039/c5ra14939h [ * * , ( ) ] link:\doibase 10.1002/ange.201505425 [ * * , ( ) ] link:\doibase 10.1080/0144235x.2016.1147816 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.085406 [ * * , ( ) ] link:\doibase 10.1080/21663831.2016.1174163 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/18/7/073016 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1039/c6tc00115 g [ * * , ( ) ] link:\doibase 10.1007/s12274 - 016 - 1148 - 0 [ ( ) ] link:\doibase 10.1021/jp305545z [ * * , ( ) ] link:\doibase 10.1063/1.4774082 [ * * , ( ) ] link:\doibase 10.1021/nn302696v [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1039/c6tc01328 g [ * * , ( ) ] link:\doibase 10.1088/0022 - 3719/15/3/008 [ * * , ( ) ] link:\doibase 10.1021/ja003233z [ * * , ( ) ] link:\doibase 10.1021/ja0026962 [ * * , ( ) ] link:\doibase 10.1016/s0925 - 8388(02)00764 - 8 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.195102 [ * * , ( ) ] link:\doibase 10.1021/cr300356 t [ * * , ( ) ] link:\doibase 10.1038/srep02351 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.255502 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.184102 [ * * , ( ) ] link:\doibase 10.1021/ja0631246 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.76.054115 [ * * , ( ) ] link:\doibase 10.1021/acs.nanolett.6b00070 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.134106 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.scriptamat.2015.07.021 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.104102 [ * * , ( ) ] link:\doibase 10.1103/physrevb.57.10421 [ * * , ( ) ] link:\doibase 10.1021/acsnano.5b03394 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.125428 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.458517 [ * * , ( ) ] link:\doibase 10.1002/anie.199201871 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.4236/mnsms.2013.31002 . [ * * , ( ) ] link:\doibase 10.1557/jmr.1990.0971 [ * * , ( ) ] link:\doibase 10.1007/bf02744780 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.035413 [ * * , ( ) ]
|
a new two - dimensional ( 2d ) material , borophene ( 2d boron sheet ) , has been grown successfully recently on single crystal ag substrates by two parallel experiments [ mannix _ et al . ,
science _ , 2015 , * 350 * , 1513 ] [ feng _ et al . ,
nature chemistry _ , 2016 , * advance online publication * ] .
three main structures have been proposed ( @xmath0 , @xmath1 and striped borophene ) .
however , the stability of three structures is still in debate . using first principles calculations
, we examine the dynamical , thermodynamical and mechanical stability of @xmath0 , @xmath1 and striped borophene .
free - standing @xmath0 and @xmath1 borophene is dynamically , thermodynamically , and mechanically stable , while striped borophene is dynamically and thermodynamically unstable due to high stiffness along @xmath2 direction .
the origin of high stiffness and high instability in striped borophene along @xmath2 direction can both be attributed to strong directional bonding .
this work provides a benchmark for examining the relative stability of different structures of borophene .
|
applying an electric gate voltage to graphene provides exotic tuning of the electronic , vibrational , and optical properties of graphene samples @xcite . since the beginning of graphene s discovery
, electronic gating has played an important role in elucidating the room temperature quantum hall effect @xcite , the klein tunneling @xcite , and many body coupling effects @xcite .
similar gating techniques are extensively applied not only to monolayer , but also to multilayer graphene to obtain tunable transport @xcite , a tunable band gap @xcite , p - n junctions @xcite , and photodetectors @xcite .
all of these exciting phenomena could be observed due to the ability of tuning graphene s fermi energy @xmath0 through the applied gate voltage .
a combination of electronic gating and inelastic scattering of light , known as the gate modulated raman spectroscopy @xcite , opens up a new possibility to understand more thouroughly the interplay of the electron , photon and phonon exitations in graphene .
several phenomena have been probed by gate modulated raman spectroscopy in graphene , such as the kohn anomaly ( ka ) effect or the phonon frequency renormalization @xcite , the quantum interference effect @xcite , electron - electron interaction @xcite , and the fano resonance in the raman spectra of graphene @xcite .
studying gate modulated raman spectroscopy in graphene has also enriched our knowledge of phonon spectra characterization @xcite , experimental evaluation of electron - phonon coupling @xcite , and various edge characterization effects @xcite .
some theoretical works have been previously performed to understand the kohn anomaly ( ka ) effect for the first - order raman ( g band ) spectra with a raman shift of @xmath1@xmath2 in graphene , such as those by ando and the mauri s groups @xcite . in the ka process
, phonon renormalization occurs through the excitation of an electron - hole pair by the electron - phonon interaction . as a result , the phonon energy is modified and the phonon lifetime becomes shorter .
the previous theories mention that the phonon frequency shows a logarithmic singularity at @xmath3 k when the absolute value of the fermi energy @xmath0 matches half of the phonon energy @xmath4 . for @xmath5 ,
the frequency shift is linearly proportional to @xmath6 .
these predictions were already confirmed by raman measurements @xcite .
recently , additional experimental results allow us to study the ka effect in the second - order raman spectra , also .
in contrast to the first - order raman spectra that consist of only a single @xmath7 value of the phonon momentum , the second - order raman spectra deals with the whole range of phonon momenta in the brillouin zone satisfying the double resonance raman condition @xcite .
raman spectral features such as the @xmath8 or 2d band ( @xmath9 ) and the @xmath10 or @xmath11 band ( @xmath12 ) are observed as the second - order raman spectra for @xmath13 .
the nonzero momentum phonon leads to a different ka effect compared with that for the @xmath7 phonon .
araujo _ et al . _ and mafra _ et al . _
have shown that the frequency shift of the @xmath8 band as a function of @xmath0 is monotonically decreasing as a function of @xmath6 which is opposite to that of the g band @xcite .
the other band at @xmath14 is , however , dispersionless as a function of @xmath0 .
yan _ et al .
_ show opposite results , that the @xmath8 band frequency as a function of @xmath0 has the same trends as that of the g band @xcite .
further , das _ et al .
_ show an asymmetric @xmath8 band dispersion , i.e. , a positive ( negative ) slope of frequency shift at negative ( positive ) @xmath0 , which is inconsistent with a symmetric dispersion shown experimentally by araujo _ et al .
_ @xcite . based on the controversies in experimental results , we present calculated results of the second - order raman spectra as a function of @xmath0 from which we understand the origin of the controvertial results .
sasaki et al , attempted to understand why the frequency shift of the @xmath8 band ka has an opposite slope when compared with that of the @xmath15 band from the viewpoint of the competition of interband and intraband electron - hole excitation in phonon perturbation @xcite . however , since the theory is done within the effective mass approximation , it is not sufficient to explain the asymmetry of the @xmath8 band frequency shift at positive and negative @xmath0 .
moreover , since the raman intensity as a function of @xmath0 is not calculated , different dispersion of raman peaks as a function of @xmath0 can not be explained from that theory . observing the change of raman intensity as a function of @xmath0 reveals the quantum interference effect . when @xmath16 , some raman scattering paths are suppressed due to the pauli exclusion principle . even with the reduced number of scattering paths , the raman intensity surprisingly increases at a particular value of @xmath0 when constructive raman phases among various scattering paths are enhanced .
chen _ et al .
_ show that the g band raman intensity gives a maximum value when @xmath17 , where @xmath18 is the laser energy @xcite .
however , the theoretical analysis in their work assumes a constant matrix element therefore neglecting the change of the raman phase due to the electron - phonon matrix elements .
previous theoretical calculations show that the electron - phonon matrix elements change sign along electronic equi - energy lines in graphene and therefore can change the raman phase @xcite .
a comprehensive calculation is , therefore , necessary to understand how the quantum interference effect affects the first- and the second - order raman intensity . in this work
we calculate the @xmath0 dependece of the first - order and the second - order raman spectra .
the calculated spectral quantities are the raman peak shift , spectral linewidth , and the raman intensity as a function of @xmath0 .
the ka correction including both the phonon frequency shift and the linewidth is modeled based on second - order perturbation theory .
the ka of the first - order raman spectra or of the @xmath7 phonon is calculated so as to reproduce the existing theoretical and experimental results and to compare with the ka of the @xmath19 phonon.we now focus on the intervalley scatterings which give three prominent peaks in the experimental spectra , namely , the @xmath8 , the @xmath10 , and the @xmath20 bands , and are relevant to @xmath13 .
the @xmath18 dependence of those raman peak positions is compared with experimental results in order to justify the present calculation methods .
finally the @xmath0 dependence of those three raman spectra are analized and compared to the experimental results .
( color online ) schematics of ( a ) the first- and ( b ) the second - order raman process . in ( b ) , ee means two consecutive electron - phonon interactions while eh means electron - phonon interaction followed by hole - phonon interaction.,width=302 ] the organization of this paper is as follows .
section [ subsec : ramanintens ] shows the method for calculating the raman intensities for the first- and second - order raman spectra .
section [ subsec : ka ] explains the method for numerically calculating the ka effect .
section [ subsec : first ] presents the calculated results of the ka effect for the @xmath7 phonon and the @xmath0 dependence of the g band intensity .
section [ subsec : second ] presents the calculation results of the ka effect for @xmath19 , and the @xmath18 and @xmath0 dependences of the second - order raman spectra .
finally , the conclusion is given in sect .
[ sec : conclusion ] .
the first - order raman process as shown in fig.[fig1](a ) consists of ( 1 ) excitation of an electron - hole pair by the electron - photon interaction , ( 2 ) phonon emission by means of the electron - phonon interaction , and ( 3 ) electron - hole recombination and photoemission by the electron - photon interaction . based on the three subprocesses , the raman intensity formula for the first - order raman process
is given by @xmath21 } { ( \elaser - e_\kvec^{\cond\val}-i\gamma ) ( \elaser - e_\kvec^{\cond\val}-\omega^\nu_0-i\gamma)}\right|^2 \nonumber\\ & \times\delta(\elaser-\omega^\nu_0-e_{\rm s } ) , \end{aligned}\ ] ] where @xmath18 is the laser excitation energy , @xmath22 is the scattered photon energy , @xmath23 is the electron energy difference between the conduction ( @xmath24 ) and the valence ( @xmath25 ) bands at a wave vector @xmath26 , @xmath27 is the carrier scattering rate due to the electron - phonon interaction @xcite , and @xmath28 is the fermi distribution function which depends on temperature .
@xmath29 is the electron - photon matrix element acounting for the optical transition of an electron in a state @xmath26 from a valence band to a conduction band , @xmath30 is the carrier - phonon interaction considering an electron ( @xmath31 ) in a conduction band or a hole ( @xmath32 ) in a valence band making a transition from a state @xmath33 to a state @xmath34 by emitting a phonon with momentum @xmath35 , mode @xmath36 , and frequency @xmath37 .
hereafter , @xmath38 is used , so that @xmath37 has units of energy . for the case of a one phonon process , only zero momentum or
the @xmath39 point phonon is relevant .
the summation over @xmath26 in eq . is taken to occur in a uniform square mesh , with a mesh spacing @xmath40 , and @xmath41 is the slope of the electron energy dispersion of graphene and @xmath42 is the weight of the integration .
we set a cutoff energy @xmath43 for @xmath44 so as to reduce the number of mesh points in the brillouin zone integration .
it is important to note that both the numerator and denominator of eq .
are complex numbers , thus the summation of @xmath26 before taking the square plays an important role in the quantum interference effect @xcite .
the electronic structure of graphene is calculated within an extended tight binding method considering up to 20 nearest neighbors and the four atomic orbitals ( @xmath45 , @xmath46 , @xmath47 , @xmath48 ) @xcite .
calculation of the phonon dispersion relations is performed within a force constant model with the interatomic potential including up to 20 nearest neighbors which is fitted from a first - principles calculation @xcite .
figure [ fig - phdisp](a ) shows the calculated results of the phonon dispersion relations ( solid lines ) and the corresponding experimental phonon dispersion relations ( red dots ) for comparison from refs . .
because of the ka effect , the dispersion of the in - plane tangential optic ( ito ) branch near the @xmath49 point is discontinuous along the @xmath50 path which can not be reproduced by the force constant model @xcite .
we fit the ito frequency from the experiment @xcite and use the following fitting formula for the raman spectra calculation [ fig .
[ fig - phdisp](b ) ] : @xmath51 where @xmath52 is defined using polar coordinates @xmath53 whose center is at the k point and @xmath54 is measured from the km direction .
eq . is valid only for @xmath55 , and when @xmath56 , we use the results from the force constant model for @xmath57 .
( color online ) ( a ) calculated ( solid lines , this work ) and experimental ( red dots , from refs . ) phonon dispersion relations .
( b ) fitting of eq .
( blue line ) to the ito branch from ref .
( dots ) near the k point.,width=302 ] in the electron - photon interaction , we adopt a dipole approximation , i.e. , assuming a slowly varying electric field in space because the laser wavelength is much greater than the inter - atomic distance @xcite .
the electron - phonon interaction is calculated using the tight binding method with the deformation potential fitted from the gw method for the k and @xmath39 points @xcite . in the second - order raman processes , phonons with modes @xmath36 and @xmath58 and momenta @xmath52 and @xmath59 , respectively ,
are emitted [ fig .
[ fig1](b ) ] . depending on the carriers taking part in the scattering event ,
the raman intensity formula is given by : @xmath60 where @xmath61 is a raman amplitude for each process : ( 1 ) an electron ( e ) , first , emits a @xmath36 phonon with momentum @xmath52 and , ( 2 ) a hole ( h ) emits the @xmath58 phonon with momentum @xmath59 . here ,
@xmath61 and @xmath62 are not equivalent to each other due to the different time order of the two phonon emission .
next , we show examples of the raman amplitude formula for @xmath63 and @xmath61 : @xmath64 } { ( \elaser - e_\kvec^{\cond\val}-i\gamma ) ( \elaser - e_{\kvec+\qvec}^{\cond}+e_{\kvec}^{\val}-\omega^\nu_{-\qvec}-i\gamma ) ( \elaser - e_{\kvec}^{\cond\val}-\omega^\nu_{-\qvec}-\omega^\mu_{\qvec}-i\gamma)},\ ] ] @xmath65 } { ( \elaser - e_\kvec^{\cond\val}-i\gamma ) ( \elaser - e_{\kvec+\qvec}^{\cond}+e_{\kvec}^{\val}-\omega^\nu_{-\qvec}-i\gamma ) ( \elaser - e_{\kvec+\qvec}^{\cond\val}-\omega^\nu_{-\qvec}-\omega^\mu_{\qvec}-i\gamma)}.\ ] ] the minus sign in eq . corresponds to the opposite charge of the hole from the electron in the hole - phonon matrix elements @xcite .
( color online ) ( a ) a schematic of the @xmath66 kohn anomaly process .
a phonon with zero wave vector ( @xmath66 ) and frequency @xmath67 vertically excites an electron - hole pair via the electron - phonon interaction .
the electron - hole pair then recombines by emitting a phonon with frequency @xmath68 .
( b ) a schematic of the @xmath69 kohn anomaly process .
an electron exists at the @xmath70 point leaving a hole behind at the @xmath49 point with a distance in reciprocal space of @xmath71 .
if the @xmath70 point is then translated by a vector @xmath72 , we can then imagine a virtual vertical transition of electron and hole .
when @xmath73 , both interband ( c ) and intraband ( d ) transitions are expected.,width=302 ] kohn mentions that conduction electrons are able to screen phonons in a metal @xcite .
this screening leads to a phonon frequency change , given by : @xmath74 where @xmath75 is the unperturbed phonon energy from the phonon dispersion relation . here , @xmath76 is the correction term taken from the second - order perturbation of the electron - phonon interaction by the excitation and recombination of an electron - hole pair ( fig .
[ fig2a ] ) : @xmath77 } { \omega^{(0),\nu}_{\qvec}-e_{\kvec+\qvec}^{s'}+e_\kvec^{s}+i\eta},\ ] ] where the prefactor @xmath78 in eq .
accounts for the spin degeneracy , while the valley degeneracy is considered in the summation over @xmath26 .
the value of @xmath76 is a complex number , in which @xmath79 [ @xmath80 gives the phonon frequency shift [ phonon linewidth ] . in eq .
, the contribution of the interband ( intraband ) electron - hole pair appears at @xmath81 ( @xmath82 ) . in a conventional 2d electron gas ,
the ka effect occurs at @xmath83 , where @xmath84 is the fermi wave vector . in graphene , due to its unique linear energy bands , the ka occurs at @xmath85 and @xmath86 .
the schematics of the ka process for @xmath87 and @xmath69 are shown in figs .
[ fig2a](a ) and ( b ) , respectively . in the @xmath88 ka
, a phonon with frequency @xmath67 vertically excites an electron - hole pair via the electron - phonon interaction [ fig .
[ fig2a](a ) ] .
the electron - hole pair then recombines by emitting a phonon with frequency @xmath68 . in the @xmath86 ka
, an electron exists at the @xmath70 point , leaving a hole behind at the @xmath49 point with a distance in reciprocal space @xmath89 [ fig .
[ fig2a](b ) ] .
if the @xmath70 point is translated by a vector @xmath72 , we can imagine a virtually vertical excitation of an electron - hole pair .
when @xmath73 , both the interband [ fig .
[ fig2a](c ) ] and intraband [ fig .
[ fig2a](d ) ] transitions are expected .
( color online ) the calculated ( dotted line ) and experimental ( open circles ) results for the g band peak shift ( a ) and the g band linewidth ( b ) as a function of the fermi energy , respectively , for @xmath90 k.,width=302 ] employing eq . at @xmath7 , we can obtain the frequency shift [ fig .
[ fig2](a ) ] and phonon linewidth [ fig .
[ fig2](b ) ] for the g band as a function of the fermi energy at @xmath91 @xmath49 . in fig . [ fig2 ]
we show the calculated ( dotted line ) and experimental ( open circles ) results @xcite of the g band peak shift and linewidth as a function of the fermi energy , respectively .
the calculated results are in good agreement with the experimental results . in fig .
[ fig2](a ) , we see dips when @xmath92 for the calculation , while the experimental results do not show such dips . these dips are originated from the logarithmic singularities at @xmath3 @xmath49 and are related to interband resonances @xcite . for @xmath93 ,
the g band frequency increases linearly as a function of the fermi energy . at @xmath94
k , the phonon linewidth shows a step function @xmath95 .
the step function indicates that when @xmath93 , the phonon linewidth from the ka effect becomes zero since no excited electron - hole pair meets the resonance condition of eq . .
at finite @xmath96 , on the other hand , the fermi distribution function becomes a smooth function and that is why we get a smooth function of the linewidth as a function of @xmath0 .
it is noted that we add an extrinsic broadening of @xmath97 in our calculations in order to fit with experimental results @xcite in fig .
[ fig2](b ) .
( color online ) calculated results of the real and imaginary parts of the first - order ( a ) lo and ( b ) ito raman amplitudes in a two dimensional brillouin zone near the @xmath49 and @xmath70 points with @xmath98 ev.,width=302 ] next , we calculate the raman spectra of the g band using eq . .
the g band consists of both the @xmath87 longitudinal optic ( lo ) and in - plane - tangential optic ( ito ) modes . in order to understand their contributions to
the raman amplitudes at each @xmath26 point , we plot the real and imaginary parts of the raman amplitude in eq . for lo and ito phonons in fig .
[ fig3](a ) and ( b ) , respectively .
here we use @xmath98 ev and take @xmath99 so as to reduce the total points of integration for saving computational time .
it will be clear that neglecting the contributions from energies above @xmath100 in the integration is reasonable since the raman intensity is quickly decreasing when @xmath101 . in fig .
[ fig3 ] , deformed triangles near the @xmath49 and the @xmath70 points indicate equi - energy lines that match the resonant conditions .
the lower ( higher ) resonant condition corresponds to the scattered ( incident ) resonance when @xmath102 ( @xmath103 ) which is shown by a large amplitude at the inner ( outer ) line .
we see changes in the sign for both the real and imaginary parts of the lo and ito raman amplitudes in both the radial and azimuthal directions .
the change of sign at the radial direction is related to an opposite phase between the scattered resonance and the incident resonance .
meanwhile , the change of sign in the azimuthal direction is related to the sign of the electron - phonon matrix element as reported by jiang et al @xcite .
the lo ( ito ) phonon has a zero matrix element at the 0 ( @xmath104 ) phase .
( color online ) ( a ) calculated results ( black dots ) and experimental results ( blue squares from ref . and
red squares from ref . ) of the g band raman intensity as a function of the reduced fermi energy .
( b ) schematic diagram showing an opposite phase between the incident ( i ) and scattered ( s ) resonances . when @xmath105 , the scattered resonance is suppressed , and therefore , the raman intensity gives a maximum value.,width=302 ] the opposite phase of the scattered resonance to the incident resonance is essential insofar as both terms give destructive interference . therefore , only taking the resonant term for calculating the g band intensity is not sufficient . we need to at least consider up to @xmath106 to get a realistic intensity .
moreover if we plot the raman intensity as a function of the fermi energy as shown in fig .
[ fig4](a ) , it becomes clear that destructive interference between the scattered resonance and the incident resonance can be suppressed when we set the fermi energy close to the laser excitation energy .
when @xmath105 , the scattered resonance can not occur due to the pauli blocking effect [ fig .
[ fig4](b ) ] . therefore , in fig .
[ fig4](a ) we see the largest g band intensity at @xmath105 as pointed out by chen _ et al _ @xcite .
the difference of the intensity at positive and negative @xmath0 which comes from the electron - hole asymmetry has been confirmed by liu _ et al _ @xcite .
anisotropy in the azimuthal direction should give destructive interference , but the effect is negligible . in fig .
[ fig5 ] , we show the calculated results of the @xmath19 ka effect from eq . .
first , let us consider the case of @xmath107 in figs .
[ fig5](b ) and ( e ) . if we compare respectively figs . [ fig5](b ) and ( e ) with figs .
[ fig2](a ) and ( b ) , both the frequency shift and phonon linewidth show the same trends as that of @xmath7 ka because both @xmath7 and @xmath107 are dominated by the interband electron - hole excitation .
the reason why the interband excitation is dominant at @xmath108 , is that the @xmath49 and @xmath70 points coincide upon translation of the @xmath70 point by a vector @xmath109 [ @xmath110 in fig .
[ fig2a](c ) ] . therefore , at the @xmath111 ka ,
only virtually vertical interband excitation , the same as at @xmath87 ka , is possible @xcite .
the previous work did not consider the interband contribution , therefore assigning the @xmath107 phonon frequency shift to be dispersionless as a function of @xmath0 @xcite .
next , if we shift by @xmath112 from @xmath108 , competition between the intraband and interband excitations take place as shown in figs .
[ fig5](a ) , ( c ) , ( d ) , and ( f ) . according to the analytical formula @xcite ,
the intraband contribution to the frequency shift is proportional to @xmath113 by assuming @xmath114 , where @xmath115 is the slope of the linear energy dispersion of graphene which is @xmath116 .
the phonon linewidth is increasing linearly with @xmath6 in the case of the intraband excitation [ figs .
[ fig5](d ) and ( f ) ] because the electron - phonon scattering rate is proportional to the carrier concentration .
the asymmetry at positive and negative @xmath0 is related to electron - hole band asymmetry considered in the tight binding calculation . after considering the ka effect on the @xmath117 phonon , in fig [ fig7 ]
we show the calculated raman spectra from eq . .
figure [ fig7](a ) shows three bands , respectively , assigned as the @xmath8 @xmath118 , @xmath10 @xmath119 , and @xmath20 @xmath120 for @xmath121 .
we confirm the origin of the @xmath8 bands from the overtone of the ito ( 2ito ) modes while the @xmath10 bands come from a combination of ito and la modes .
the major contributions to the @xmath8 intensity come from the @xmath122 and @xmath123 terms as shown by @xmath124 in fig .
[ fig7](b ) .
this confirms the previous calculation that the @xmath125 and @xmath126 terms are negligible [ @xmath127 in fig .
[ fig7](b ) ] because of the quantum interference effect during the @xmath26 integration @xcite .
figure [ fig7](c ) shows the second - order raman intensities for @xmath128 .
the intensities of all these raman bands are inversely proportional to @xmath18 because of the increase of the electron - phonon scattering rate @xmath129 as a function of @xmath18 @xcite . assuming that each band can be represented by a single peak , the @xmath8 , @xmath10 , and @xmath20 peak dispersions as a function of @xmath18 are shown in fig .
[ fig7](d ) .
the @xmath8 band shows a positive slope as a function of @xmath0 , i.e. , @xmath130 in this work , @xmath131 in ref . , and @xmath132 in ref . .
meanwhile the @xmath10 band shows a negative slope , i.e. , @xmath133 in this work and @xmath133 in ref .
and the @xmath20 band slope is @xmath134 in this work , @xmath135 in ref . , and @xmath136 in ref .
[ not shown in fig .
[ fig7](d ) ] .
good agreement between theory and experiment in the slope of the @xmath18 dispersion indicates the reliability of our phonon dispersion used in the calculation .
however , discrepancies with the experiments of about @xmath137 in the @xmath8 and the @xmath20 bands for a given @xmath18 show that the calculated electronic energy dispersion underestimates the experimenal results .
this can be seen insofar as the @xmath8 and @xmath20 peaks at @xmath138 in theory give relatively the same value for @xmath139 in the experiment , thus the present electronic energy dispersion near @xmath18 underestimate the real value by @xmath140
. this might be because we neglect the many body effects in the band calculations .
nevertheless , the overal agreement is sufficient for us to proceed and consider the @xmath0 dependence of the raman intensity for a particular @xmath18 .
figure [ fig8 ] shows the evolution of the second order raman spectra for several values of @xmath0 : @xmath141 ( black solid lines ) , @xmath142 ( red dashed lines ) , @xmath143 @xmath144 ( green dashed lines ) , @xmath145 @xmath146 ( blue dashed lines ) , @xmath147 @xmath148 ( violet dashed lines ) .
we use the same @xmath149 as araujo _ et al _
. figures [ fig8](a ) and ( d ) show that the @xmath8 raman spectrum is redshifted , decreases its peak intensity [ indicated by a blue arrow in fig .
[ fig8](d ) ] and broadens as @xmath0 increases .
these results are consistent with the experiment @xcite .
however , the raman intensity of the @xmath10 and @xmath20 bands dramatically increases especially when @xmath150 [ indicated by the blue arrows in figs .
[ fig8](b ) and ( c ) ] .
we find that the increase of the intensity comes from the quantum interference effect and not from the ka effect .
although the ka effect broadens the second - order raman spectra at a finite @xmath0 , however the broadening order should not be @xmath151 .
figures [ fig8](a ) , ( b ) , and ( c ) show that the spectra where combination of the phonon modes are located ( @xmath152 ) dramatically broadens like a continous spectrum .
the broadening should come from the increased intensity of many different @xmath52 phonon modes that are suppressed at @xmath153 during integration on @xmath26 but now appears at @xmath154 thanks to the suppression of the destructive interference .
these results give clues about how the raman phase governed by the electron - phonon matrix elements distinguishes between the @xmath0 dependence of the raman intensity of the combination modes from that of the overtone modes as is also observed in experiments @xcite .
although @xmath155 is too high for experiments on graphene , one can reduce the @xmath18 to become @xmath156 to satisfy the condition of @xmath157 to get proper conditions for observing the quantum interference effect .
the calculated raman spectra at @xmath158 and @xmath149 in fig .
[ fig8](c ) and ( d ) clearly show that the @xmath10 and @xmath8 bands can not be fit by a single lorentian for each band .
figure [ fig9](a ) shows the lorentzian fitting results on the second order raman spectra for @xmath153 .
the dotted line is the calculated raman intensity fitted by six lorentzians .
we fit the @xmath8 bands with two lorentzians labelled by @xmath159 ( blue ) and @xmath160 ( red ) which reffer to @xmath8 bands from outer ( @xmath52 in km direction ) and inner ( @xmath52 in @xmath161 direction ) scattering processes , respectively @xcite .
three lorentzians are needed to fit the @xmath10 band , labelled by @xmath162 ( green ) , @xmath163 ( blue ) , and @xmath164 ( red ) .
finally one lorentzian is used to fit the @xmath20 band .
after lorentzian fitting , we compare both the peak shift and the spectral linewidth as a function of @xmath0 as shown in figs .
[ fig9](b)-(e ) .
we do not show the @xmath159 and @xmath162 for simplicity because there is no experimental data available for comparison .
the calculated results in fig .
[ fig9](b)-(e ) can not fit the experimental value of both the peak position and the linewidth due to the underestimation of the electronic energy dispersion as previously discussed in the @xmath18 dependence of the second - order raman spectra [ see fig .
[ fig7](d ) ] . but
we can discuss the change of both quantities as a function of @xmath0 , where the ka effect takes place . in figs .
[ fig9](b)-(e ) , both the spectral peak position and the linewidth as a function of @xmath0 are plotted in the same range , comparing the theory and experiments .
reasonable agreements between experiments and theory are achieved .
the three major peaks , i.e. , the @xmath160 , @xmath164 , and @xmath20 bands show `` @xmath165 '' ( `` v '' ) shapes of the raman peak shift ( spectral linewidth ) as a function of @xmath0 .
these behaviors exist because of the intraband electron - hole excitation renormalization of phonons as shown in fig .
[ fig5 ] . the @xmath163 band in fig .
[ fig9](d ) is relatively dispersionless in @xmath0 because it is located in the shoulder of the @xmath8 band where 2ito @xmath166 exists .
therefore , for these bands , the competition between interband and intraband electron - hole excitations are expected .
the calculated results overestimate the experimental spectral linewidths of all bands , which is related to the choice of @xmath167 in the @xmath26 integration .
we can tackle this issue by reducing the value of @xmath168 by @xmath169 ; however , the computational burden becomes @xmath170 times larger .
in conclusion , we calculated the first- and second - order raman spectra as a function of @xmath0 .
the opposite effect of the kohn anomaly that is found experimentally between the first- and the second - order raman spectra occurs because the ka effect on the first - order raman spectra is dominated by the renormalization of phonons by the interband electron - hole excitation , while in the second - order raman spectra , the intraband electron - hole excitation dominates over the ka effect .
we also discussed the quantum interference effect observed in the change of the raman intensity as a function of @xmath0 .
both the first- and the second - order raman spectra exhibit an impact of the quantum interference effect , especially when @xmath171 .
present calculated results found that not only is the resonance condition important , but also the explicit consideration of the electron - phonon matrix elements are essential to determine the @xmath0 dependence of the raman spectral lineshape .
e.h . is supported by a mext scholarship .
acknowledges the interdepartmental doctoral degree program for material science leaders at tohoku university for providing a financial support .
r.s . acknowledges mext grants no .
25286005 and no .
m.s.d . acknowledges nsf - dmr grant no .
15 - 07806 .
|
intensity of the first- and the second - order raman spectra are calculated as a function of the fermi energy .
we show that the kohn anomaly effect , i.e. , phonon frequency renormalization , in the first - order raman spectra originates from the phonon renormalization by the interband electron - hole excitation , whereas in the second - order raman spectra , a competition between the interband and intraband electron - hole excitations takes place . by this calculation
, we confirm the presence of different dispersive behaviors of the raman peak frequency as a function of the fermi energy for the first- and the second - order raman spectra , as observed in experiments .
moreover , the calculated results of the raman intensity sensitively depend on the fermi energy for both the first- and the second - order raman spectra .
these results thus also show the importance of quantum interference effect phenomena .
|
success of density functional theory ( dft ) @xcite for the ground - state calculations had prompted search @xcite for similar theories for the excited - states . over the past decade , time - dependent density - functional theory ( tddft ) @xcite has become a standard tool @xcite for obtaining transition energies and the associated oscillator strengths .
however , despite its widespread use , the theory is not without limitations .
for example , calculating excitation energies for double excitation of electrons still remains @xcite a challenge in the tddft approach . at the same time , the charm of getting the excitation energy as the difference between two total energies remains .
this is because one can choose the excited - state at will , promoting as many electrons as one wishes to a set of chosen orbitals , calculate the total corresponding energy and find the excitation energy by subtracting the ground - state energy .
thus research in the direction of performing a kohn - sham like calculation for the excited - states continues . a ground - state like dft approach to obtain the total energy of an excited - state has been developed by grling @xcite and by levy and nagy @xcite .
the theory is based on the constrained - search approach @xcite and proposes that the energy of an excited - state can also be written as a functional @xmath0 = f[\rho,\rho_{0 } ] + \int\rho({\bf r})v_{ext}({\bf r})d{\bf r } \label{1}\ ] ] of the excited - state density @xmath1 . here
@xmath2 $ ] is a bi - density functional that depends on the ground - state density @xmath3 also , and @xmath4 is the external potential that the electrons are moving in .
the bi - density functional for the density @xmath5 of the @xmath6 excited - state is defined via the constrained - search formulation as @xmath7 = { \rm min}_{\psi \to \rho } \left < \psi | { \hat t } + { \hat v}_{ee}|\psi \right > \ ; , \label{2}\ ] ] where @xmath8 is orthogonal to the lower @xmath9 states of the hamiltonian , already determined by the density @xmath3 . such a way of obtaining the functional @xmath2 $ ] makes it non - universal and also state - dependent .
the exchange - correlation energy functional @xmath10 $ ] for the excited - state is then defined as the difference of @xmath2 $ ] and the non - interacting kinetic energy @xmath11 $ ] corresponding to @xmath5 .
the latter is defined in a manner similar to eq . [ 2 ] by dropping the operator @xmath12 from the right hand side .
thus ( for brevity , from here onwards we drop @xmath3 from the argument of the functional ) @xmath13 = f[\rho]-t_{s}[\rho]\;. \label{3}\ ] ] with the assumption that the excited - state density is non - interacting v - representable , the density is obtained by solving the excited - states kohn - sham equation ( atomic units are used throughout the paper ) @xmath14\phi_{i}({\bf
r})= \epsilon_{i}\phi_{i } ( { \bf r } ) \label{4}\ ] ] as @xmath15 where @xmath16 is the occupation number of the orbital @xmath17 . in eq .
[ 4 ] the various terms have their standard meaning with @xmath18 representing the exchange - correlation potential for the excited - state .
it is determined by taking the functional derivative of the excited - state exchange - correlation energy functional . that a kohn - sham like calculation can be performed for the excited - states
was first proposed by harbola and sahni @xcite on physical grounds , and has been put @xcite on a rigorous mathematical footing recently on the basis of differential virial theorem @xcite .
calculations of excited - states energies based on the harbola - sahni work have yielded excellent results @xcite .
the near exact exchange - correlation potential for the singlet @xmath19 excited - state of helium has also been constructed @xcite recently .
however , we are not aware of any work where an exchange - correlation functional for the excited - states has been reported ; in performing excited - state calculations @xcite , either the ground - state functionals or the orbital based - theories @xcite have been employed .
the proposition for the construction of an excited - state exchange - correlation functional is indeed a difficult one since the functional is non - universal and also state - dependent .
thus a general functional form for it may not exist .
against such a background , we ask if it is at all possible to obtain a simple lda - like functional for the excited - states . to keep matters simple
, we have been looking at this problem within the exchange - only approximation . in this paper
we show that it is indeed possible to construct an exchange energy functional that gives transition energies comparable to the exact exchange - only theories such as hartree - fock @xcite , optimized potential @xcite or the harbola - sahni @xcite theory .
the construction of the functional is based on the homogeneous electron - gas and in finding the final form of the functional we are guided mostly by qualitative plausibility arguments .
our work is thus exploratory in nature and represents probably the first attempt to construct an excited - state exchange - energy functional in terms of the density .
the evidence of the accuracy of the functional constructed by us is given by the results of the transition energies of a large number of excited - states .
we also refer the reader to ref .
@xcite for an expression for the change in the exchange energy in terms of the ground - state kohn - sham orbitals when an electron is promoted from a lower energy orbital to a higher one . in the present work
we take a particular class of excited states in which some core orbital are filled , then there are some vacant orbitals and again there are some filled orbitals .
we construct an lda - like functional for such states in the following section .
as stated above , we now consider such excited - states where the occupation of the orbitals is such that the electrons occupy some core orbitals and some shell orbitals , leaving the orbitals between the core and the shell region vacant .
this is shown schematically in fig .
[ f : co - she ] .
such an excited - state would be obtained , for example , if an electron from the filled orbitals of the ground - state is excited to just above the occupied levels .
the exact exchange energy for a set of occupied orbitals is given as @xmath20 so that the excited - state exchange energy when an electron is transferred from one of the orbitals occupied in the ground - state to the lowest unoccupied level is given as @xmath21 where @xmath22 represents the orbital from which the electron has been removed and @xmath23 where the electron is added .
the sum over the index @xmath24 in the second term on the right hand side runs over all the orbitals , including @xmath22 and @xmath23 , up to the highest occupied orbital in the excited - state . on the other hand
the sum in the fifth term runs over all the orbital except @xmath23 .
we now attempt to make an lda - like approximation for the excited - state exchange energy so that the difference ( the last four terms in the equation above ) between the approximate excited- and ground - state exchange energies is close to that given by the exact expression above . in making this approximation accurate ,
it is evident that the self - energy terms ( third and fourth terms on the right hand side of eq .
[ 7 ] ) for the orbitals @xmath22 and @xmath23 are to be treated accurately . as the first step towards an excited - state functional , we make the correspondence between the excited - states that we are considering and similar excitations in a homogeneous electron gas ( heg ) .
if the heg is in it s ground state , the electrons are filled up to the fermi level so that the electrons occupy wave - vectors in k - space from @xmath25 to @xmath26 , where @xmath5 is the electron density . on the other hand ,
in an excited state of the system the electrons will occupy k - space differently compared to the ground state .
for the kind of excited - states that we consider in this paper , the corresponding occupation in the k - space is as follows : the electrons occupy orbitals from @xmath25 to @xmath27 and @xmath28 to @xmath29 with a gap in between as shown in fig .
( [ k - space ] ) .
space occupation in the ground and the excited state configuration . ] so that the excited state density is given by @xmath30 @xmath31 and @xmath32 where @xmath33 in eq .
[ 8 ] @xmath34 and @xmath35 are the core ( corresponding to the electrons occupying @xmath36space from zero to @xmath27 ) and the shell ( corresponding to the electrons occupying @xmath36space from @xmath28 to @xmath29 ) density , respectively , and @xmath5 is the total density . the exchange energy for the heg that occupies the k - space as described above can be obtained exactly and is given as ( mlda stands for modified local density approximation ) @xmath37 where @xmath38 \label{13}\ ] ] is the exchange energy of the core electrons , @xmath39 \label{14}\ ] ] is the exchange energy of the electrons in the shell , and @xmath40 \nonumber \\ & & - \left(k_3 ^ 2 - k_1 ^ 2)^2 \ln \left(\frac{k_3 + k_1}{k_3 - k_1 } \right ) \right ] \label{15}\end{aligned}\ ] ] represents the exchange energy of interaction between the core and the shell electrons . here
@xmath41 is the volume of the heg . after adding the three terms
, the exchange - energy can also be written in the form @xmath42d{\bf r } + log\ ; terms \label{16}\ ] ] where @xmath43 represents the exchange - energy per particle when the heg is in its ground - state with the fermi momentum equal to @xmath44 .
the equation above has a nice interpretation : the integral on the right - hand side represents the exchange energy of the system of electrons with density @xmath5 when per electron energy is written as @xmath45 $ ] , i.e. the per electron energy is given according to the occupation in the k - space ( compare with eq . [ 10 ] ) .
the log terms , on the other hand , have no such simple interpretation .
they have the kinetic energy density in them but we have not been able to write the terms in as easy a form as the first term . that the functional above has all the right limits if we take @xmath46 or @xmath47 is easily verified .
finally , the modified local spin density ( lsd ) functional @xmath48 $ ] in terms of the spin densities @xmath49 and @xmath50 is easily obtained from the functional above as @xmath51 = \frac{1}{2}e_{x}^{mlda } [ 2\rho^{\alpha } ] + \frac{1}{2}e_{x}^{mlda}[2\rho^{\beta } ] \label{lsda}\ ] ] having derived the exchange functional for the heg , we now apply it to the excited - states of various atoms to check if the functional above gives exchange energy differences accurately .
the excited - states chosen are such that they can be represented by a single slater determinant so that the lda is a good approximation @xcite for them .
the different radii in the k - space , @xmath27 , @xmath28 and @xmath29 , needed to evaluate the exchange energy are found by eqs .
[ 8 ] , [ 9 ] and [ 10 ] . for each state
( ground and excited ) , the same set of orbitals @xcite is employed to get the hartree - fock and the lsd exchange energies .
we calculate the lsd and mlsd exchange energies using spherical spin densities since the effect of non - sphericity on the total exchange energy is small @xcite . in table
i we show the difference between the excited - state exchange energy and the ground - state exchange energy for some atoms and ions . in the first column
we give the difference as obtained by the hartree - fock expression for the exchange energy .
in the second coulumn , the numbers are given for both the excited - state and the ground - state exchange energies obtained by employing the ground - state lsd functional .
the third column gives the energy difference when the excited - state exchange energy is calculated using the functional of eq .
it is clearly seen that the ground - state lsd approximation underestimates this energy difference .
this is not surprising since the ground - state functional would give a larger exchange energy for the excited - state than what a proper excited - state functional should give .
however , when the functional of eq .
[ 12 ] is employed to calculate the exchange energy for the excited - states we find , to our surprise , that for the majority of the atoms the functional overestimates the differences by a large amount , whereas we expected to find the error to be about @xmath52 which is the general lda exchange energy error .
we note that this large difference can not come because we have spherical densities .
if non - spherical densities are used , the difference may increase even further .
for example , for the fluorine atom , the ground - state exchange energy will become more negative for non - spherical densities . on the other hand ,
the excited - state exchange energy will remain unchanged since the density is already spherical .
this will result in an even larger difference in the exchange energies of the two states .
we now look for possible sources of error in the exchange - energy differences when the functional of eq .
[ 12 ] is employed to get the exchange energy for the excited - states .
for this we examine eq . [
7 ] in which the last four terms on the right hand side represent the exchange energy difference .
thus @xmath53 it is the lda values to this term that are given in table i. the sources of error in this term we suspect are the self - exchange energies of the orbitals @xmath22 and @xmath23 involved in the electron transfer .
we now argue that the self - energy correction for both the terms sould be in the same direction .
thus for both the orbitals the self - interaction correction ( sic ) is made by subtracting @xcite @xmath54 = \frac{1}{2}\int\int\frac{|\phi({\bf r}_{1})|^{2 } |\phi({\bf r}_{2})|^{2})}{|{\bf r}_{1}-{\bf r}_{2}|}d{\bf r}_{1}d{\bf r}_{2 } - e_{x}^{lsd}[\rho(\phi)]\ ; , \label{18}\ ] ] where @xmath55 is the orbital density for the orbital @xmath56 .
the argument goes as follows .
the lda should be reasonablty accurate when the integral over @xmath44 is continuous .
this makes the first term in the energy difference accurate .
we do have a choice of writing the first and the second terms as @xmath57 but then the first term above will not be accurate as the lda to it would involve integration in k - space with a break from @xmath27 to @xmath28 . therefore to keep the lda accurate , we keep the summation continuous and write the self - interaction energy of the electron removed with a negative sign in front . by including the self - interaction correction for the removed electron only , we find that the error in the exchange energy difference reduces to about @xmath52 of the corresponding hf value . to make the difference even more acuurate
, we now consider the term for the orbital @xmath23 where the electron is added . there
the electron comes in with its self - interaction so for the added orbital too @xmath58 should be subtracted to make the results for the energy difference comparable to the hartree - fock results .
thus the final expression for the exchange - energy that we have is @xmath59-e_{x}^{sic}[\phi_{add } ] \label{20}\ ] ] where @xmath60 is the energy functional given by eq .
[ 12 ] and @xmath61 $ ] is given by eq . [ 18 ] .
we now compute the exchange energy differences given by the functional in eq . [
20 ] and show them in table i. as is evident from the numbers displayed there , the functional of eq .
[ 20 ] gives highly accurate exchange - energy differences for all the systems considered .
when the exchnage - energy difference between the ground- and the excited - state is small , the hf , lsd and the functionals derived above , all give roughly the same results .
however , when this difference is large , the lda underestimates the magnitude of the difference by a large amount whereas the functional of eq .
q:[12 ] overestimates it .
only when the latter is corrected for the self - interaction is the difference almost the same as the hartree - fock difference . having obtained the functional to obtain accurate exchange energy difference , we now apply it to a large number of excited - states of the class considered here and find that we get the transition energies very close to those given by the hartree - fock theory .
we now employ the exchange functional @xmath62 proposed above to obtain the transition energies for a variety of excitations in different atoms . we find that for all the systems the transition energies obtained by us are very close to the corresponding hartree - fock energies @xcite .
our calculations proceed as follows : we get the ground - state energy by solving the kohn - sham equation with the effective exchange potential calculated using the dirac formula @xcite .
we then solve the kohn - sham equation with the same ( corresponding to the ground - state ) functional for the excited - state configuration .
this gives us the excited - state energy @xmath63 .
the difference between @xmath63 and the ground - state energy gives us the transition energy @xmath64 .
we then employ the kohn - sham orbitals from this calculation to get the modified lda exchange energy including sic by employing the functional @xmath62 of eq .
[ 19 ] . making appropriate corrections
we then get the total excited - state energy corresponding to this functional , and its difference with the ground - state energy gives us @xmath65 .
although we have not performed self - consistent calculations with the new functional , self - consistency is not expected to affect the results significantly .
this is because , as we shall see in the results , the major difference in the transition energies given by different functionals arises from the difference in the value of the exchange energy itself .
shown in table ii are the transition energies @xmath66 , @xmath64 and @xmath65 for some light atoms and ions when one of their inner electrons is excited to the lowest available orbital .
the excitation energy in these systems is such that for some of them @xmath64 is close to @xmath66 but for others it is not .
however , @xmath65 is uniformly accurate for all the systems .
we note that the error in @xmath64 is almost fully from the error in the corresponding exchange energy difference .
this is evident from a comparison of the numbers in table i ( for the exchange energy differences ) and in table ii . in table iii
we look at the excitation energies of the alkali atoms and @xmath67 by exciting an electron from the uppermost orbital to an outer orbital .
these are weakly bound systems and as such their excitation energies are relatively smaller .
thus they provide a good testing ground for the proposed functional . an interesting point about these systems
is that the lsd itself gives excitation energies close to the hf excitation energies .
it is therefore quite gratifying to see that the transition energies obtained by the new functional also are of very good quality , although the present method tends to slightly overestimate the transition energies .
next we consider some bigger atoms where we can excite the electron from more than one inner orbital . shown in tables iv and
v are the excitation energies for the atoms in the third row of the periodic table . in table
iv we consider an electron being excited from the @xmath68 orbital to the @xmath69 orbital . in all these case @xmath64 is smaller than the true energy difference whereas the present functional gives highly accurate estimates of the transition energy .
notice again that the error in the value of @xmath64 arises mainly from the error in the exchange energy . in table
v , we show the transition energies for the same set of atoms and ions as in table iv , but for the electron now being excited from the @xmath70 orbital to the @xmath69 orbital . consequently the energy of excitation is much larger in this case .
the lsd in all these cases underestimates the excitation energy , whereas the present functional gives accurate results although slightly overestimating them .
however , the error with respect to the lsd is reduced by a factor of @xmath71 or more . shown in table vi
are the excitation energies for a group of atoms for which the lsd gives transition energies very close to the hf excitation energies . in all the cases we find that the functional proposed here is able to give accurate excitation energies .
thus we find that when the lsd results are accurate , so are the results given by the new functional .
what is significant , however , is that when the lsd results are poor , the new functional properly corrects the error in lsd .
finally , we consider the cases where two electrons are excited to the higher orbitals . as already pointed out , double excitations are difficult to deal with in the tddft approach to finding excitation energies . results for different excitations for a variety of atomic systems are shown in table vii .
as is evident from the table , for all the systems , our method gives excellent results whereas the lsd underestimates the energies .
in the above we have presented a new lda - like functional for obtaining the excitation energies .
it has been employed to investigate over @xmath72 excited states .
the results show that our procedure gives accurate excitation energies for all of them , whereas for most of the systems the lsd underestimates the energy difference .
we have worked within the exchange - only approximation and have chosen a particular class of excited - states .
what we have learnt through the study reported here is that a simple extension of the lda to the excited - states overestimates the energy differences .
this is due to the self - interaction of the electron inherent in the lda . when corrected for the self - interaction through a careful analysis , the resulting functional gives highly accurate answers for the excited - states
. thus if more accurate functionals than the lda are employed , our method of developing excited - state functionals should give good excited - state functionals .
in the present work , we have developed a functional for a particular class of excited states and demonstrated that it is possible to construct excited - state energy functionals that are capable of giving transition energies close to the exact theory .
we are now working on functionals for states other than considered in this paper .
as pointed out in the introduction , excited - state functionals are not universal and therefore have to be dealt with separately for different kinds of excited - states . in this work
, we have also not looked at the correlation energy functionals .
can correlation energy functionals be developed along similar lines ?
we trust that it should be possible and are working on this problem . finally , we also wish to look at the ultimate excitation i.e. the ionization of the system . if the electron is removed from the uppermost orbital , the lda is known to give good ionization energies when calculated through the total energy difference . in these cases
too our functional would give results the same as those obtained from the lda : if we go through the arguments presented earlier , we find that in these cases our functional reduces to the lda functional for the core electrons .
this is because in calculating the exchnage energy , the summation over the occupied orbitals is continuous and therefore we do not have to make any self - interaction correction for the removed electron .
further the shell electron density vanishes so the contribution from the added electron is zero .
interestingly , we find that if we ionize the atom by removing an electron from the inner orbitals , we obtain accurate removal energies if we think of the process in two spteps - first removing an electron from the top level and then exciting the resulting ion by exciting an electron from the inner orbital to the top of filled orbitals .
more work on such different kinds of excitation is in progress and will be reported in the future .
dreizler and e.k.u .
gross , _ density functional theory _ , ( springer - verlag , berlin , 1990 ) r.g .
parr and w. yang , _ density - functional theory of atoms and molecules _ , ( oxford university press , oxford , 1999 ) .
t. ziegler , a.rauk and e.j .
baerends , theor .
acta * 43 * , 261 ( 1977 ) .
o. gunnarsson and b.i .
lundquist , phys .
b * 13 * , 4274 ( 1976 ) .
u. von barth , phys .
a * 20 * , 1693 ( 1979 ) .
perdew and m. levy , phys .
b * 31 * , 6264 ( 1985 ) . r.k .
pathak , phys .
a * 29 * , 978 ( 1984 ) .
theophilou , j. phys .
c * 12 * , 5419 ( 1979 ) .
gross , l.n .
oliviera and w. kohn , phys .
a * 37 * , 2809 ( 1988 ) ; l.n .
oliviera , e.k.u .
gross and w. kohn , phys .
a * 37 * , 2821 ( 1988 )
nagy , j. phys .
b * 29 * , 389 ( 1996 ) .
e. runge and e.k.u .
gross , phys .
lett . * 52 * , 997 ( 1984 ) .
casida , _ recent advances in density functional methods , part 1 _ , edited by d.p .
chong ( world scientific , singapore,1995 ) .
m. petersilka , u.j .
gossmann and e.k.u gross , phys .
76 * , 1212(1996 ) .
maitra , f. zhang , r.j . cave and k.
burke , j. chem .
* 120 * , 5932 ( 2004 ) .
a. grling , phys .
a * 59 * , 3359 ( 1999 ) .
m. levy and .
nagy , phys .
lett . * 83 * , 4361 ( 1999 ) ; .
nagy and m. levy , phys .
a * 63 * , 052502 ( 2001 ) .
m. levy , proc .
usa * 76 * , 6062 ( 1979 ) .
harbola and v. sahni , phys .
lett . * 62 * , 489 ( 1989 ) .
v. sahni , l. massa , r. singh and m. slamet , phys .
rev . lett .
* 87 * , 113002 ( 2001 ) .
a. holas and n.h .
march , phys .
a * 55 * , 2040 ( 1995 ) .
k.d . sen , chem .
phys . lett . *
188 * , 510 ( 1992 ) .
r. singh and b.m .
deb , phys . rep . * 311 * , 47 ( 1999 ) . m.k .
harbola , phys . rev .
a * 69 * , 042512 ( 2004 ) .
harbola , phys .
a * 65 * , 052504 ( 2002 ) .
sharp and g.k .
horton , phys .
rev . * 90 * , 3876 ( 1953 ) ; j.d .
talman and w.f .
shadwick , phys .
a * 14 * , 36 ( 1976 ) .
fischer , _ the hartree - fock method for atoms _ , ( john wiley , new york 1977 ) .
v. sahni , _ quantal density functional theory _
( springer berlin , 2004 ) .
a. grling , phys .
a * 54 * , 3912 ( 1996 ) .
the orbitals employed are those obtained by solving the kohn - sham equations with the harbola - sahni exchange potential ( ref .
@xcite above ) .
these orbitals are very close ( see , for example , v. sahni , y. li and m.k .
harbola , phys .
a * 45 * , 1434 ( 1992 ) and refs .
@xcite above ) to the hartree - fock orbitals .
janak and a.r .
williams , phys .
b * 23 * , 6301 ( 1981 ) .
perdew and a. zunger , phys .
b * 23 * , 5048 ( 1981 ) .
the energies are calculated by solving the kohn - sham equation with the harbola - sahni potential .
the resulting multiplet energies are essentially the same ( see ref .
@xcite ) as those of hartree - fock theory . p.a.m .
dirac , proc .
cambridge phil .
* 26 * , 376 ( 1930 ) .
.difference in the exchange energies of the ground- and excited - states of some atoms and ions .
the first column gives the atom / ion and the transition , the second column the difference @xmath73 as obtained in hartree - fock theory , the third column the difference @xmath74 given by the ground - state energy functional .
the fourth and the fifth column describe the difference as obtained with the functional proposed in this paper .
the fourth column gives the exchange - energy difference @xmath75 obtained by employing the functional of eq .
[ 12 ] whereas the fifth column gives that given by the functional of eq .
[ 19 ] @xmath76 .
numbers given are in atomic units . [
cols="<,^,^,^,^",options="header " , ]
|
an exchange energy functional is proposed and tested for obtaining a class of excited - state energies using density - functional formalism .
the functional is the excited - state counterpart of the local - density approximation functional for the ground - state .
it takes care of the state - dependence of the energy functional and leads to highly accurate excitation energies .
|
all graphs in this paper are simple of finite orders , i.e. , graphs are undirected with no loops or parallel edges and with finite number of vertices .
the _ complement _
@xmath14 of a graph @xmath0 , is a graph with the same vertex set as @xmath0 and with the property that two vertices are adjacent in @xmath14 if and only if they are not adjacent in @xmath0 . for any vertex @xmath15 ,
the _ open neighborhood _ of @xmath16 is the set @xmath17 and the _ closed neighborhood _ of @xmath16 is the set @xmath18=n(v)\cup \{v\}$ ] . for a set @xmath19 ,
the open neighborhood of @xmath20 is @xmath21 and the closed neighborhood of @xmath20 is @xmath22=n(s)\cup s$ ] .
a set @xmath19 is a _ dominating set _ if @xmath22=v$ ] , or equivalently , every vertex in @xmath23 is adjacent to at least one vertex in @xmath20 .
the _ domination number _ @xmath24 , is the minimum cardinality of a dominating set in @xmath0 . for a detailed treatment of domination theory ,
the reader is referred to @xcite .
let @xmath25 be the family of dominating sets of a graph @xmath0 with cardinality @xmath4 and let @xmath26 .
the _ domination polynomial _
@xmath27 of @xmath0 is defined as @xmath28 ( see @xcite ) .
this polynomial is the generating polynomial for the number of dominating sets of each cardinality .
calculating the domination polynomial of a graph @xmath0 is difficult in general , as the smallest power of a non - zero term is the domination number @xmath29 of the graph , and determining whether @xmath30 is known to be np - complete @xcite .
but for certain classes of graphs , we can find a closed form expression for the domination polynomial .
two graphs @xmath0 and @xmath31 are said to be _ dominating equivalent _ , or simply @xmath32-equivalent , written @xmath33 , if @xmath34 .
it is evident that the relation @xmath35 of being @xmath32-equivalence is an equivalence relation on the family @xmath36 of graphs , and thus @xmath36 is partitioned into equivalence classes , called the _
@xmath32-equivalence classes_. given @xmath37 , let @xmath38=\{h\in { \cal g}:h\sim g\}.\ ] ] we call @xmath39 $ ] the equivalence class determined by @xmath0 .
a graph @xmath0 is said to be dominating unique , or simply @xmath9-unique , if @xmath39 = \{g\}$ ] @xcite .
determining @xmath9-equivalence class of graphs is one of the interesting problems on equivalence classes .
a question of recent interest concerning this equivalence relation @xmath40 $ ] asks which graphs are determined by their domination polynomial .
it is known that cycles @xcite and cubic graphs of order @xmath41 @xcite ( particularly , the petersen graph ) are , while if @xmath42 , the paths of order @xmath1 are not @xcite . in @xcite
, authors completely described the complete @xmath43-partite graphs which are @xmath9-unique .
their results in the bipartite case , settles in the affirmative a conjecture in @xcite .
let @xmath1 be any positive integer and @xmath5 be @xmath1-barbell graph with @xmath6 vertices which is formed by joining two copies of a complete graph @xmath7 by a single edge . in this paper , we consider @xmath1-barbell graphs and study their domination polynomials .
we prove that for every @xmath8 , @xmath5 is not @xmath9-unique .
more precisely , in section 2 , we show that for every @xmath1 , @xmath10 $ ] contains many graphs , which one of them is @xmath44 and another one is the complement of book graph of order @xmath11 , @xmath12 . in section 3 ,
we present many graphs in @xmath45 $ ] .
in this section , we study the @xmath9-equivalence classes of some graphs .
first we consider the domination polynomial of the complement of book graph .
the @xmath1-book graph @xmath46 can be constructed by bonding @xmath1 copies of the cycle graph @xmath47 along a common edge @xmath48 , see figure [ figure6 ] .
the book graph @xmath46.,height=94 ] the following theorem gives a formula for the domination polynomial of @xmath46 .
@xcite for every @xmath49 , @xmath50 domination polynomials , exploring the nature and location of roots of domination polynomials of book graphs has studied in @xcite . here , we consider the domination polynomial of the complement of the book graphs
. we shall prove that the @xmath1-barbell graph @xmath5 and @xmath12 have the same domination polynomial .
the turn graph @xmath51 is a complete multipartite graph formed by partitioning a set of @xmath1 vertices into @xmath43 subsets , with sizes as equal as possible , and connecting two vertices by an edge whenever they belong to different subsets .
the graph will have @xmath52 subsets of size @xmath53 , and @xmath54 subsets of size @xmath55 .
that is , a complete @xmath43-partite graph @xmath56 the turn graph @xmath57 can be formed by removing a perfect matching , @xmath1 edges no two of which are adjacent , from a complete graph @xmath58 . as roberts ( 1969 ) showed ,
this graph has boxicity exactly @xmath1 ; it is sometimes known as the robert s graph @xcite .
if @xmath1 couples go to a party , and each person shakes hands with every person except his or her partner , then this graph describes the set of handshakes that take place ; for this reason it is also called the cocktail party graph .
so , the cocktail party graph @xmath59 of order @xmath60 is the graph with vertices @xmath61 in which each pair of distinct vertices form an edge with the exception of the pairs @xmath62 .
the following result is easy to obtain .
[ lem ] for every @xmath63 , @xmath64 .
figure [ cbook ] shows the complement of the book graph @xmath65 .
.,width=336 ] the vertex contraction @xmath66 of a graph @xmath0 by a vertex @xmath67 is the operation under which all vertices in @xmath68 are joined to each other and then @xmath67 is deleted ( see@xcite ) .
the following theorem is useful for finding the recurrence relations for the domination polynomials of arbitrary graphs .
[ theorem1]@xcite let @xmath0 be a graph .
for any vertex @xmath67 in @xmath0 we have @xmath69 , x ) - ( 1 + x)p_u(g , x),\ ] ] where @xmath70 is the polynomial counting the dominating sets of @xmath71 which do not contain any vertex of @xmath68 in @xmath0 .
the following theorem gives a formula for the domination polynomial of the complement of the book graph .
[ theorem6 ] for every @xmath63 , @xmath72 * proof .
* consider graph @xmath73 and vertex @xmath16 in the figure [ cbook ] . by theorem [ theorem1 ]
, we have : @xmath74 , x ) - ( 1 + x)p_v(b_n^c , x)\\ & = & ( x + 1 ) d(b_n^c - v , x ) + xd(k_{n+1},x ) - ( 1 + x)(d(k_{n+1},x)-(n+1)x - nx^2)\\ & = & ( x + 1 ) d(b_n^c - v , x ) - d(k_{n+1},x)+x(1+x)(1+n(1+x))\\ & = & ( x + 1 ) d(b_n^c - v , x ) - ( ( 1+x)^{n+1}-1)+x(1+x)(1+n(1+x)),\end{aligned}\ ] ] where @xmath75 .
now , we use theorem [ theorem1 ] to obtain the domination polynomial of the graph @xmath76 .
we have @xmath77 , x ) - ( 1 + x)p_u(b_n^c - v , x).\end{aligned}\ ] ] since @xmath78 and using lemma [ lem ] , we have @xmath79 consequently , @xmath80 @xmath81 the @xmath1-barbell graph is the graph on @xmath6 vertices which is formed by joining two copies of a complete graph @xmath7 by a single edge , known as a bridge , shown in figure [ figure1 ] .
we denote this graph by @xmath5 . for this graph
, we shall calculate this domination polynomial .
we need the following definition and theorems .
the barbell graph of order 16 , @xmath82.,width=207,height=94 ] an irrelevant edge is an edge @xmath83 , such that @xmath84 , and a vertex @xmath85 is domination - covered , if every dominating set of @xmath86 includes at least one vertex adjacent to @xmath16 in @xmath0 @xcite .
we need the following theorems to obtain the domination polynomial of barbell graph @xmath5 .
@xcite[dcov ] let @xmath87 be a graph .
an vertex @xmath88 of @xmath0 is domination - covered if and only if there is a vertex @xmath89 $ ] such that @xmath90 \subseteq n[v]$ ] .
@xcite[irre ] let @xmath87 be a graph .
an edge @xmath91 is an irrelevant edge in @xmath0 , if and only if @xmath67 and @xmath16 are domination - covered in @xmath92 .
[ theorem5 ] for every @xmath8 and @xmath63 , @xmath93 * proof .
* let @xmath94 be an edge joining two @xmath7 in barbell graph . by theorem [ dcov ] two end vertices of edge @xmath94
are domination - covered in @xmath95 .
so , by theorem [ irre ] the edge @xmath94 is an irrelevant edge of @xmath5 . therefore @xmath96
@xmath81 the following corollary is an immediate consequence of theorems [ theorem6 ] and [ theorem5 ] .
[ korolari1 ] for each natural number @xmath1 , @xmath5 and @xmath12 have the same domination polynomial .
more precisely , for every @xmath1 , @xmath10\supseteq \{bar_n , b_{n-1}^c , k_n\cup k_n\}$ ] , and @xmath97\supseteq \{bar_n , b_{n-1}^c , k_n\cup k_n\}$ ] . here , we present some other families of graphs whose are in the @xmath10 $ ] .
let to define the generalized barbell graphs .
as we know , the @xmath5 is formed by joining two copies of a complete graph @xmath7 by a single edge .
we like to join two copies with more edges as follows : suppose that @xmath98 and @xmath99 are the vertices of two copies of complete graph of order @xmath1 , @xmath7 and @xmath100 .
the generalized barbell graph is denoted by @xmath101 and is a graph with @xmath102 and @xmath103 where @xmath104 . as examples see two non - isomorphic graphs @xmath105 in figure [ gen ] .
notice that @xmath12 is one of the specific case of @xmath106 .
the left graph in figure [ gen ] , is @xmath107 .
two generalized barbell graphs @xmath105 .
, width=302,height=90 ] we have the following theorem .
[ theorem7 ] for every @xmath108 and @xmath63 , @xmath109 * proof . *
we prove this theorem by induction on @xmath110 .
suppose that @xmath111 , then by theorem [ theorem5 ] , the result holds .
assume that the result holds for @xmath112 .
let @xmath113 and @xmath94 be the additional edge of @xmath101 to the @xmath114 . by theorem [ dcov ] two end vertices of edge @xmath94
are domination - covered in @xmath95 .
so , by theorem [ irre ] the edge @xmath94 is an irrelevant edge of @xmath114 .
therefore by the induction hypothesis we have the result .
@xmath81 the following corollary is an immediate consequence of theorems [ theorem5 ] and [ theorem7 ] . for each natural number @xmath1 and @xmath115 , @xmath5 and @xmath101 have the same domination polynomial .
the following example shows that , except for the generalized barbell graphs , there are other graphs in @xmath9-equivalence classes of @xmath5 .
[ last ] all connected graphs in @xmath116 $ ] are the graphs @xmath117 and two graphs in figure [ othe ] .
two graphs in @xmath116 $ ] .
, width=226,height=90 ]
we observed that , for each natural number @xmath1 and @xmath115 , the domination polynomials of @xmath5 and @xmath101 is @xmath118 . in this section ,
we present graphs whose domination polynomials are @xmath119 . for this purpose , we construct families of graphs from a path @xmath120 which we denote them by @xmath121 in the following definition .
[ def1 ] the graph @xmath121 is a graph which obtain from a path @xmath120 with the vertices @xmath122 , by substituting a graph @xmath123 of order @xmath124 , for every vertex @xmath125 of @xmath120 , such that * for @xmath126 , the graphs @xmath123 have at least one vertex of degree @xmath127 and other @xmath123 s have at least two vertices of degree @xmath127 , and * in the graph @xmath121 , the end vertices of each edge @xmath128 in the path graph , @xmath120 are one vertex of degree @xmath127 in graphs @xmath129 and @xmath123 . *
* let @xmath128 @xmath134 be the edge joining @xmath129 and @xmath123 in @xmath121 . by theorem [ dcov ] two end vertices of edge @xmath128
are domination - covered in @xmath135 .
so , by theorem [ irre ] every edge @xmath128 is an irrelevant edge of @xmath121 .
therefore we have the result .
@xmath81 we shall generalize the graphs @xmath121 in definition [ def1 ] such that this generalized graphs and @xmath121 have the same domination polynomial .
suppose that @xmath136 be a family of graphs in the form of @xmath137 such that the complete graphs @xmath138 with @xmath139 and @xmath140 with @xmath141 are joined with @xmath110 following edges @xmath142 for each natural number @xmath110 , all graphs in the family of @xmath136 have the same domination polynomial .
more precisely , the domination polynomial of each @xmath31 in @xmath136 is equal to @xmath119 . * conclusion . * in this paper , we studied the @xmath9-equivalence classes of barbell graphs @xmath5 .
we showed that , for each natural number @xmath1 , @xmath44 , @xmath5 , @xmath101 and the complement of the book graph of order @xmath11 , @xmath12 have the same domination polynomial , i.e. , @xmath10=[bar_{n , t}]=[b_{n-1}^c]=[k_n\cup k_n]$ ] .
example [ last ] , implies that except for these kind of graphs , there are another graphs in this class .
therefore , exact characterization of graphs in @xmath10 $ ] remains as an open problem .
also we presented many families of graphs whose are in @xmath143 $ ] , but similar to example [ last ] , there are another graphs in this class .
so , exact characterization of graphs in @xmath143 $ ] remains as another open problem .
|
let @xmath0 be a simple graph of order @xmath1 .
the domination polynomial of @xmath0 is the polynomial @xmath2 , where @xmath3 is the number of dominating sets of @xmath0 of size @xmath4 .
the @xmath1-barbell graph @xmath5 with @xmath6 vertices , is formed by joining two copies of a complete graph @xmath7 by a single edge .
we prove that for every @xmath8 , @xmath5 is not @xmath9-unique , that is , there is another non - isomorphic graph with the same domination polynomial .
more precisely , we show that for every @xmath1 , the @xmath9-equivalence class of barbell graph , @xmath10 $ ] , contains many graphs , which one of them is the complement of book graph of order @xmath11 , @xmath12 .
also we present many families of graphs in @xmath9-equivalence class of @xmath13 .
department of mathematics , yazd university , 89195 - 741 , yazd , iran + [email protected] * keywords : * domination polynomial ; @xmath9-unique ; equivalence ; generalize barbell graphs .
* ams subj . class .
: * 05c60 , 05c69
|
the threat of global spreading of epidemics , like the pandemic flu from 2009 , or spreading of computer viruses which are endangering the functioning of internet dependent facilities , are responsible for the enormous increase in public interest on immunization during the last years .
much progress has been achieved in understanding epidemic spreading @xcite , and various models have been developed suggesting possible ways of efficient immunization @xcite .
however , the search for even more effective immunization strategies must be pursued since any improvement of immunization efficiency can save human lives and resources . in this paper
we introduce a novel immunization strategy based on inverse targeting , which proves to be effective and numerically more efficient than any proposed previous one .
epidemics can spread in human population through networks of social contacts , and viruses can propagate on computer networks .
we will implement an immunization process on networks by immunizing the nodes which represent either people that are to be vaccinated , or computers that should be equipped with specially developed antivirus programs . we suppose that the dynamics of both , the epidemic spreading and the immunization process are much faster than the growth and change of the network itself .
the question we want to answer is , given the topology of the network through which the epidemics can spread , what is the best possible way to immunize its nodes .
the immunization process should require the smallest possible number of immunization doses .
at the same time , the size of the largest connected cluster of non - immunized nodes should stay as small as possible throughout the immunization process .
+ + an immunization strategy determines the sequence in which nodes of a network should be immunized to prevent most efficiently the spreading of epidemics .
the simplest immunization strategy is the one in which the nodes are chosen at random @xcite .
this strategy has however proven to be inefficient , requiring a large fraction of nodes ( in many cases , such as internet , nearly 100% ) to be immunized in order to stop epidemics spreading @xcite .
targeted immunization strategies turn out to be much more efficient @xcite .
here one immunizes first those individuals that are most important for the disease spreading . in terms of network properties
this importance can be expressed through the number of links of a node , or through its betweenness centrality @xcite . although it has been shown that these two properties are correlated @xcite , their role in the immunization process is typically quite different .
most studied targeted strategies are based on these two node traits . in the high degree based ( hd)@xcite , and the high betweenness based ( hb ) @xcite targeting the sequence in which nodes will be immunized is based on their degree or betweenness in the initial network .
they can be improved by adaptive strategies @xcite , high degree adaptive ( hda ) , and high betweenness adaptive ( hba ) strategy . here
the degree or betweenness of nodes is recalculated for the remaining network of non - immunized nodes each time a node has been immunized .
the node with highest value of this recalculated degree ( for hda ) or betweenness ( for hba ) is immunized next .
these dynamic strategies have proven to be very effective @xcite .
other strategies , like immunizing neighbors of a randomly chosen node according to some rule @xcite have the advantage of requiring only local information , but are also less effective . a further approach , recently introduced in ref .
@xcite , is the improvement of existing strategies through optimization procedures .
this approach leads to very effective node and link immunization strategies .
another strategy based on equal graph partitioning @xcite is especially effective when the number of available immunization doses is fixed . for any given number of nodes
this strategy finds an effective way to partition the network in clusters such that the size of the largest cluster is the smallest possible . in ref .
@xcite the authors show that for a given number of doses this strategy is more effective than targeting hd , hb or hda strategy .
however , if the supply of immunization doses is not known in advance or increases with time , one must take into account that when new doses become available the network of non - immunized nodes can be quite different from the initial network .
the nodes which can now be immunized are then different from those that would give an ideal graph partitioning if we had this larger quantity of immunization doses from the beginning of the process . in such cases dynamic
hda remains the best known immunization strategy because the hba , which needs more global information , turn out to be numerically extremely demanding . as a consequence , although the hba may be theoretically the more efficient strategy @xcite , the complexity of its algorithm makes it uninteresting for practical purposes and limits its use to very small systems . in this paper
we will introduce a strategy which is more efficient than hda and in many cases even better than hba , being at the same time numerically much faster than hba allowing to apply it to large , real systems .
in any successful immunization strategy , nodes which are not relevant for keeping the largest cluster of non - immunized nodes together , should only be immunized during later stages . in our method
we start by recognizing such nodes first .
this way we reveal the sequence of nodes that are going to be immunized in reversed order . to illustrate our method ,
let the nodes and the links of the network be black .
every node has an index ( i d ) ( fig.[fig : topology ] ) . at each step
we check for each black node how would turning it and its links green increase the size of the largest connected green cluster and then choose the one with the smallest contribution to its size .
when more nodes give the same contribution , the one with the smallest number of links connecting it to black nodes is chosen .
if there are several nodes which fulfill both criteria , one of them is chosen at random . after a node is chosen we put its i d in a list and continue the process until all nodes and links are colored green .
the list of nodes ids describing the reconstruction process gives us the inverse sequence of nodes that are to be immunized .
the complexity of the inverse targeting algorithm is at most @xmath2 , or @xmath3 for sparse networks , where @xmath4 is the number of links and @xmath5 the number of nodes .
this is much faster then the hba strategy .
the best known algorithm for determining betweenness of nodes goes as @xmath2 @xcite and since it has to be applied at each step of the immunization process , the complexity of the algorithm for the hba strategy is @xmath6 , or @xmath7 for sparse networks .
we compare the effectiveness of our inverse targeting immunization strategy to hda and hba strategies .
we start by comparing the fractions @xmath8 of immunized nodes for which the network of non - immunized individuals breaks apart , which is the measure usually used in studies of epidemics . in addition , we use network s robustness to immunization as the measure of how effective the immunization procedure is .
a good immunization strategy should make such a robustness as small as possible .
the robustness sums up the sizes of the largest connected clusters @xmath9 of the networks of non immunized nodes remaining after immunization of q nodes : @xmath10 where @xmath5 is the size of the network .
this measure captures the network response to immunization throughout the immunization process , and not only at the percolation point at which the network of non immunized nodes becomes disconnected @xcite .
the process of immunization usually takes some time .
the initially slow production of needed vaccine soon after the disease starts spreading , slows down the immunization process in these initial phases .
therefore it is important that the immunization process not only completely breaks up the network of contacts among susceptible , non - immunized individuals , but also that the size of the largest connected cluster , that is , the number of individuals being at highest risk of getting infected , is as small as possible during the whole immunization process .
this property is expressed by a small robustness @xmath11 .
we have applied our strategy together with hda and hba strategies to three different model networks , so called random regular graphs @xcite ( in which nodes are connected randomly keeping the number of links per node constant ) , erds - rnyi networks @xcite and scale free networks created with the configuration model @xcite . in figure [ fig : robustness ] we show how the fraction of nodes in the largest connected non - immunized cluster @xmath12 changes with the fraction @xmath13 of immunized nodes during different immunization processes on these three model networks .
@xmath8 is the value for which the largest connected cluster is of order @xmath14 , and the robustness @xmath11 is the area under the curves in figure [ fig : robustness ] .
we can see that for all three types of model networks our immunization strategy has a smaller @xmath8 than hda and has approximately the same value as for the hba immunization . at the same time
the robustness of networks in the case of our new strategy is lower than in other strategies making it the most efficient of the three strategies .
we have also applied our strategy to two real networks , the internet at the level of autonomous system @xcite , as example for a network through which computer viruses can spread , and the collaboration network of high energy physicist ( hep ) as an example of social networks @xcite . here
we could compare our strategy only to the hda immunization strategy since hba would take too much computer time .
the results are presented in figure [ fig : robustnessreal ] , where we can see that in both cases the new immunization strategy is according to both criteria , @xmath8 and @xmath11 , much more efficient than the hda strategy .
summing up the results presented in figures [ fig : robustness ] and [ fig : robustnessreal ] , we have found that our immunization strategy outperforms the other two .
comparing to the hda ( hba ) strategy , in the case of the scale free network , @xmath8 is @xmath15 and @xmath11 @xmath16 smaller , for the erds - rnyi network , @xmath8 is @xmath17 and @xmath11 @xmath18 smaller , and for the random regular network @xmath8 is @xmath19 and @xmath11 @xmath20 smaller . for real networks , for hep network , @xmath8 is @xmath21 and @xmath11 @xmath22 smaller and for the as internet , @xmath8 is @xmath23 and @xmath11 @xmath1 smaller . for model networks
we have studied the effect of the system size on the efficiency of our immunization strategy and have compared it with the other two strategies . in figure
[ fig : dda ] we can see that our immunization strategy is for all network sizes more effective than hda ( gives the smallest @xmath11 value ) . for erds - rnyi and scale free networks it is always more efficient than the hba strategy and only for small regular networks with less than 1000 nodes hba is more efficient . the way the effectiveness of immunization strategies changes with the density of links in the model networks is shown in figure [ fig : threshold ] .
we can see that the new immunization strategy is the most efficient for all average degrees @xmath24 studied .
in all cases studied here ( for different @xmath24 and different @xmath5 ) the difference in the effectiveness of different strategies is smallest for scale free networks .
we have introduced a novel immunization strategy based on inverse targeting . to estimate the efficiency of our strategy we have not only used the percolation transition @xmath8 but also an additional measure taking into account the history of the immunization process .
we find that the new strategy is much more efficient than the high degree based adaptive targeted strategy .
it is also at least as efficient as the high betweenness based targeting strategy which was for a long time claimed to be the most effective immunization strategy @xcite . due to its numerical inefficiency
, the hba strategy can not be used for larger systems and is therefore not applicable to real world problems .
our strategy is in contrast also numerically efficient and can therefore become the strategy of choice for real applications .
the algorithm we introduced could find other applications besides immunization .
it could for example be used to determine attack strategies for the efficient destruction of criminal networks .
we acknowledge financial support from the eth competence center coping with crises in complex socio - economic systems ( ccss ) through eth research grant ch1 - 01 - 08 - 2 and the swiss national science foundation under contract 200021 126853 , the swiss national science foundation under contract 200021 126853 and funcap .
lloyd and r.m .
may , _ science _ * 292 * , 1316 - 1317 ( 2001 ) .
newman , _ phys .
e _ * 66 * , 016128 ( 2002 ) r. cohen , s. havlin and d. ben - avraham _ phys .
lett . _ * 91 * , 247901 ( 2003 ) .
p. holme , _ europhys .
lett . _ * 68 * , 908 - 914 ( 2004 ) .
schneider , t. mihaljev , s. havlin and h.j .
herrmann , _ phys .
, in press ( 2011 ) y. chen _ et al _ , _ phys .
* 101 * , 058701 ( 2008 ) .
et al _ , _ phys .
e _ * 65 * , 056109 ( 2002 ) .
et al _ , _ phys .
lett _ * 85 * , 4626 ( 2000 ) .
r. albert , h. jeong and a.l .
barabasi , _ nature _ * 406 * , 378 ( 2000 ) .
et al _ , _ phys . rev .
lett _ * 85 * , 5468 ( 2000 ) .
r. pastor - satorras and a. vespignani , _ phys .
lett . _ * 86 * , 3200 ( 2001 ) .
_ , _ phys .
lett . _ * 86 * , 3682 ( 2001 ) .
r. pastor - satorras and a. vespignani , _ phys .
e _ * 65 * , 036104 ( 2002 ) .
freeman , _ sociometry _ * 40 * , 35 ( 1977 ) m.e.j .
newman and m.girvan , _ phys .
e _ * 69 * , 026113 ( 2004 ) .
newman , _ phys .
e _ * 64 * , 016131 ( 2001 ) .
u. brandes , _ j. math .
sociol . _ * 25 * , 163 ( 2001 ) . c.m .
et al _ , _ pnas _ in press ( 2011 ) h.j .
et al _ , _ j. stat .
_ p01027 ( 2011 ) a.bekessy , p.bekessy , and j.komlos , _ stud .
math . hung . _ * 7 * , 343 ( 1972 ) p. erds and a. rnyi , _ publ .
inst . hung .
sci . _ * 5 * , 17 - 60 ( 1960 ) .
m. molloy and b. reed , _ random struct .
algorithms _ * 6 * , 161 ( 1995 ) .
|
we propose a new method to immunize populations or computer networks against epidemics which is more efficient than any method considered before .
the novelty of our method resides in the way of determining the immunization targets .
first we identify those individuals or computers that contribute the least to the disease spreading measured through their contribution to the size of the largest connected cluster in the social or a computer network .
the immunization process follows the list of identified individuals or computers in inverse order , immunizing first those which are most relevant for the epidemic spreading .
we have applied our immunization strategy to several model networks and two real networks , the internet and the collaboration network of high energy physicists .
we find that our new immunization strategy is in the case of model networks up to @xmath0 , and for real networks up to @xmath1 more efficient than immunizing dynamically the most connected nodes in a network .
our strategy is also numerically efficient and can therefore be applied to large systems .
|
observations of accreting neutron stars in low - mass x - ray binaries ( lmxbs ) with the _ rossi x - ray timing explorer _ have revealed that the persistent x - ray emission of at least fifteen show remarkably coherent quasi - periodic brightness oscillations ( qpos ) , with frequencies @xmath5 ranging from @xmath0300 hz to @xmath01200 hz .
these kilohertz qpos are the highest - frequency oscillations ever seen in any astrophysical object .
eight of the fourteen identified sources in which kilohertz qpos have been detected are `` atoll '' sources ( , , , , , , , and ) .
kilohertz qpos have also been detected in all six of the originally identified `` z '' sources ( 1 , , , , , and ) .
( for the definitions of atoll and z sources , see hasinger & van der klis 1989 . ) highly coherent brightness oscillations with frequencies ranging from @xmath0360 hz to @xmath0580 hz have been detected during type i ( thermonuclear ) x - ray bursts from four kilohertz qpo sources .
another neutron star in the direction of the galactic center shows burst oscillations with a frequency of 589 hz , but this source has not yet been positively identified .
the frequency ranges and rms amplitudes of the currently known kilohertz qpos and burst oscillations are listed in table 1 , with references .
two simultaneous kilohertz qpos have so far been seen in six of the eight atoll sources in which kilohertz qpos have been detected ( all except and ) and in all six of the z sources .
the differences @xmath6 between the frequencies of the two qpos seen in these sources all fall in the range @xmath7350 hz . in the atoll sources ( ford et al .
1996 , 1997a ) , ( wijnands et al .
1997b ) , and ( strohmayer et al .
1996a , 1996b , 1996c ) and in the z sources ( van der klis et al .
1996e ) , ( wijnands et al .
1997c ) , and ( wijnands et al .
1998 ) @xmath6 is constant in time , within the errors . in the z source 1 ,
@xmath6 changed from @xmath0250 hz when the accretion rate was probably near and sometimes even slightly greater than the eddington critical rate to @xmath0300 hz when the accretion rate was significantly lower ( van der klis et al.1997b ) . in , a brightness oscillation with a frequency consistent with the difference
@xmath8 hz between the frequencies of the two kilohertz qpos seen simultaneously in this source was detected with marginal significance during one 30-min interval ( ford et al .
1996 , 1997a ) . in , a strong
, relatively coherent brightness oscillation was detected in six of the twelve x - ray bursts so far observed from this source , with a frequency that has remained constant , within the errors , for more than a year and is consistent with the difference @xmath9 hz between the frequencies of the two kilohertz qpos seen simultaneously in this source ( strohmayer et al .
1996b , 1996c ; strohmayer , zhang , & swank 1997 ; strohmayer 1997 , personal communication ) .
burst oscillations have also been seen in , with a frequency @xmath10580 hz ( zhang et al .
1996 , 1997 ) , and in , with a frequency @xmath10520 hz ( morgan & smith 1996 ; smith , morgan , & bradt 1997 ) . in both cases the frequency of the burst oscillation is approximately _
twice _ the difference between the frequencies of the two kilohertz qpos observed simultaneously in these sources .
the peaks in power density spectra associated with these kilohertz qpos are relatively narrow , with fwhm @xmath11 as small as @xmath12 . when integrated over all photon energies , the rms relative amplitudes of the kilohertz qpos seen in the atoll sources range up to @xmath13% of the total 260 kev x - ray countrate and are typically much greater than the @xmath14% amplitudes of the kilohertz qpos observed in the z sources .
the amplitudes of the kilohertz qpos increase steeply with photon energy , up to at least 15 kev , in all the atoll and z sources where the photon - energy dependence of the kilohertz qpos has been studied (: berger et al.1996 ; : zhang et al .
1996 ; : strohmayer et al . 1996c
; : wijnands & van der klis 1997 ; : van der klis et al .
1996e ; : wijnands et al .
1997c ; : wijnands et al .
1998 ) .
the frequencies of the kilohertz qpos have been seen to vary by as much as a factor @xmath15 .
the frequencies of the two kilohertz qpos observed in the persistent x - ray emission between type i x - ray bursts of the atoll sources ( ford et al . 1996 , 1997a ) , ( strohmayer et al . 1996a , 1996b , 1996c ) , ( wijnands & van der klis 1997 ) , and ( smale et al.1996 , 1997 ) increase steeply with increasing countrate ( @xmath16 ) . the frequencies of the pair of kilohertz qpos seen during two different observations of ( ford et al
1997a ) correlated tightly and positively with the energy of the _ peak _ in the kev x - ray spectrum ( ford et al .
1997b ) , even though the tracks in the frequency-_countrate _ plane made by the source during the two observations differed significantly .
the frequencies of the pair of high - frequency qpos observed in the z sources 1 , , , and increase steeply with the inferred accretion rate in these sources ( van der klis et al.1996a , 1996b , 1996d , 1996e , 1997a ; wijnands et al . 1997c , 1998 ;
as discussed by van der klis [ 1989 ] and lamb [ 1989 , 1991 ] , there is no simple relation between countrate and accretion rate in the z sources ) .
two simultaneous kilohertz qpos are so common that attention has focused on models that might be able to explain the occurrence of such kilohertz qpo pairs .
strohmayer et al .
( 1996c ) and ford et al .
( 1997a ) have suggested adapting the magnetospheric beat - frequency model to explain these qpos .
this model was originally proposed by alpar & shaham ( 1985 ) and lamb et al .
( 1985 ) and developed further by shibazaki & lamb ( 1987 ) to explain the _ single _ 1560
hz horizontal - branch oscillation ( hbo ) discovered earlier in the z sources ( see van der klis 1989 ) . as we discuss in
2 , it is very difficult to understand several key features of the kilohertz qpos , including the simultaneous occurrence of two kilohertz qpos , in terms of the magnetospheric beat - frequency model .
here we propose a model for the kilohertz qpos that is fundamentally different from the magnetospheric beat - frequency model . in the model presented here ,
the x - ray source is a neutron star with a surface magnetic field @xmath17 g and a spin frequency of a few hundred hertz , accreting gas via a keplerian disk . in this model
the key characteristics of the kilohertz qpos are explained as follows ( details are presented in
3 ) : _ accretion flow_.magnetoturbulence , differentialcooling , and radiation forces create density inhomogeneities ( `` clumps '' ) in the gas in the nearly keplerian disk flow .
these clumps are dissipated by the shear in the azimuthal velocity combined with gas pressure forces and turbulence .
if the neutron star has a magnetic field @xmath18 g , some of the accreting gas is channeled out of the disk flow by the stellar magnetic field .
channeling occurs at several stellar radii , if the stellar magnetic field is @xmath19@xmath20 g , but only very close to the stellar surface , if the field is @xmath21@xmath22 g. even if the field is as strong as @xmath19@xmath20 g , some of the accreting gas continues to drift inward in a nearly keplerian disk flow until either radiation drag forces become important , the gas reaches the innermost stable circular orbit , or the gas begins to interact viscously with the stellar surface .
the fraction of the accreting gas that remains in the keplerian flow close to the star depends primarily on the strength of the stellar magnetic field and the accretion rate .
_ role of radiation drag and general relativity_.for the accretion rates typical of the z and atoll sources , the nearly keplerian flow ends and the inward radial velocity of the gas in the disk increases sharply as the gas nears the star , either because of the azimuthal drag force exerted on the gas in the disk by the radiation coming from near the stellar surface or if the drag force is weak and the radius of the neutron star is less than the radius of the innermost stable circular orbit because of the general relativistic corrections to newtonian gravity that produce an innermost stable orbit . in either case , the inward radial velocity of the gas becomes supersonic within a very short radial distance from the radius at which it begins to increase .
_ effect of inhomogeneities in the flow_.roughly speaking , clumps that form outside the radius where the radial inflow becomes supersonic are dissipated before gas from them reaches the stellar surface , whereas gas from clumps that form near the radius where the flow becomes supersonic falls inward and collides with the stellar surface before the clumps are destroyed . where the streams of denser gas from the clumps orbiting near the sonic point collide with the stellar surface , the streams create arc - shaped areas of brighter x - ray emission that move around the star s equator with a frequency equal to the orbital frequency of the clumps , which is approximately the keplerian frequency at the sonic point .
_ higher - frequency kilohertz qpo_.the frequency @xmath23 of the higher - frequency of the two kilohertz qpos is approximately equal to the keplerian frequency at the sonic point .
a detectable quasi - periodic oscillation of the x - ray flux and energy spectrum may be produced at this frequency by the periodically changing aspect ( and , for most viewing directions , by the neutron star s periodic eclipse ) of the arc - shaped areas of brighter x - ray emission , as they move around the star with a frequency equal to the orbital frequency of clumps in the keplerian disk flow near the sonic point .
the oscillation appears quasi - periodic because the clumps have finite lifetimes and slightly different orbital frequencies .
we find that @xmath24 can be as large as 100 for these oscillations , consistent with the observed coherence of the kilohertz qpos .
the frequency of the sonic - point keplerian qpo is almost independent of the spin rate of the neutron star ( frame - dragging causes the keplerian frequency at the sonic point , and hence the qpo frequency , to depend weakly on the star s spin rate ) . _
lower - frequency kilohertz qpo_.the frequency @xmath25 of the lower - frequency of the two kilohertz qpos is equal to one ( or possibly two ) times the difference between the keplerian frequency at a radius near the sonic point and the stellar spin frequency .
a qpo may be produced at this beat frequency if the magnetic field of the neutron star is weak enough that a keplerian disk flow penetrates close to the star but strong enough to channel some of the accreting gas before it collides with the stellar surface . for the accretion rates of the z and atoll sources , this requires surface magnetic fields @xmath26 in the range @xmath27 g. channeling of gas out of the disk generates a qpo at the beat frequency because it creates slightly brighter regions that _ rotate with the star _ ( these slightly brighter regions are distinct from the brighter spots that move around the star at the sonic - point keplerian frequency ) .
the slightly brighter regions that rotate with the star produce a weakly - beamed radiation pattern that rotates at the stellar spin frequency .
the radiation in this pattern produces a slight increase of the azimuthal drag force on the gas in the clumps orbiting near the sonic point , once or twice each beat period ( depending on the symmetry of the slightly brighter region ) .
this oscillation of the drag force in turn creates an oscillation of the inward flux of gas from a given clump at the sonic point once or twice each beat period .
this modulation of the mass flux in the gas streams flowing inward from all clumps orbiting near the sonic point generates a quasi - periodic oscillation of the luminosity and spectrum of the x - ray emission from the stellar surface , with a frequency equal to one ( or possibly two ) times the difference between the keplerian frequency at a radius near the sonic point and the stellar spin frequency .
the difference between the frequencies of the two qpos in a pair is therefore close to ( but not necessarily _ exactly _ equal to ) the stellar spin frequency . _
oscillation amplitudes_.the amplitudes of theqpos at the sonic - point keplerian frequency and at the beat frequency depend on the strength of the neutron star s magnetic field and the mass accretion rate , and hence one or both of these qpos may sometimes be undetectable .
oscillations at the stellar spin frequency and its overtones are expected to be weak but may sometimes be detectable .
_ similarity of kilohertz qpo frequency behavior in different sources._the similar frequency ranges and the similarity of the frequency - accretion rate correlations of the kilohertz qpos observed in the z and atoll sources which have accretion rates and magnetic fields that differ by factors of ten or more is explained by the crucial role of radiation drag , which is only effective within one or two stellar radii , the approximate proportionality of the angular momentum carried by the accreting matter and the radiation drag force , the fact that the stellar magnetic fields as well as the accretion rates are larger in the z sources , and the tendency for the vertical thickness of the disk to increase with increasing mass flux . the radius where the radiation drag or general relativistic corrections to newtonian gravity cause the radial velocity to increase abruptly plays a key role in the model of the kilohertz qpos proposed here .
once the radial velocity begins to increase , the flow becomes supersonic within a very short radial distance ( the inward radial velocity increases
so abruptly that the location of the sonic point is insensitive to the precise definition used ) , so the sonic point is a useful indicator of the point in the disk flow where the radial velocity begins to increase steeply .
hence , for convenience we shall call the point where the radial velocity increases sharply the `` sonic point '' and refer to this model for the kilohertz qpos as the `` sonic - point model '' .
we show that if the frequency of the higher - frequency qpo in a pair of kilohertz qpos is the orbital frequency of gas in a stable keplerian orbit around the neutron star , as in the sonic - point model proposed here , the frequencies of these qpos provide interesting new upper bounds on the masses and radii of the neutron stars in the z and atoll sources and new constraints on the equation of state of matter at high densities .
any model of how the kilohertz qpos are produced must be consistent with the known properties of the atoll and z sources .
hence , before analyzing the sonic - point qpo model of the kilohertz qpos in more detail , we first summarize in 2 the previously known x - ray spectral and lower - frequency x - ray variability properties of the atoll and z sources and the physical picture of these sources that has been developed based on these properties . in
3 we analyze the physics of the sonic - point model and show that it is consistent with the basic properties of the kilohertz qpos . in
4 we demonstrate that the sonic - point model is also consistent with the existing physical picture of the z and atoll sources and with many of the more detailed properties of the kilohertz qpos . in
5 we show how to derive upper bounds on the masses and radii of the neutron stars in the kilohertz qpo sources from the frequencies of stable keplerian orbits in the kilohertz range and discuss the constraints on the properties of neutron - star matter that follow from these bounds ; the bounds we derive include the effects of frame - dragging . finally , in 6
we discuss several specific predictions of the sonic - point model .
the atoll sources are lmxbs that , over time , trace atoll - shaped patterns in x - ray color - color diagrams ( hasinger & van der klis 1989 ) .
they have luminosities @xmath28 , i.e. , 110% of the eddington critical luminosity @xmath29 of a neutron star .
power - density spectra of their brightness variations show broad , band - limited noise components at frequencies below @xmath30 hz .
no qpos with frequencies @xmath31 hz have so far been detected in any of the atoll sources with the exception of 1 , which has a qpo that increases in frequency from 1 to 30 hz as the countrate increases , when the source is very bright ( oosterbroek et al
. 1995 ; shirey et al . 1996 ; bradt , shirey , & levine 1998 ) .
recent comparisons of models of the x - ray emission of neutron stars in lmxbs with the x - ray spectra of atoll sources observed with ( psaltis & lamb 1998a , 1998b , 1998c ) suggest that these sources can be subdivided into two groups , the `` 4u '' atoll sources ( such as , , , , and ) and the `` gx '' atoll sources ( such as , , , and ) , based on the strengths of their inferred magnetic fields ; one atoll source , , has intermediate spectral properties and therefore probably has a magnetic field of intermediate strength .
the z sources are lmxbs that produce a characteristic z - shaped track in x - ray color - color diagrams ( hasinger & van der klis 1989 ) .
they have luminosities @xmath32 , i.e. , comparable to @xmath29 .
the three branches of the z are called the horizontal , normal , and flaring branches . when a z source is on the horizontal branch , a qpo with a frequency in the range 1560 hz is observed ( van der klis et al 1985 ; van der klis 1989 ) .
this `` horizontal branch oscillation '' ( hbo ) is also detectable in some z sources when they are on the upper part of the normal branch .
the relative width @xmath33 of the hbo peak in power density spectra is typically 0.10.3 and its centroid frequency typically increases with increasing countrate . as a z source moves down the normal branch ,
the hbo becomes weaker and eventually disappears into the noise continuum .
near the middle of the normal branch , a different qpo appears .
the peaks in power density spectra produced by this qpo have relative widths @xmath34 and centroid frequencies in the range 48 hz ( middleditch & priedhorsky 1986 ; van der klis 1989 ) .
the properties of this qpo do not vary appreciably on the lower normal branch . as a z source moves from the lower normal branch to the flaring branch , the frequency of this second qpo increases abruptly to @xmath35 hz and the qpo becomes weaker and less coherent , eventually disappearing into the noise ( van der klis 1995 ; dieters & van der klis 1997 ) .
this qpo is the `` normal / flaring branch oscillation '' ( n / fbo ) .
in addition to these two types of quasi - periodic oscillations , power spectra of the brightness variations of the z sources also show three distinct band - limited noise components at frequencies below 200 hz ( hasinger & van der klis 1989 ) .
the properties of the qpo and noise components vary systematically with the position of a source on its z track ( hasinger & van der klis 1989 ; see kuulkers 1995 , dieters & van der klis 1997 , and wijnands et al .
1997a for detailed studies of this behavior and possible exceptions ) .
recent analyses of archival data ( kuulkers 1995 ; kuulkers et al .
1995 , 1996 ) suggest that the z sources can be subdivided into two groups , based on the morphology of their z tracks and power spectra : the `` cyg - like '' z sources ( 2 , , and ) and the `` sco - like '' z sources ( 1 , , and ) . at a more detailed level , shares some of the characteristics of the `` cyg - like '' sources and therefore probably has intermediate properties ( hasinger & van der klis 1989 ; wijnands et al .
1997a ) . the atoll and z sources are neutron stars accreting gas from a keplerian disk fed by a low - mass companion star ( see van der klis 1989 ) . the magnetic fields and the accretion rates of these neutron stars are thought to be the most important parameters that determine their x - ray spectral and temporal characteristics .
according to the most complete and self - consistent current model of these sources , the so - called _ unified model _ ( lamb 1989 , 1991 ) , the atoll sources have dipole magnetic fields @xmath36 g and luminosities 110% of the eddington critical luminosity , whereas the z sources have dipole fields @xmath37 g and luminosities very close to ( and sometimes slightly above ) the eddington critical luminosity .
_ x - ray spectra._the x - ray spectra of the z and atoll sources and the low upper limits on the amplitudes of any periodic variations of their persistent x - ray brightness ( less than 1% in some cases ; see , e.g. , vaughan et al .
1994 ) constrain the properties of these neutron stars .
the low upper limits on brightness variations at their spin frequencies can be understood if the magnetic fields of these neutron stars are @xmath38 g and the magnetospheres and inner disks are surrounded by a central coronae with electron scattering optical depths @xmath39 , even at low accretion rates ( lamb et al.1985 ; lamb 1989 , 1991 ) , because such coronae strongly suppress the x - ray flux oscillation produced by any radiation pattern that rotates with the star ( brainerd & lamb 1987 ) .
this effect is discussed further in 3.6 .
when , as in the z sources , the total luminosity of the neutron star , inner disk , and central corona becomes comparable to @xmath29 , the vertical radiation force drives gas upward , out of the disk and radiation drag removes its angular momentum in less than one orbit , creating a region of approximately radial inflow that extends out to @xmath40 km ( lamb 1989 , 1991 ) .
oscillations in this radial flow are thought to be responsible for the n / fbo ( lamb 1989 ; fortner et al . 1989 ; fortner 1992 ; miller & lamb 1992 ) .
detailed physical modeling of the x - ray spectra of the atoll and z sources indicates that soft ( @xmath41 kev ) photons are produced by optically - thick bremsstrahlung and other processes at the surface of the star and by self - absorbed , high - harmonic cyclotron emission in the inner magnetosphere ( psaltis , lamb , & miller 1995 ; psaltis & lamb 1998a , 1998b , 1998c ) .
as we discuss below , the accretion rate and the strength of the stellar magnetic field determine whether emission from the stellar surface or cyclotron emission in the magnetosphere is the dominant source of soft photons .
these soft photons are upscattered by electrons in the magnetosphere and the central corona that surrounds the neutron star to produce the x - ray spectrum that emerges from the corona . in the z sources ,
scattering by cooler electrons in the region of approximately radial inflow that surrounds the inner disk and corona further deforms the x - ray spectrum .
cyclotron emission is the dominant source of soft photons if upscattered cyclotron photons are able to supply the full accretion luminosity ( see psaltis et al .
the range of stellar magnetic fields for which cyclotron emission is dominant can be estimated as follows . within the magnetosphere
, the spectrum of self - absorbed cyclotron photons can be approximated by a blackbody spectrum truncated at the energy at which cyclotron emission becomes optically thin .
comptonization within the magnetosphere and central corona increases the energy of a typical photon by a factor @xmath42 , where @xmath43 is the compton @xmath44 parameter , @xmath45 and @xmath46 are the electron temperature and rest mass , @xmath47 is the electron scattering optical depth , and @xmath48 is the boltzmann constant ( see rybicki & lightman 1979 , pp 195223 ) . hence cyclotron emission is the dominant source of soft photons if @xmath49 or , equivalently , @xmath50 where @xmath51 is the mass accretion rate onto the neutron star surface , @xmath52 is the effective radius of the cyclotron photosphere , @xmath53 is the harmonic number at which the transition from optically thick to optically thin emission occurs , and @xmath54 is the radius of the neutron star .
figure [ fig : parameterspace ] shows how @xmath55 depends on for the range of electron temperatures expected in the magnetospheres and central coronae of the atoll and z sources .
as figure [ fig : parameterspace ] shows , if the star s magnetic field is and the accretion rate is , the dominant source of photons is the thermal emission from the surface of the neutron star .
these photons are then upscattered by the electrons in the magnetosphere and the central corona .
numerical calculations of the x - ray spectra produced by stars with these field strengths and accretion rates agree well with observations of the spectra of the `` 4u '' atoll sources when they are in the so - called `` banana '' spectral state ( psaltis & lamb 1998c ) . if instead the star s magnetic field is @xmath56 g and the accretion rate is @xmath57 , cyclotron emission is the dominant source of photons .
these photons are then upscattered by the electrons in the magnetosphere and the central corona .
numerical calculations of the x - ray spectra produced by stars with these field strengths and accretion rates agree well with observations of the spectra of the `` gx '' atoll sources ( psaltis & lamb 1998a , 1998b , 1998c ) .
if the stellar magnetic field is and the mass accretion rate is , electron cyclotron emission in the magnetosphere is very efficient in producing soft photons . at these accretion rates soft photons
are comptonized not only by the hot electrons in the magnetosphere and central corona but also by the cool electrons in the approximately radial inflow .
numerical computations of the x - ray spectra and color tracks predicted by this physical model agree well with @xmath58 and @xmath59 measurements of the x - ray spectra and color tracks of the z sources ( psaltis et al . 1995 ; psaltis & lamb 1998a , 1998b , 1998c ) .
_ accretion flows_.if the magnetic field of the neutron star is @xmath60 g and the accretion rate is @xmath61 , the keplerian disk flow may extend inward all the way to the surface of the neutron star without being channeled by the stellar magnetic field .
( here and below we refer to accretion flows confined near the orbital plane as _ disk flows _ , whether or not the azimuthal velocity field is nearly keplerian , and as _ keplerian disk flows _ if the azimuthal velocity field is nearly keplerian , that is , if the flow is nearly circular . )
if instead the stellar magnetic field is @xmath62 g , most of the gas in the disk will couple to the stellar field and be funneled out of the disk plane toward the magnetic poles of the star at a characteristic cylindrical radius @xmath63 ( ghosh & lamb 1979a , 1979b , 1992 ) that is much larger than the radius of the star .
hence , the disk flow around such a star ends far above the stellar surface . for stellar magnetic fields of intermediate strength ,
some of the accreting gas is likely to couple to the stellar magnetic field beginning at @xmath63 , which will then channel it toward the star s magnetic poles , but some is also likely to continue as a disk flow inside @xmath63 , as a result of rayleigh - taylor instability and incomplete coupling of the flow in the inner disk to the stellar magnetic field ( see , e.g. , scharlemann 1978 ; ghosh & lamb 1979a , 1991 ; lamb 1984 ; spruit & taam 1990 ) . in this case
we expect some of the gas in the disk flow to be channeled out of the disk by the magnetic field over a range of radii inside @xmath63 , with the rest of the gas remaining in a geometrically thin keplerian flow that penetrates close to the stellar surface , as shown in figure [ fig : sideviews ] .
this last case is the one that we expect to be relevant to the z sources and most of the atoll sources .
ghosh & lamb ( 1992 ) derived expressions for the characteristic cylindrical radius @xmath63 at which gas in the disk couples strongly to the stellar magnetic field and begins to be channeled out of the disk flow , as a function of the stellar dipole magnetic moment @xmath64 , the mass accretion rate through the inner disk , and the neutron star gravitational mass @xmath65 , both for gas - pressure - dominated ( gpd ) and for radiation - pressure - dominated ( rpd ) inner disks , assuming @xmath63 is much larger than the radius @xmath54 of the star .
the gpd model is expected to be valid for the atoll sources .
the disks around the z sources are expected to be rpd .
the scaling of @xmath63 with @xmath64 , , and @xmath65 given by the analytical rpd model is expected to be valid for the z sources , but the vertical thickness of the boundary layer given by the model is not expected to be reliable .
we have therefore reduced the vertical thickness of the boundary layer in the model so that the keplerian frequency at the radius where the gas couples strongly to the magnetic field agrees quantitatively with the keplerian frequency inferred from spin frequency of 2 , based on the frequencies of its kilohertz qpos ( wijnands et al .
1998 ; psaltis et al .
1998 ) and the magnetospheric beat - frequency interpretation of its hbo ( psaltis et al .
1998 ) , for a magnetic moment equal to the moment inferred from its x - ray spectrum ( psaltis & lamb 1998c ) .
specifically , the keplerian frequency at the coupling radius was assumed to be 380 hz , as indicated by the sum of the spin and hbo frequencies when 2 is on the normal branch , for a magnetic moment of @xmath66 and a mass flux through the inner disk equal to the mass accretion rate onto the neutron star that produces the eddington critical luminosity .
the resulting two expressions for @xmath67 can be solved for the dipole moment @xmath68 that gives a keplerian frequency @xmath69 at the coupling radius ; this characteristic dipole moment is @xmath70 here @xmath71 . if the star s dipole moment exceeds @xmath68 , most of the gas in the disk flow will couple to the magnetic field and be channeled out of the disk at a radius where the keplerian frequency is less than @xmath69 .
figure [ fig : parameterspace ] shows the dipole moment that gives keplerian frequencies of 500 and 1100 hz at the main coupling radius , as a function of the mass accretion rate .
these two curves show that for the neutron star magnetic fields inferred from x - ray spectral modeling , the keplerian frequency at the gas coupling radius can not be made simultaneously consistent with the @xmath105001100 hz ranges of the kilohertz qpos in the 4u and z sources . accreting gas that is not channeled out of the disk plane at @xmath63
is expected to continue to drift slowly inward in a keplerian disk flow until ( a ) it is channeled out of the disk at smaller radii , ( b ) it begins to flow rapidly inward , either because it loses angular momentum to the radiation drag force or because it crosses the radius of the marginally stable orbit , or ( c ) it collides with the stellar surface .
gas that is flowing rapidly inward remains in a ( non - keplerian ) disk flow unless it is channeled out of the disk by the stellar magnetic field .
we emphasize that _ the accretion disk ends above the stellar surface only if all the gas in the disk is channeled out of the disk flow before it reaches the star_. in the magnetospheric beat - frequency model of the 1560 hz hbo observed in the z sources ( alpar & shaham 1985 ; lamb et al . 1985 ; shibazaki & lamb 1987 ) , a substantial fraction of the gas in the keplerian disk flow is channeled out of the disk at @xmath72 stellar radii and flows toward the star s magnetic poles .
this model therefore requires that the z sources have magnetic fields @xmath73 g. the atoll sources have not previously shown any direct indications that they have dynamically important magnetic fields ( e.g. , no hbo - like qpos have so far been observed in an atoll source ) , and hence their magnetic fields are thought to be 5 g. however , the existence of kilohertz qpos at the difference between an orbital frequency and the spin frequency of the neutron star indicates that most if not all of the atoll sources have magnetic fields @xmath74 g. this is consistent with the detailed modeling of their x - ray spectra described above , which indicates that some atoll sources have magnetic fields as strong as 5 g. these sources may produce weak hbos , which may be detected in the future . as noted in 1 , strohmayer et al .
( 1996c ) and ford et al .
( 1997a ) have suggested adapting the magnetospheric beat - frequency qpo model to explain the two simultaneous kilohertz qpos . as described in 2.2 , the magnetospheric beat - frequency model was originally proposed by alpar & shaham ( 1985 ) and lamb et al .
( 1985 ) and developed further by shibazaki & lamb ( 1987 ) to explain the _ single _ , variable frequency , 1560 hz hbo seen in the z sources .
strohmayer et al . and ford et al .
have suggested that the frequency of the higher - frequency qpo in a kilohertz qpo pair might be the keplerian orbital frequency at the radius where gas in the accretion disk begins to couple to the stellar magnetic field and that the frequency of the lower - frequency qpo is the beat between this frequency and the stellar spin frequency .
there are several difficulties with this application of the magnetospheric beat - frequency model .
_ widespread occurrence of two simultaneous , highly coherent kilohertz qpos_. the occurrence of two simultaneous , strong ( amplitudes up to @xmath13% rms ) , and highly coherent ( @xmath75100 ) kilohertz qpos appears inconsistent with the magnetospheric beat frequency model , because a strong , highly coherent oscillation with a frequency equal to the orbital frequency where gas first couples to the stellar magnetic field is not expected , nor is it observed in other types of sources where the magnetospheric beat frequency mechanism is thought to be operating . the magnetospheric beat - frequency mechanism is expected to produce a single , strong , moderately coherent oscillation with a frequency equal to the beat frequency ( power is of course also expected at overtones of this frequency ) , but it is not expected to produce a strong , highly coherent oscillation with a frequency equal to the orbital frequency at the initial gas coupling radius , for two reasons .
first , no effect is known that would select a sufficiently narrow range of radii near this radius .
second , even if a narrow range of radii were selected , no mechanism is known that would make the orbital frequencies at these radii visible .
gas is expected to leave the disk along field lines that thread a very narrow annulus where the gas couples strongly to the magnetic field of the star ( ghosh & lamb 1979a ) .
the range of orbital frequencies in this annulus is likely to be very small , consistent with the @xmath0510 hz observed width of the hbo ( lamb et al . 1985 ) .
the magnetic field is expected to couple more weakly to the gas in the disk over a wider annulus ( ghosh & lamb 1979a ) . however , in both annuli the gas - magnetic field interaction repeats each beat period , not each orbital period ( lamb et al .
1985 ) , so there is no obvious way that this interaction would make the orbital period visible in x - rays . no other mechanism is known that would single out a narrow range of radii in the inner disk .
even if a narrow range of radii were singled out , little luminosity is expected to be generated in such a narrow annulus , as was explained in the context of the hbo by lamb ( 1988 ) .
generation of a qpo by beaming of radiation via periodic occultation of the emission from the neutron star by clumps orbiting at the initial coupling radius would require us to be viewing _
all _ the kilohertz qpo sources from within a small range of special inclinations and would also require the height of the disk at its inner edge to vary with accretion rate in just the right way ( again see lamb 1988 ) .
moreover , such beaming would be strongly suppressed by scattering in the ionized gas that surrounds the neutron star ( see lamb et al .
1985 ; brainerd & lamb 1987 ; kylafis & phinney 1989 ; lamb 1988 , 1989 ; psaltis et al .
1995 ) . a qpo that appears to be
the magnetospheric beat frequency qpo has been observed in several accretion - powered pulsars ( see , e.g. , angelini , stella , & parmar 1989 ; lamb 1989 ; shibazaki 1989 ; shinoda et al .
1990 ; lamb 1991 ; finger , wilson , & harmon 1996 ; ghosh 1996 ) . in these pulsars only a single , fairly coherent oscillation has been seen , with properties which indicate that it is the magnetospheric beat - frequency qpo ; no qpo is observed at what would be the keplerian frequency at the coupling radius . to summarize this point ,
the theory of the magnetospheric beat - frequency qpo mechanism predicts , and observations of accretion - powered pulsars confirm , that this mechanism produces only a _
strong , moderately coherent oscillation , with a frequency equal to the beat frequency . in order to explain the two simultaneous kilohertz qpos ,
the magnetospheric beat - frequency mechanism would have to generate two strong , _ highly coherent _ oscillations , at the beat frequency _ and _ at the keplerian frequency at the radius where gas in the disk first strongly couples to the magnetic field .
this is the most important reason why the magnetospheric beat - frequency mechanism appears unpromising as an explanation of the two kilohertz qpos .
_ simultaneous occurrence of two kilohertz qpos and the hbo in the z sources_. identification of the 1560 hz hbo in the z sources as the magnetospheric beat - frequency oscillation is supported by the fact that ( 1 ) the magnetic field strengths required are in agreement with those inferred from the spin - down rates of the millisecond recycled rotation - powered pulsars and detailed modeling of the x - ray spectra of the z sources ( see alpar & shaham 1985 ; lamb et al .
1985 ; ghosh & lamb 1992 ; and 2.2 ) , ( 2 ) the @xmath0200350 hz neutron star spin frequencies predicted by the magnetospheric beat - frequency model ( lamb et al . 1985 ; ghosh & lamb 1992 ) are consistent with the spin rates of the recycled rotation - powered pulsars ( ghosh & lamb 1992 ) and the @xmath0250350 hz spin frequencies inferred from the difference between the frequencies of the two kilohertz qpos seen in the z sources ( see 1 ) , and ( 3 ) only the fundamental and ( sometimes ) the first overtone of the hbo have ever been detected in the z sources when they are on the horizontal branch , as one would expect if the hbo is the magnetospheric beat - frequency qpo ( see van der klis 1989 ) .
if the hbo _ is _ the magnetospheric beat - frequency oscillation in the z sources , then _ neither _ of the two kilohertz qpos can be the magnetospheric beat - frequency oscillation , because the hbo was always present when the kilohertz qpos were detected ( van der klis et al .
1996e ) .
_ steep increase of kilohertz qpo frequencies with increasing mass accretion rate_. in at least some sources , the frequencies of both kilohertz qpos increase more steeply with inferred mass flux through the inner disk than is expected in the magnetospheric beat - frequency model , which predicts @xmath76 for a gpd disk and @xmath77 for an rpd disk ( ghosh & lamb 1992 ; see figure 2 ) .
it is therefore difficult to explain why the frequencies of the kilohertz qpos increase so steeply with using this model .
the magnetospheric beat - frequency model _ can _ explain the steep increase of the hbo frequency with observed in the z sources , because these neutron stars are likely to have been spun up to frequencies comparable to the keplerian frequency at the gas - magnetic field coupling radius , so the hbo frequency in a z source is expected to be small compared to either the star s spin frequency or the keplerian frequency at the gas coupling radius ( see lamb et al .
1985 ; shibazaki & lamb 1987 ) .
this expectation is consistent with the @xmath0250350 hz spin frequencies inferred from the differences between the frequencies of the qpos in the kilohertz qpo pairs .
in contrast to the comparatively low frequency of the hbo , the frequency of the lower - frequency qpo in the kilohertz qpo pairs ( which is the beat frequency in the magnetospheric beat - frequency interpretation ) is up to @xmath00.7 times the frequency of the higher - frequency qpo ( which in this interpretation is the keplerian frequency where the gas in the disk couples to the magnetic field ) .
it follows that in this interpretation the neutron star spin frequency is in some cases only @xmath78 times the keplerian frequency at the coupling radius .
hence the frequencies of both of the two kilohertz qpos should increase as @xmath79 increases in the same way that @xmath80 increases , which is much more slowly than observed , as figure 2 shows . to summarize , whereas the inferred dependence of the hbo on the mass flux through the inner disk is explained naturally by the magnetospheric beat - frequency model of the hbos ,
the inferred dependence of the kilohertz qpos on this mass flux is difficult to explain using this model .
_ observed range of kilohertz qpo frequencies_. the frequencies of the qpos at what would be the keplerian orbital frequency in the magnetospheric beat - frequency interpretation all range up to @xmath010001200 .
this is understandable if these qpos are generated within a few kilometers of the surface of the neutron star , as in the sonic - point model , whereas in the magnetospheric beat - frequency model the frequencies of these qpos would be expected _ a priori _ to range from @xmath050 hz up @xmath01500 hz for the wide range of neutron star magnetic field strengths and accretion rates inferred in the atoll and z sources .
the expected range of orbital frequencies could be reduced if the mass accretion rate and the strength of the star s dipole magnetic field are tightly correlated in just the right way . however , the correlation that is indicated by spectral modeling ( psaltis et al .
1995 ; psaltis & lamb 1998a , 1998b , 1998c ) gives magnetospheric beat frequencies in the range @xmath020200 hz and orbital frequencies in the range @xmath0300500 hz , much smaller than the frequencies of the kilohertz qpos .
_ anticorrelation between kilohertz qpo amplitudes and inferred magnetic field strengths_. in the magnetospheric beat - frequency interpretation , the kilohertz qpos should have higher amplitudes in systems with stronger neutron star magnetic fields . however , as discussed further in 4.4 , the rms amplitudes of the kilohertz qpos in the atoll sources , which are thought to have the weakest neutron star magnetic fields , are considerably stronger than the rms amplitudes of the kilohertz qpos in the z sources , which are thought to have the strongest neutron star magnetic fields .
in the next section we describe a model that is consistent with both the physical picture of the atoll and z sources developed over the past decade and with the properties of the kilohertz qpos .
we now analyze the physics of the sonic - point model of the kilohertz qpos . in
3.1 we describe the key elements of the sonic - point model and summarize the results of the general relativistic gas dynamical and radiation transport calculations that are described in more detail later in this section . in 3.2 we investigate the motion of the gas in the inner disks of the z and atoll sources and
show that at a characteristic angular momentum loss radius @xmath81 that is typically several kilometers larger than the radius of the neutron star , radiation drag extracts angular momentum from the gas so quickly that the gas accelerates sharply inward . as a result ,
density fluctuations in the gas near the sonic point create brighter footprints that rotate around the stellar equator at the sonic point keplerian frequency @xmath82 , producing a qpo at this frequency .
we show further that a decrease in @xmath81 with accretion rate is to be expected . in 3.3 we show that the sonic - point keplerian frequency is expected to be between @xmath0300 hz and @xmath01200 hz in all sources and should increase with increasing mass flux through the inner disk .
the model therefore explains one of the most important features of the kilohertz qpos .
we also show that the sonic - point mechanism naturally produces a second qpo at the beat frequency between @xmath82 and the stellar spin frequency @xmath83 . in
3.4 we show that this mechanism can produce qpos with @xmath84 , comparable to the largest q values observed . in 3.5 we consider radiation transport in the footprints and surrounding hot central corona and show that the amplitude of the qpo at the sonic - point keplerian frequency is expected to increase steeply with increasing photon energy .
also , the keplerian - frequency oscillations at higher photon energies are expected to be delayed by @xmath85sec relative to the oscillations at lower energies , to the extent that the effects of comptonization dominate . in
3.6 we consider the effect of gas surrounding the neutron star and magnetosphere on the amplitudes of the kilohertz qpos and demonstrate that qpo amplitudes as large as @xmath86% are understandable in the sonic - point model .
the light travel time across a neutron star is a fraction @xmath87 of the time required for accretion to change the star s mass and spin rate , so the exterior spacetime is stationary to extremely high accuracy .
the spacetime outside a steadily and uniformly rotating axisymmetric star with gravitational mass @xmath65 and angular momentum @xmath88 is unique to first order in the dimensionless angular momentum parameter @xmath89 ( hartle & thorne 1968 ) and is the same as the kerr spacetime to this order ( the spacetime outside a rotating star differs from the kerr spacetime in second and higher orders ; see , e.g. , cook , shapiro , & teukolsky 1994 ) .
all , or almost all , of the atoll and z sources appear to have spin frequencies 350 hz ( see 2 and
4 ) , in which case @xmath90 is for these stars ( see 5.2 ) .
hence the kerr spacetime is a reasonably accurate approximation to the exterior spacetime .
therefore , in this work we use the familiar boyer - lindquist coordinates . _
all analytical expressions given in this paper are accurate only to first order in @xmath90_. these expressions would therefore be the same if written in terms of the circumferential radius .
calculation of physical quantities to higher order is straightforward , given a stellar model and interior and exterior metrics valid to the higher order , but no analytical expressions are known , so such calculations must be carried out numerically . for conciseness ,
in this section ( only ) we use units in which @xmath91 , where @xmath92 is the speed of light , @xmath93 is the gravitational constant , and @xmath48 is boltzmann s constant . before presenting our calculations of the most important elements of the sonic - point model ,
we first describe the model in more detail , summarizing the results of our calculations and indicating where these results are presented in the subsections that follow .
we shall assume that the spin axes of the neutron stars in the kilohertz qpo sources are closely aligned with the rotation axes of their accretion disks .
this is expected to be the case in these lmxbs , because mass transfer is expected to produce an accretion disk in which the gas circulates in the same sense as the orbital motion of the system .
the torque on the neutron star created by such a disk will align the spin axis of the star with the axis of the accretion disk in a time short compared to the duration of the mass transfer phase in such systems ( see daumerie et al . 1996 ) .
_ accretion flow near the star_.as discussed in 2.2 , if the stellar magnetic field is of intermediate strength we expect a keplerian disk flow to penetrate inside the magnetosphere .
what happens to the flow there depends on the magnitude of the radiation drag force ( miller & lamb 1993 , 1996 ) and on whether the radius @xmath54 of the neutron star is larger or smaller than the radius @xmath94 of the innermost stable circular orbit .
there are four possibilities ( see table [ table : flowcases ] for a summary ) .
suppose first that @xmath95
. this will be the case if the equation of state of neutron - star matter is relatively soft and the star in question has an intermediate - to - high mass and is not spinning near its maximum rate .
there are then two possibilities ( 3.2 ) : ( 1a ) if the drag force exerted by radiation coming from near the neutron star s surface is strong enough to remove @xmath96% of the angular momentum of the gas in the keplerian disk flow just outside @xmath94 , gas in the inner disk drifts slowly inward in nearly circular orbits until it reaches the critical angular momentum loss radius @xmath81 , where it transfers so much angular momentum to the radiation so rapidly that centrifugal support fails and the radial velocity of the gas increases sharply .
inside @xmath81 the gas spirals inward in a disk flow in which the radial velocity is supersonic until the gas is channeled by the star s magnetic field or collides with the star s surface .
( 1b ) if instead the radiation drag is so weak that it removes @xmath97% of the angular momentum of the gas in the keplerian disk flow by the time the gas reaches @xmath94 , gas in the inner disk drifts slowly inward in nearly circular orbits until it approaches @xmath98 , where general relativistic corrections to newtonian gravity cause the radial velocity of the flow to increase steeply ( see , e.g. , muchotrzeb 1983 ; muchotrzeb - czerny 1986 ) . inside @xmath94 ,
the gas spirals inward in a disk flow in which the radial velocity is supersonic until the gas is channeled by the star s magnetic field or collides with the star s surface . whether case 1a or 1b applies depends primarily on the strength of the star s magnetic field and on the mass accretion rate .
suppose now that @xmath99
. this will be the case if the equation of state of neutron - star matter is relatively stiff and the star in question has an intermediate - to - low mass or is spinning very rapidly .
there are again two possibilities ( 3.2 ) : ( 2a ) if the radiation drag force is at least moderately strong , gas in the inner disk drifts slowly inward until it reaches the critical radius @xmath81 , where it accelerates rapidly inward .
inside @xmath81 the gas spirals inward in a disk flow in which the radial velocity is supersonic until the gas is channeled by the star s magnetic field or collides with the star s surface .
( 2b ) if instead the radiation drag is weak , gas in the inner disk drifts slowly inward in a disk flow in which the radial velocity is supersonic until the gas is channeled by the star s magnetic field or interacts directly with the star s surface .
again , whether case 2a or 2b applies depends primarily on the strength of the star s magnetic field and on the mass accretion rate .
regardless of whether the slow inward drift of the gas in the disk is terminated by loss of angular momentum to the radiation field or because the gas crosses the radius of the innermost stable circular orbit , the inward radial velocity of the gas increases abruptly at @xmath81 or @xmath94 and becomes supersonic within a very small radial distance . as mentioned in 1 and earlier in this section and described in detail below , the radius where either radiation drag or general relativistic corrections to newtonian gravity cause the radial velocity to increase sharply and become supersonic plays a key role in the model of the kilohertz qpos proposed here .
the radius of the sonic point is a useful indicator of the location of this transition .
radiation drag can remove at most only a fraction of the specific angular momentum of the gas and therefore can create a radially supersonic inflow in the inner disk only within a few stellar radii ( 3.2 ) .
the reason is that the specific angular momentum of gas in circular keplerian orbits at radii much larger than the radius of the neutron star is much greater than in orbits just outside the stellar surface , so removal of a small fraction of the angular momentum of the gas at large radii can not cause a large fraction of it to fall to the stellar surface .
in contrast , the specific angular momentum of gas orbiting near the star is not much greater than the angular momentum of gas orbiting at the stellar surface , so radiation drag can cause gas in orbit near the star to plunge to its surface . radiation drag can cause gas to plunge inward from a radius larger than one would estimate using the newtonian approximation , because ( 1 ) special and general relativistic effects on the gas dynamics and radiation transport significantly increase the fraction of the angular momentum of the accreting gas that can be removed by the radiation and ( 2 ) in general relativity the specific angular momentum of gas in circular keplerian orbits with radii varies only slowly with @xmath100 .
for this reason , the sonic - point model predicts that if radiation drag produces a transition to rapid radial inflow , the transition will occur within ( for a @xmath101 neutron star ) and hence that the keplerian frequency at the transition radius will be ( 3.3 ) .
inside the sonic point , the vertical optical depth of the disk flow falls steeply with decreasing radius , usually to a value small compared to unity .
the optical depth of the disk flow measured radially from the stellar surface to the sonic radius typically also becomes less than unity within a very small radial distance , unless the accretion rate is very high or the geometrical thickness of the disk in the vertical direction is very small .
when the change from `` optically thick '' to `` optically thin '' disk flow is caused by the radiation drag force , the transition is somewhat analogous to the ionization front at the boundary of an hii region . in the disk flow transition
, the photon mean free path becomes longer because the radiation is removing angular momentum and the flow is accelerating inward , causing the density to fall sharply , whereas in an ionization front the radiation is removing bound electrons from atoms and molecules , causing the opacity to fall sharply .
_ generation of the qpo at the sonic - point keplerian frequency_.suppose first that the magnetic field of the neutron star is dynamically negligible .
an element of gas in the disk flow outside the sonic radius diffuses inward subsonically .
when it reaches the sonic radius , it falls supersonically to the stellar surface along a spiral trajectory like that shown in figure [ fig : spirals]a .
this trajectory was computed in full general relativity for a nonrotating , isotropically emitting star , using the numerical algorithm described in miller & lamb ( 1996 ) .
( the shape of the trajectory depends on the luminosity , spin , and other properties of the source ; the example shown in fig .
[ fig : spirals]a is only illustrative . )
the surface density of the disk flow is much smaller inside the sonic radius than outside because of the sharp increase in the radial velocity at the sonic point .
gas falls inward from the sonic radius and impacts the star all around its equator , producing a bright equatorial ring of x - ray emission .
the flow in the inner part of the accretion disk is expected to have density fluctuations ( `` clumps '' ) produced by a variety of mechanisms , such as thermal instability , kelvin - helmholtz instability , and magnetoturbulence ( see lamb et al .
1985 ; shibazaki & lamb 1987 )
. there may be as many as several hundred such clumps at a given radius in the disk outside the sonic point .
the velocity - independent radially outward component of the radiation force and the velocity - dependent radiation drag force both tend to cause gas in the shadow of a clump to overtake the clump in azimuth and radius , and hence the radiation force tends to increase clumping . as discussed in 3.4 , a clump that forms outside the sonic radius is dissipated by gas pressure forces , turbulence , and the shear in the velocity field before the gas from it can reach the stellar surface ,
so such clumps do not produce significant inhomogeneities in the inflow from the sonic radius .
in contrast , a clump that forms near the sonic radius can persist for up to @xmath102 times the infall time to the stellar surface , so such a clump generates a stream of denser infalling gas from the sonic point to the stellar surface that lasts as long as the clump survives .
the shape of the pattern formed by the denser gas falling inward from a clump depends on how the angular velocity of the infalling gas varies with radius .
for example , if the angular velocity of the gas were independent of radius , the pattern of higher gas density would be a straight , radial line from the clump to the stellar surface .
in general , the pattern formed by the denser gas from a clump is a fairly open curve .
the shape of the pattern , like the trajectories that produce it , depends on the luminosity , spin , and other properties of the source .
figure [ fig : spirals]b shows the pattern formed by inflow of gas along spiral trajectories with the shape shown in figure [ fig : spirals]a . in this example
the pattern of denser gas is also a spiral .
it is more open than the spiral trajectories of the individual elements of denser gas that produce it , because the source of the denser gas ( the clump ) orbits the star only slightly slower than the infalling gas .
we expect the time - averaged radiation field around the star to be nearly axisymmetric .
moreover , the gas from a given clump typically orbits the star @xmath103 times before colliding with the stellar surface , so the effects of any azimuthal variations in the radiation drag force are averaged out .
therefore , the spiral pattern of denser gas produced by the inflow from a given clump rotates nearly uniformly around the star with a rotation frequency equal to the orbital frequency of the clump that is producing it , which is the keplerian frequency at a radius near the sonic point .
the time evolution that generates an x - ray oscillation at the sonic - point keplerian frequency is illustrated by the four panels of figure [ fig : keplerqpo ] . in this figure a single clump is shown advancing in its orbit by 90 from one panel to the next .
the pattern of higher density gas rotates uniformly around the star with a frequency equal to the orbital frequency of the clump , so it also advances by 90 from one panel to the next .
where the denser gas from a clump collides with the stellar surface , it produces an arc - shaped area of brighter x - ray emission .
this arc - shaped brighter `` footprint '' moves around the star s equator with a frequency equal to the rotation frequency of the pattern , which is the orbital frequency of the clump and is therefore approximately equal to the keplerian frequency @xmath104 at the sonic point . in reality
, many clumps are crossing the sonic radius at any given time , and hence there are many bright footprints moving around the star s equator .
the radiation from these footprints carries the kinetic energy of infall that is released when the gas that has fallen inward from the clumps at the sonic radius collides with the stellar surface . as seen by a distant observer
whose line of sight is inclined with respect to the orbital axis of the disk , the aspect presented by a given bright footprint varies periodically and the footprint is eclipsed with a frequency equal to the rotation frequency of the pattern , which is approximately the keplerian orbital frequency at the sonic point .
footprints come and go as clumps form near the sonic point and then dissipate .
also , the orbital frequencies of the clumps that are producing footprints at any given time differ slightly . as a result , a distant observer sees a strong , _ quasi - periodic _ oscillation of the x - ray flux and spectrum at the pattern rotation frequency , which is close to @xmath104 .
_ this is the the sonic - point keplerian - frequency qpo_. the spin of the neutron star is not involved in generating the sonic - point keplerian frequency qpo .
the precise frequency of the sonic - point keplerian qpo does depend weakly on the spin frequency of the neutron star , because the frame - dragging caused by the star s spin has a small effect on the frequencies of keplerian orbits near the neutron star .
the sonic - point keplerian frequency qpo mechanism described here , which generates a qpo with a frequency equal to the orbital frequency at the sonic point , differs fundamentally from the suggestion of paczynski ( 1987 ) , who speculated that when the sonic point is located at the radius of the marginally stable orbit , a qpo might be produced by unsteady flow through it with a low ( @xmath105 ) frequency that would have nothing directly to do with the orbital frequency of the marginally stable orbit . _
generation of the qpo at the sonic - point beat frequency_.suppose now that the magnetic field of the neutron star is weak enough that a prograde keplerian disk flow penetrates near the surface of the star but strong enough that close to the star some of the gas is channeled by the field .
the channeled gas produces slightly brighter spots where it collides with the stellar surface .
if these slightly brighter spots are offset from the star s spin axis , the enhanced radiation from them generates a beamed pattern of radiation that rotates with the star .
this rotating radiation pattern creates a periodic oscillation in the radiation force that is acting on the clumps of gas crossing the sonic radius .
hence , if conditions near the star are such that the nearly keplerian motion of the gas in the clumps orbiting near the sonic radius is terminated by loss of angular momentum to the radiation , the inward flux of mass from the clumps near the sonic radius will increase once or twice each beat period , depending on the symmetry of the radiation pattern that rotates with the star . as a specific
example , suppose that the radiation drag force experienced by the gas in a given clump orbiting near the sonic radius peaks each time the beam of radiation that rotates with the star sweeps across the clump . this will occur with a frequency equal to one or two times the difference between the orbital frequency of the clump , which is the keplerian orbital frequency near the sonic point , and the spin frequency @xmath83 of the neutron star .
_ we call @xmath106 the sonic - point beat frequency and denote it @xmath107_. if the transition to supersonic inflow is caused by radiation drag , the oscillation of the radiation drag force will cause the supersonic inflow of gas from the clumps orbiting near the sonic point to oscillate quasi - periodically with a frequency equal to one or two times the sonic - point beat frequency .
the time evolution that produces this oscillation is shown schematically in the six snapshots of figure [ fig : beatqpo ] .
for clarity , only a single clump is shown orbiting at the sonic radius and the six snapshots show the sequence of events as seen in a frame corotating with the star .
in panel ( a ) the clump is in the beam of radiation that rotates with the star and hence the inward flux of denser gas from the clump is greater than average . in panel
( b ) the clump has moved out of the beam and hence the inward flux of gas from the clump is smaller than average ; the pulse of gas that was dragged off the clump by the stronger radiation flux in frame ( a ) is now falling inward from the clump . in panel
( c ) the clump has moved ahead of the beam by 180 ; the inward flux of gas from the clump remains smaller than average and the pulse of denser gas that began to fall inward in frame ( a ) is further away from the clump . in panel
( d ) the clump is now ahead of the beam by 270 ; the inward flux of gas from the clump remains smaller than average and the pulse of denser gas that began to fall inward in frame ( a ) is approaching the stellar surface . in panel
( e ) the clump has now moved back into the beam of radiation and a new pulse of denser gas is leaving the clump ; the pulse of denser gas that began to fall inward in frame ( a ) is now very close to the stellar surface . in panel
( f ) the clump has again moved ahead of the beam by 90 and the inward flux of gas from the clump is again smaller than average ; the pulse of denser gas that began to fall inward in frame ( a ) is now colliding with the stellar surface , producing a brighter beam of radiation from the footprint of the stream .
in reality , the radiation drag force , and hence the inward mass flux from a clump , may be greatest not when the clump is fully illuminated by the beam that rotates with the star , as assumed in this illustration , but at some other relative phase , because the radiation drag force depends in a complicated way on the various components of the radiation stress - energy tensor ( see miller & lamb 1996 ) . in any case , the inward mass flux from the orbiting clumps oscillates quasi - periodically with a frequency approximately equal to the sonic - point beat frequency @xmath108 , producing a quasi - periodic oscillation in the luminosity and spectrum of the x - ray emission from the stellar surface .
_ this is the sonic - point beat - frequency qpo_. _ properties of the sonic - point qpos_.in 3.3 we show that the sonic - point mechanism naturally generates keplerian - frequency qpos with frequencies @xmath109 hz for a wide range of stellar magnetic fields and accretion rates and that it is natural for the frequencies of the sonic - point keplerian and beat - frequency qpos to increase steeply with increasing mass flux through the inner disk .
the angular distribution of the emission from each emitting point within a footprint at the stellar surface is broad , so the power at @xmath104 is likely to be much greater than the power at overtones of @xmath104 .
the modulation of the inflow from clumps orbiting at the sonic radius is unlikely to be perfectly sinusoidal , so overtones of the sonic - point beat frequency @xmath107 as well as @xmath107 itself may be detectable . in 3.4 we argue that the sonic - point mechanism can produce qpos with @xmath24 ratios as large as @xmath102 , i.e. , similar to those observed in the atoll and z sources .
the oscillation at the sonic - point beat frequency is generated by the beating of the stellar spin frequency , which is periodic , against the quasi - periodic sonic - point keplerian frequency oscillation .
it is therefore natural to expect the @xmath24 values of the oscillations in a pair to be roughly similar , in the absence of other disturbing effects . in 3.5
we show that the steep increase in qpo amplitude with photon energy observed in many kilohertz qpos is understandable in the sonic - point model .
the visibilities of the various qpos generated by the sonic - point mechanism depend on the mass flux in the keplerian disk flow that penetrates to the sonic point , the number and distribution of clumps in the disk at the sonic point , the brightness of the footprints produced by the streams from the clumps relative to the brightness of the rest of the stellar surface , the geometry and optical depth of the scattering material around the star , and the inclination of the system .
if radiation forces are required to produce large clumps with substantial density contrasts , the amplitudes of the qpos at the sonic - point keplerian and beat frequencies are likely to become much smaller and may become undetectable once the angular momentum loss radius @xmath110 has retreated inward to the radius @xmath98 of the innermost stable circular orbit , because once this has happened , the gas in the disk near the sonic point is increasingly shielded from radiation coming from the stellar surface by gas in the disk closer to the star . however , if large clumps with substantial density contrasts are formed by other mechanisms , the sonic - point keplerian frequency qpo may still be detectable if the sonic point is at @xmath94 , even if the sonic - point beat frequency qpo which can only be generated if radiation forces have a significant dynamical effect at the sonic radius has disappeared . in 3.6
we show that scattering by the electrons in the central corona that surrounds the neutron stars in the z and atoll sources strongly attenuates the already intrinsically weak beaming oscillation at the stellar spin frequency as well as the weak beaming oscillations at other frequencies , making all but the strongest beaming oscillation
the one at the sonic - point keplerian frequency difficult to detect with current instruments .
in contrast , scattering only weakly attenuates luminosity oscillations , so the luminosity oscillations at the sonic - point beat frequency and its overtones are likely to be detectable . taking into account both the generation of these various oscillations and their attenuation in the circumstellar environment , keplerian - frequency and beat - frequency qpos with rms amplitudes as large as 15% appear possible . as discussed in 2 , the z sources appear to have significantly stronger magnetic fields than the atoll sources .
the magnetic fields of the z sources are therefore likely to channel a larger fraction of the accreting gas out of the disk plane before it penetrates close to the stellar surface , so we expect the amplitudes of the kilohertz qpos to be smaller in the z sources than in the atoll sources
. the magnetic field of the neutron star in some atoll sources may be so weak that the qpo at the sonic - point beat frequency is undetectable , even though a qpo at the sonic - point keplerian frequency is visible .
the sonic - point beat frequency may also be undetectable at some times if the scattering optical depth between the stellar surface and the sonic point is too large .
in other sources or at other times , the luminosity oscillation at the sonic - point beat frequency may be detectable even though the beaming oscillation at the sonic - point keplerian frequency has been suppressed by scattering in the corona surrounding the star . in both the atoll and the z sources , we expect radiation drag to cause gas in the keplerian disk flow to make an abrupt transition from slow inward drift to rapid radial inflow several kilometers above the surface of the neutron star .
this transition occurs at the radius where the drag exerted by radiation from the star removes enough angular momentum from the gas quickly enough that it falls to the surface of the star unimpeded by a significant centrifugal barrier .
we first describe the consequences of the transfer of angular momentum from the gas in the inner disk to the radiation and then present a fully general relativistic calculation of the gas dynamics and radiation transport in the inner disk that shows the nature of this transition in the atoll sources , which have luminosities much less than the eddington critical luminosity .
we next give approximate analytical expressions for the location and width of this transition , then discuss why the nature and location of the transition may be similar in the z sources even though their accretion rates and luminosities are much higher than the accretion rates and luminosities of the atoll sources , and finally summarize the implications for the sonic - point qpos .
_ transfer of angular momentum to the radiation field_.the azimuthal drag force exerted by radiation from the star can create a radially supersonic inflow in the inner disk . however , this is possible only within a few stellar radii if the luminosity of the star is produced by accretion and most of the accreting gas flows through the inner disk .
the reason is that radiation drag can remove at most a fraction of the specific angular momentum of the accreting gas ( see fortner et al .
1989 ; for more detailed discussions , see miller & lamb 1993 , 1996 ) .
the specific angular momentum of gas in circular keplerian orbits at radii large compared to the radius @xmath54 of the neutron star is much larger than at @xmath54 , so radiation drag can not cause more than @xmath102030% of the gas at such large radii to fall to the stellar surface . in contrast , the specific angular momentum of gas in circular keplerian orbits very near the star is not much larger than the angular momentum of orbits at the stellar surface , so radiation drag can cause gas in keplerian orbits near the star to plunge to the surface .
radiation drag can create a radially supersonic inflow at a larger radius than one would estimate using the newtonian approximation , because special and general relativistic effects significantly increase the fraction of the angular momentum of the accreting gas that is transferred to the radiation ( miller & lamb 1993 ) . also , in general relativity the specific angular momentum of gas in circular keplerian orbits with radii varies only slowly with radius .
the latter point is illustrated by figure [ fig : angmomvsradius ] , which shows the specific angular momentum of gas in circular keplerian orbit around a nonrotating and a slowly rotating @xmath101 neutron star , as a function of the radius of the orbit .
the specific angular momentum varies fairly steeply at , but only slowly with radius between and .
for example , the specific angular momentum of circular orbits changes by less than @xmath111% from ( 25 km ) to ( 12 km ) .
these curves are of course physically meaningful only at radii greater than the radius @xmath54 of the neutron star .
the relatively flat shape of the angular momentum curve for circular orbits near a neutron star means that if gas orbiting there loses even @xmath86% of its angular momentum , it will plunge to the stellar surface .
we can estimate the outer boundary of the region where radiation coming from near the stellar surface can remove sufficient angular momentum from the gas orbiting in the disk so that it falls supersonically to the stellar surface by comparing @xmath112 , the largest fraction of angular momentum that can be removed by radiation coming from the surface of the neutron star , with @xmath113 , the fraction of the specific angular momentum of gas in a circular keplerian orbit at radius @xmath100 that must be removed in order for the gas to fall from @xmath100 to @xmath94 ( if the radius of the star is less than @xmath94 , gas that has reached @xmath94 can fall to the stellar surface without losing any more angular momentum ) .
these two quantities are compared in figure [ fig : angmomfractionvsradius ] .
here @xmath112 was computed by assuming that the gravitational binding energy at the stellar surface is converted into radiation there , taking into account the relevant angular , special relativistic , and general relativistic factors ( see miller & lamb 1996 ) , and assuming that all of the radiation emitted in the direction of the disk interacts with the disk flow ( @xmath114 in the notation of miller & lamb [ 1993 ] ) .
figure [ fig : angmomfractionvsradius ] shows that @xmath112 exceeds @xmath113 only inside ( ) . this characteristic radius depends on the radius of the neutron star , because the luminosity is proportional to the gravitational binding energy at the stellar surface , but this dependence is relatively weak ( the characteristic radius decreases from @xmath115 to @xmath116 as the stellar radius increases from @xmath117 to @xmath118 ) .
hence , _ the sonic - point model predicts that if radiation drag produces a transition to rapid radial inflow , the transition will occur within and hence that the keplerian frequency at the transition radius will be _ ( see 3.3 ) . determining the actual radius of the transition requires gas dynamical and radiation transport calculations .
_ model calculations_.the nature of the transition to supersonic radial inflow is illustrated by the following fully general relativistic calculations of the gas dynamics and radiation transport in the innermost part of the accretion disk flow . in these calculations
the azimuthal velocity of the gas in the disk is assumed to be nearly keplerian far from the star .
internal shear stress is assumed to create a constant inward radial velocity @xmath119 in the disk , as measured in the local static frame , of @xmath120 ( here and below we take @xmath119 to be positive for radially inward flow ) .
the half - height @xmath121 of the disk flow at radius @xmath100 is assumed to be @xmath122 at all radii , where @xmath123 is a constant and @xmath100 is the radius , and the kinetic energy of the gas that collides with the surface of the star is assumed to be converted to radiation that emerges from a band around the star s equator with a half - height equal to @xmath124 . for simplicity , and to show the effects of radiation forces more clearly ,
any effect of the stellar magnetic field on the dynamics of the disk flow near the sonic transition is neglected in this model calculation .
the surface density of the disk flow as measured by a local observer comoving with the gas is related to the surface density @xmath125 of the disk flow as measured by a local static observer by a lorentz boost , so @xmath126 where @xmath127^{-1/2}$ ] and we have neglected @xmath128 . for stationary disk accretion , @xmath129 at radius @xmath100 is given by ( see novikov & thorne 1973 , eq .
5.6.3 ) @xmath130 where @xmath131 is the mass accretion rate through the disk , as measured at infinity .
close to the star , the photon mean - free - path is limited primarily by thomson scattering and is therefore @xmath132 , where @xmath133 is the electron number density in the disk flow as measured in the frame comoving with the flow and @xmath134 @xmath135 is the thomson scattering cross section .
hence , for a steady flow , the radial optical depth from the stellar surface at radius @xmath54 through the disk to radius @xmath100 is @xmath136^{-1 } } { \gamma_r(r')\gamma(r')}\ , \frac{(1 - 2m / r')^{-1}}{r'h(r')v^{\hat r}(r ' ) } \,dr',\nonumber\\ \label{eq : radialopticaldepth } \end{aligned}\ ] ] where @xmath137 is the mass accretion rate through the inner disk , as measured at infinity .
the radial coordinate @xmath138 in the first expression on the right is the proper radial distance in the frame comoving with the gas . in the second expression on the right
we have estimated @xmath133 in the disk flow between @xmath100 and the stellar surface by assuming for simplicity that the gas in the disk is fully ionized hydrogen and using the continuity equation .
then @xmath139 , where @xmath140 is the proton mass and @xmath141 is the half - thickness of the disk flow .
the differential proper radial distance @xmath142 in the frame comoving with the gas is related to the differential radial distance in the local static frame by a lorentz boost , so @xmath143 , where @xmath119 is the inward radial velocity and @xmath144^{-1/2}$ ] .
the differential radial distance in the local static frame is in turn related to the differential radial distance @xmath145 in the global ( boyer - lindquist ) coordinate system by @xmath146 , so @xmath147 . once the drag force exerted by the radiation from the stellar surface begins to remove angular momentum from the gas in the keplerian disk , centrifugal support
is lost and the gas falls inward , accelerating rapidly .
radiation that comes from near the star and is scattered by the gas in the disk is generally scattered out of the disk plane and hence does not interact further with the gas in the disk . moreover ,
second and successive scatterings do not contribute proportionately to the azimuthal radiation drag force on the gas because the radiation field is aberrated by the first scattering and afterward carries angular momentum ( see miller & lamb 1993 , 1996 ) .
we therefore treat the interaction of the radiation with the gas in the disk by assuming that the intensity of the radiation coming from the star is attenuated as it passes through the gas in the disk , diminishing as @xmath148 , where @xmath149 is given by equation ( [ eq : radialopticaldepth ] ) , and that scattered radiation does not contribute to removal of angular momentum from the gas . in calculating the radiation drag force ,
we assume for simplicity that the differential scattering cross section is isotropic in the frame comoving with the accreting gas ( this gives results very close to those using a thomson differential cross section ; see lamb & miller 1995 ) . the radiation field and the motion of the gas
are computed in full general relativity .
the optical depth of the disk flow near the star is generally much less than one would estimate by calculating it without taking into account the radiation forces .
the reason is that the loss of centrifugal support caused by transfer of angular momentum to the radiation causes the inward radial velocity of the gas in the disk at radii less than the radius @xmath81 of the radiation - induced transition to supersonic inflow to be orders of magnitude higher , and the density of the gas to be orders of magnitude lower , than if radiation forces were neglected ( see miller & lamb 1996 ) .
hence the mean free path in the disk flow is much larger , and the optical depth from the stellar surface to a given radius in the flow is much smaller , than they would be in the absence of radiation drag . in this way
the radiation increases the transparency of the disk flow , so its effects are felt much farther out in the flow than one would estimate from the properties of the undisturbed flow .
figure [ fig : simpleexamples ] shows results obtained by solving self - consistently for the gas dynamics and the radiation field in this simple model , for a disk flow of semi - thickness @xmath150 , a nonrotating neutron star with gravitational mass and radius , and accretion rates through the inner disk of 0.02 , 0.03 , 0.04 , and @xmath151 , where @xmath152 is the mass accretion rate that produces the eddington critical luminosity at infinity ( see lamb & miller 1995 ) .
these accretion rates are typical of the lower - luminosity atoll sources .
the calculations were carried out using the numerical algorithm described in miller & lamb ( 1996 ) .
the results shown in figure [ fig : simpleexamples ] demonstrate that radiation forces and general relativistic effects create a sharp transition to supersonic inflow several kilometers above the stellar surface and that the transition radius decreases with increasing mass flux when the transition is caused by radiation drag .
these results are not intended to represent the inflow in any particular source . for and @xmath153 ,
the transition to rapid radial inflow occurs at @xmath154 and @xmath155 , respectively , and is caused by transfer of angular momentum from the gas to the radiation field . for and @xmath151 ,
most of the radiation from the star does not penetrate as far out as the radius @xmath156 of the innermost stable circular orbit , so the transition to rapid radial inflow occurs at @xmath94 and is caused by the absence at of circular orbits with the specific angular momentum of the orbit at .
the radial velocity profiles plotted in figure [ fig : simpleexamples]a show that at the sonic point the radial velocity typically increases by about two orders of magnitude in a very small radial distance .
the angular velocity profiles plotted in figure [ fig : simpleexamples]b show that the effect of radiation forces on the angular velocity is smaller , but is still very significant . for and @xmath151 , the angular velocities outside @xmath157
are nearly keplerian , because the gas there is shielded from the radiation by the gas further in .
however , for and @xmath153 , the angular velocity drops below the keplerian value well outside @xmath157 .
it is this departure from keplerian orbital motion , and the associated loss of centrifugal support , that causes the flow to accelerate radially inward .
all the angular velocity profiles turn downward near the stellar surface because of the strong azimuthal radiation drag force there .
the decrease in the kinetic energy and angular momentum of the inflowing gas is exactly compensated by the increase in the energy and angular momentum carried outward by the escaping radiation ( see miller & lamb 1993 , 1996 ) .
figure [ fig : simpleexamples]c shows that the vertical column density of all four disk flows drops abruptly from @xmath158 outside the sonic radius to @xmath15910 inside . the radial optical depth measured outward from the stellar surface
is shown in figure [ fig : simpleexamples]d ; it decreases by about two orders of magnitude at the sonic radius , but not as sharply as the vertical optical depth .
even when the transition to rapid radial inflow occurs at @xmath94 , the flow between @xmath94 and the stellar surface is generally strongly affected by radiation forces , as shown by the differences between the velocity and optical depth profiles for and @xmath151 .
we caution that these results are only illustrative .
for example , x - rays may be emitted from a larger fraction of the stellar surface than we have assumed , in which case @xmath81 will be larger for a given accretion rate
. _ radius of the transition_.in the model of gas dynamics and radiation transport just described , under conditions such that loss of angular momentum to radiation drag is important outside @xmath94 , loss of centrifugal support and rapid radial inflow begins about five photon mean - free - paths from the stellar surface , i.e. , the transition occurs where . in this model
the inward radial velocity of the gas in the disk inside @xmath81 typically rises very sharply to @xmath160 and then changes more slowly until the gas collides with the stellar surface ( see also miller & lamb 1993 , 1996 ) .
hence we can estimate @xmath81 by setting in equation ( [ eq : radialopticaldepth ] ) , scaling @xmath119 and @xmath161 in units of @xmath162 and @xmath163 , respectively , and solving for the radius .
the result is @xmath164 this estimate is in rough agreement with the transition radius found in our numerical calculations . _
width of the transition_.under conditions such that the transition to rapid radial inflow is caused by radiation drag , the width @xmath165 of the radial velocity transition is determined by @xmath166 , the photon mean - free - path in the keplerian disk flow at @xmath81 , because this is the characteristic distance over which the gas first becomes exposed to radiation from the star , and by the radial distance @xmath167 over which the gas in the disk loses its angular momentum once it is exposed to radiation from the star . from expressions similar to those used above to compute the radial optical depth in the keplerian disk flow ,
we find @xmath168 where we have scaled the quantities that enter this expression in terms of their approximate values in the transition region .
this estimate is consistent with the results of our numerical calculations . from the expression for the radiation drag rate given by miller & lamb ( 1993 , 1996 ) ,
we find @xmath169 where we have scaled @xmath170 in terms of the typical value for a gas - pressure - dominated inner disk .
these estimates show that the radial velocity is expected to increase over a very small radial distance @xmath171 @xmath172 .
this expectation is in accord with the results of our numerical calculations presented above . as noted in 3.1 ,
this transition is somewhat analogous to the transition at the boundary of a strmgren sphere , except that here the radiation removes angular momentum from the gas rather than stripping electrons from atoms or molecules and causes a sharp increase in the inward radial velocity rather than in the degree of ionization .
if the neutron star radius is smaller than @xmath94 and conditions are such that the transition to rapid radial inflow is caused by general relativistic corrections to newtonian gravity , the radial velocity increases rapidly near @xmath94 because gas inside @xmath98 can fall to the stellar surface without losing any angular momentum .
the precise location of the sonic transition is slightly affected by pressure forces and , if the shear stress is large , by outward angular momentum transport within the disk flow ( see muchotrzeb 1983 ; muchotrzeb - czerny 1986 ) . in the absence of shear
stresses and pressure forces , the radial velocity @xmath119 of an element of gas at radius @xmath173 , where , that has fallen inward from a circular orbit at @xmath94 is approximately @xmath174 ( miller & lamb 1996 ) .
hence , for @xmath175 , the inflow becomes supersonic over a radial distance @xmath176 .
_ summary and implications for kilohertz qpos_.for sources in which the radius @xmath54 of the neutron star is larger than the radius @xmath94 of the innermost stable circular orbit and accretion is under conditions such that angular momentum loss to the radiation field terminates the keplerian flow at a radius larger than @xmath54 , our results show that the keplerian disk flow near the star is terminated by angular momentum loss to radiation and that the radius of the transition to rapid radial inflow decreases steeply with increasing mass flux through the inner disk .
hence the orbital frequency at the sonic radius , and therefore the frequency of the associated qpo , increases steeply with increasing mass flux .
if instead conditions are such that radiation forces do not terminate the keplerian flow , the keplerian flow is probably terminated by interaction with the stellar surface , in which case generation of a coherent qpo with a frequency equal to the orbital frequency at the stellar surface appears very unlikely
. for sources in which @xmath54 is smaller than @xmath98 and accretion occurs under conditions such that radiation forces terminate the flow outside @xmath94 , our results show that the keplerian flow is terminated by angular momentum loss to radiation and that the radius of the transition to rapid radial inflow decreases steeply with increasing mass flux . hence the frequency of the qpo associated with the orbital frequency at the sonic radius again increases steeply with increasing mass flux . if instead conditions are such that radiation forces do not terminate the keplerian flow outside @xmath94 , the keplerian flow is terminated by general relativistic corrections to newtonian gravity and the radius of the transition to rapid radial inflow is approximately independent of the mass flux .
hence , under these conditions the frequency of any qpo with the orbital frequency at the sonic point will be approximately independent of the accretion rate .
although the numerical results presented here are for accretion rates typical of the atoll sources , which have very weak magnetic fields and are accreting at rates much less than the eddington critical rate , we expect the structure of the accretion flow and the behavior of the sonic - point keplerian qpo frequency to be similar in the z sources , even though they are thought to have stronger magnetic fields and are accreting at much higher rates .
there are three reasons .
( 1 ) the radius of the transition to rapid radial inflow that we are considering is necessarily close to the neutron star whatever the luminosity , because the fraction of the angular momentum in the keplerian disk flow that can be removed by radiation drag is small and general relativistic corrections to newtonian gravity are important only near the neutron star .
( although the luminosities of the z sources are much higher , the angular momentum flux that must be removed to create a radial inflow is correspondingly higher . in sources with luminosities very close to the eddington critical luminosity
, radiation drag can significantly affect the motion of up to of the accreting gas further away , but can only affect the motion of most of the accreting gas close to the star [ see fortner et al .
1989 ; lamb 1989 ] . )
( 2 ) the inferred magnetic fields of the neutron stars in the z sources are stronger than those of the neutron stars in the atoll sources , and hence a larger fraction of the disk flow is likely to be channeled out of the disk by the stellar magnetic field , causing the mass flux through the inner disks of the z sources to be comparable to the mass fluxes through the inner disks of the atoll sources .
( 3 ) the vertical thickness of the inner disk flow is likely to be larger in the z sources than in the atoll sources .
all of these effects act in the direction of making the radius of angular momentum loss in the z sources similar to its value in the atoll sources .
we do expect significant departures from the behavior found in the present calculations when the luminosity equals or exceeds the eddington critical luminosity .
these points are discussed in more detail in 4 .
we consider now the frequencies that are generated by the sonic - point mechanism .
there are two fundamental qpo frequencies : ( 1 ) the keplerian frequency at or near the sonic radius and ( 2 ) the sonic - point beat frequency , which is generated by interaction of radiation from the neutron star with the accretion flow near the sonic radius .
qpos may also be detectable at overtones of the keplerian and beat frequencies .
qpos at sideband frequencies and oscillations at the spin frequency of the neutron star and its overtones are likely to be very weak . in analyzing the frequencies generated by the sonic - point mechanism , we shall for definiteness use boyer - lindquist global coordinates , which are familiar because they are commonly used to describe the spacetime of rotating black holes .
however , all of the expressions we give are accurate only to first order in @xmath90 and hence would be unchanged if we used circumferential rather than boyer - lindquist coordinates . in boyer - lindquist coordinates ,
the circular frequency of an element of matter on the surface of the rotating star or of an element of gas in orbit around it is @xmath177 where @xmath178 and @xmath179 are the azimuthal and time coordinates of the element of matter or gas .
when measured in boyer - lindquist coordinates , any rotational frequency appears to be the same at every point in space and is equal to the frequency measured by a distant observer because of the symmetries of the spacetime outside a steadily rotating star and @xmath180 , which reflect the stationarity and axial symmetry of the spacetime ( see , e.g. , misner , thorne , & wheeler 1973 , pp .
892895 ) . in any coordinate system in which the time and azimuthal coordinates are based on @xmath181 and @xmath180 , such as the boyer - lindquist coordinate system ,
the time interval required for one rotation of an element of the star or one orbit of an element of gas orbiting the star is the same everywhere , when measured in the global time coordinate .
stated more concretely , if an element of gas emits a pulse of x - rays each time its azimuthal coordinate @xmath178 increases by @xmath182 , the time interval @xmath183 between the arrival of successive pulses will be the same everywhere , when measured in the global time coordinate @xmath179 . ] and the fact that the boyer - lindquist time coordinate is the proper time of an observer at radial infinity .
we emphasize that the frequency ( [ eq : circfreq ] ) is _ not _ the frequency that would be measured by a local observer ( i.e. , in a local orthonormal tetrad ) , unless the observer is at infinity . _
sonic - point keplerian frequency_.as explained earlier in this section , the radiation drag force acting on the gas in a clump near the sonic radius generates a supersonic stream of denser gas that spirals inward toward the stellar surface .
to the extent that the radiation field is axisymmetric , the inspiral trajectory of the gas from each such clump is identical and hence the azimuthal separation @xmath184 between any two such streams is the same at every radius inside the sonic point and equal to the azimuthal separation @xmath185 of the two clumps at the sonic radius @xmath186 that are producing the two streams .
therefore a collection of clumps distributed around the star at the sonic radius will generate a pattern of inspiral streams that rotates around the star at the orbital frequency @xmath187 of the gas at the sonic radius .
we can estimate @xmath104 by noting that , to first order in @xmath90 , the the spacetime outside a slowly and uniformly rotating star is the same as the kerr spacetime ( hartle & thorne 1968 ; the two spacetimes are not the same to higher orders in @xmath90 ) . hence , to first order in @xmath90 the orbital frequency of an element of gas in a prograde keplerian circular orbit at boyer - lindquist radius @xmath188 in the rotation equator of a slowly rotating star is ( see bardeen , press , & teukolsky 1972 ) @xmath189 \left ( \frac{m}{r_{\rm s}^3 } \right)^{\!1/2 } \nonumber\\ & \approx & 1181\,\left(m\over{1.4\,\msun}\right)^{\!1/2 } \left(r_{\rm s}\over\mbox{15\ km}\right)^{\!\!-3/2 } \,{\rm hz}. \label{eq : sonickeplerfreq } \end{aligned}\ ] ]
the _ pattern _ of the gas streaming inward from the clumps at @xmath188 rotates uniformly around the star with frequency @xmath104 .
the pattern frequency is different from the _ orbital _ frequency @xmath190 of the gas inside the sonic radius , which varies with radius . in particular ,
the pattern frequency is different from the orbital frequency of the gas just before it collides with the stellar surface .
the brightness pattern made by the collision of the inspiraling streams of denser gas with the stellar surface rotates around the star at frequency @xmath104 , as measured by an observer at infinity .
this is therefore the frequency at which the radiation pattern produced by the bright footprints of the streams rotates around the star and hence also the centroid frequency of the resulting quasi - periodic oscillation in the x - ray flux and spectrum seen by a distant observer whose line of sight is inclined to the rotation axis of the disk
. therefore the radiation pattern produced by the bright footprints of the streams generally does _ not _ corotate with the star , but instead moves around the stellar surface .
_ stellar spin frequency_.as explained in 3.1 ,
if the star s magnetic field is too weak to affect the accretion flow at the sonic point but is strong enough to partially channel the flow close to the stellar surface , it will create a weakly beamed pattern of radiation that rotates _ with the star_. the surface magnetic fields of the neutron stars in the atoll and z sources , which are thought to be several hundred million years old , may be nearly dipolar .
if the magnetic field is dipolar but offset from the center of the star , the radiation pattern it produces will have two unequal maxima around the star , whereas if the field is a centered but tilted dipole field , the radiation pattern will have two nearly equal maxima around the star .
we therefore expect that there will generally be a brightness oscillation with a frequency equal to one or two times the spin frequency @xmath83 of the star . however , as we discuss in detail in 4.3 , we expect the relatively weak magnetic fields of the atoll and z sources to produce radiation patterns that are only weakly beamed , even near the star .
when the substantial attenuation of this inherently weak beaming that is caused by scattering in the central corona is taken into account , the x - ray oscillation at @xmath83 or @xmath191 seen by a distant observer may be very weak or even undetectable with current instruments .
_ sonic - point beat frequency_.the weakly beamed pattern of radiation rotating with the star causes the radiation drag force acting on the gas in a given clump orbiting near the sonic radius to peak with a frequency equal to one or two times the difference ( beat ) frequency @xmath107 between the keplerian frequency at the sonic point and the spin frequency of the star , as explained in 3.1 .
the peak in the drag force will cause a temporary increase in the flux of denser gas streaming supersonically inward from the clump with a frequency @xmath192 where @xmath193=1 or 2 , depending on the symmetry ( see lamb et al .
the inflow time from the sonic point is typically 510 beat periods , so there are typically this many density enhancements in the stream of gas from a given clump to the stellar surface .
these density enhancements do not lie along a single gas streamline , but instead along a sequence of streamlines .
as they collide with the stellar surface they produce a sequence of brighter impact arcs around the surface in the plane of the disk . as the quasi - periodic increase in the mass flux from a given clump arrives at the stellar surface , it produces a quasi - periodic modulation in the luminosity of the star with frequency @xmath194 . because this frequency is the difference between an orbital and a spin frequency ,
the frequency of the resulting quasi - periodic modulation of the x - ray flux and spectrum seen by an observer at infinity is @xmath194 . _
expected range of sonic - point qpo frequencies._in
3.2 we showed that either radiation forces or general relativistic corrections to newtonian gravity cause an abrupt transition to rapid radial inflow and that the radius of this transition is bounded below by the radius of the marginally stable orbit , which is for a nonrotating @xmath101 star , and above by the upper bound on the fraction of the angular momentum of the accreting gas that can be transferred to the radiation field , which constrains the transition radius to be for a @xmath101 star .
as discussed in 4.2 ,
there are several effects that are likely to reduce further the allowed range of the transition radius , perhaps to .
inserting these radii in equation ( [ eq : sonickeplerfreq ] ) gives an expected range for the sonic - point keplerian frequency of . for stars with spin frequencies ,
the expected range of beat frequencies is then .
these ranges are similar to the frequencies of the kilohertz qpos detected so far .
_ overtones and sidebands of the sonic - point qpo frequencies._in addition to the beaming oscillations at @xmath104 and @xmath195 and the luminosity oscillation at @xmath107 , we expect accretion onto weakly magnetic neutron stars to generate brightness oscillations at several other frequencies . for example , the x - ray brightness modulation produced by the motion of the brighter impact footprints around the star with frequency @xmath104 is unlikely to be _ exactly _ sinusoidal , so there may be some power at overtones of @xmath104 . however , the angular distribution of the x - ray emission from the impact footprints is expected to be very broad ( see 3.1 ) , so the power at overtones of @xmath104 is likely to be much smaller than the power at @xmath104 . the radiation pattern that rotates with the star is also expected to be very broad by the time the radiation escapes from the central corona , so the power at overtones of @xmath83 is likely to be much smaller than the power at @xmath83 , which is itself expected to be small for the reasons discussed above .
however , the radiation pattern that rotates with the star is likely to be narrower at the sonic radius than outside the central corona , causing the modulation of the mass flux from the sonic point to be somewhat non - sinusoidal and giving rise to weak overtones of the sonic - point beat frequency .
luminosity oscillations are not as strongly attenuated by scattering in the central corona ( see 4.1 ) , so it may be easier to detect overtones of @xmath107 .
the sonic - point mechanism causes a modulation of the inward mass flux and hence the stellar luminosity at the sonic - point beat frequency ; from the point of view of a distant observer , this luminosity is then modulated at the sonic - point keplerian frequency by the motion of the brighter footprints around the star . as a result ,
qpos are generated at some sideband frequencies .
however , by far the strongest qpos are generated at the sonic - point keplerian and beat frequencies . to see this ,
consider as an example the x - ray flux waveform @xmath196 \times\nonumber\\&&\quad\times\ ; [ 1 + b_1\cos ( 2\pi\nu_{\rm b}t + \phi_{\rm b1})\nonumber\\ & & \quad\qquad \mbox { } + b_2\cos ( 4\pi\nu_{\rm b}t + \phi_{\rm b2 } ) ] \ ; , \label{fluxwave } \end{aligned}\ ] ] which describes these modulations . in equation ( [ fluxwave ] ) we have included for the sake of illustration only the fundamental and first overtones of the keplerian and beat frequencies ; higher harmonics may of course also be present .
the frequencies that are generated are listed in table [ table : freqgen ] .
the only first - order qpos are those at the sonic - point beat and keplerian frequencies .
there are also qpos at the second harmonics of the beat and keplerian frequencies .
the lowest - order qpos generated by the modulation of the beat - frequency `` carrier '' by the motion of the footprints around the star are of second order and are at the stellar spin frequency @xmath195 and at the very high frequency @xmath197 .
the only qpo of intermediate frequency generated by this modulation is a qpo at @xmath198 that is third - order and hence likely to be very weak .
all the other qpos generated are also third - order and in addition have very high frequencies . they are therefore likely to be intrinsically weak , strongly attenuated by scattering in the central corona , and difficult to detect . as we discuss in
4.3 , only brightness oscillations that are fairly large near the neutron star and are not strongly attenuated by the gas surrounding the star will be detectable by a distant observer . in the sonic - point model ,
only the sonic - point beat frequency qpo and its overtones are luminosity oscillations ; the other qpos are beaming oscillations and are therefore more attenuated by scattering .
for this reason , we expect strong qpos only at @xmath107 and @xmath104
. nonetheless , qpos at other frequencies , such as @xmath83 , @xmath199 , and @xmath200 may be detectable .
detection of oscillations at any of these frequencies would corroborate the sonic - point model . a key question
any theory of the kilohertz qpo sources that relates a qpo frequency to orbital motion must address is why narrow peaks are seen in power spectra of the brightness variations of these sources , rather than a broad continuum corresponding to the range of orbital frequencies in the inner disk .
indeed , the remarkably high coherence ( @xmath201 ) of some of the kilohertz qpos places very strong constraints on any model of these qpos , because there are many physical effects that tend to decrease the coherence of oscillations at kilohertz frequencies .
we discuss some of the most important effects , derive the resulting constraints on the accretion flow , and show that the sonic - point mechanism can produce narrow peaks consistent with the observed coherence of the kilohertz qpos .
gas in the inner disk orbits the neutron star with a wide range of frequencies , so at first glance one might think that fluctuations in the gas density throughout the inner disk would generate a broad spectrum of brightness variations .
for the reasons discussed in 2.3 , it is very unlikely that such fluctuations can produce large amplitude brightness variations via direct emission or occultation .
we have therefore focused our attention on the brightness variations that such density fluctuations produce indirectly , as gas from them falls inward and impacts the stellar surface .
here we show that the accretion flow in the vicinity of the sonic radius acts as a filter that _ selects _ brightness oscillations at the sonic - point orbital frequency while _ suppressing _ brightness oscillations at higher and lower frequencies . in order to see how this filtering occurs
, we analyze the x - ray emission at the stellar surface produced by clumps that form at different radii in the inner accretion disk .
_ effects of azimuthal shear_.let us suppose for the sake of argument that the thickness of the inner disk is infinitesimal and that the flow there is laminar ( we discuss the effects of turbulence shortly ) .
suppose further that a very small , roughly spherical density fluctuation ( clump ) has formed well outside the sonic radius and consider what will happen to the gas in this clump and how it will affect x - ray emission from the star .
when a small clump forms , the orbital phases of the elements of gas that comprise it are necessarily very similar , because of the small azimuthal extent of the clump .
the elements of gas that comprise the clump also have very similar orbital frequencies , because the radial extent of the clump is small . as the clump drifts inward , elements of gas at different radii are sheared relative to one another in the azimuthal direction .
the frequency of the x - ray brightness oscillation that is generated by the inspiraling gas from a clump is approximately equal to the keplerian frequency at the radius where the clump originally formed ( see the discussion in
3.3 ) . in the inner disk
to the stellar surface is large , azimuthal shearing of a clump formed at @xmath100 would by itself reduce the amplitude of the brightness variation produced by gas from the clump , but this is not so .
the reason is that to first order , the azimuthal velocity shear in a keplerian flow conserves the volume of the clump , because the velocity field is divergence - free to this order .
hence , the density of the gas in a clump relative to the density of the background gas is affected only weakly by the differential rotation .
in fact , if the flow were incompressible as well as laminar , the angular width of the brighter impact arc on the stellar surface would also be unaffected , even if the _
angular extent of the clump at all radii becomes large , because the angular extent of that part of the clump that is crossing the sonic radius at any given time is unchanged by azimuthal shear .
] a clump that forms very near the sonic radius will therefore produce an x - ray brightness oscillation at the sonic - point keplerian frequency .
if clumps can form inside the sonic radius , such clumps would produce x - ray brightness oscillations at the orbital frequencies where they formed .
this analysis shows that if the clumps were small and the flow laminar , differential rotation of the gas would not by itself pick out any particular orbital frequency .
consequently , if clumps were to form at a wide range of radii and nothing besides inward drift and azimuthal shear were to happen , gas inspiraling from the clumps and colliding with the stellar surface would produce brightness fluctuations with a wide range of frequencies up to the orbital frequency at the stellar surface ( see , e.g. , kluzniak , michelson , & wagoner 1990 ) . _
effects of clump destruction_.in reality , clumps that form _ outside _ the sonic radius are destroyed before they drift inward to the sonic point and gas in them can reach the stellar surface .
clumps ( if any ) that form _ inside _ the sonic radius orbit the star at a rapidly changing frequency and collide with the stellar surface before gas from them can produce a long wavetrain .
hence , only clumps that form very near the sonic radius can produce strong , relatively coherent brightness oscillations . in this way the disk flow filters out brightness oscillations at orbital frequencies other than the sonic - point keplerian frequency . in order to illustrate the effects of clump destruction in the keplerian flow
, we consider a simple model in which clumps form and are then destroyed by small - scale turbulence in the disk , such as that produced by the magnetoturbulence that is thought to be responsible for angular momentum transport within the disk flow ( balbus & hawley 1991 , 1992 , 1997 ; brandenburg et al.1995 , 1996 ; hawley , gammie , & balbus 1995 , 1996 ) . in this simple model
we assume that the forces that formed the clump in the first place no longer act to hold it together .
if the clump has size @xmath202 , turbulent motions on scales @xmath203 will tend to advect the clump without destroying it ; only turbulent motions on scales @xmath204 will disrupt the clump .
we assume that the effect of these small - scale turbulent motions on the clump can be described by a diffusion coefficient , where @xmath205 is a dimensionless parameter that describes the strength of the small - scale turbulence and @xmath170 is the thermal sound speed in the disk ; for turbulent motions with scales @xmath206 , @xmath205 is small compared to unity . as a result of outward angular momentum transport by the magnetoturbulence , a clump of initial size @xmath202 that forms at radius @xmath100 will drift slowly inward with radial velocity @xmath207 , where @xmath208 is the usual viscosity parameter and @xmath141 is the half - thickness of the disk .
such clumps will be dissipated by turbulent diffusion in a time @xmath209 . in this model , a clump that forms a radial distance @xmath210 _ outside
_ the sonic radius @xmath188 can reach the sonic radius and generate a supersonically inspiraling stream of gas before being destroyed only if @xmath211 , i.e. , only if the clump forms within a radial distance @xmath212 of the sonic radius . turbulence on small scales may be weak , and the effects that form the clumps in the first place such as thermal instability , magnetic stresses , and radiation forces may tend to continue to hold them together as they drift inward toward the sonic point , so that clumps can survive somewhat longer than @xmath213 .
even so , it is clear that clumps that form even a small distance outside the sonic radius will be disrupted by turbulence before they reach the sonic point .
this process suppresses brightness oscillations with frequencies less than the sonic - point keplerian frequency .
it appears improbable that clumps will form in the hypersonic radial inflow between the sonic radius and the stellar surface , but even if clumps do form in this region , they will not produce appreciable brightness oscillations , for two reasons .
first , a clump that forms inside the sonic radius will collide with the stellar surface after completing @xmath10110 orbits . a clump that lives only for a time @xmath214 will generate power over a range of frequencies @xmath215 .
hence the power produced by a clump that forms inside the sonic radius will be spread over a frequency range @xmath216 .
second , a clump that forms inside the sonic radius orbits the star at a frequency that rapidly increases from , for example , the @xmath217 orbital frequency at the sonic radius to the @xmath218 orbital frequency at the stellar surface , as the clump spirals inward . for these reasons ,
even if clumps do form inside the sonic radius , any power they generate will be spread over a wide range of frequencies rather than concentrated in a narrow peak .
_ coherence of the sonic - point keplerian frequency qpo_.clumps that form outside the sonic radius but close enough to it to reach it before being destroyed have initial orbital frequencies in the relatively narrow range @xmath219 which is consistent with the observed narrowness of the kilohertz qpo peaks .
clumps orbiting at different distances above or below the midplane of the disk will generally also have slightly different orbital frequencies .
to estimate the resulting spread in frequencies , we assume that at each distance @xmath220 above and below the disk plane the inward radial velocity increases sharply from subsonic to supersonic at some cylindrical radius @xmath221 .
we call the axisymmetric two - dimensional surface defined by @xmath222 the sonic surface .
if the sonic surface is cylindrical , the oscillation frequency generated by the clumps orbiting in the midplane of the disk is slightly greater than the oscillation frequency generated by the clumps orbiting above and below the midplane .
the spread in oscillation frequencies depends on the precise shape of the sonic surface , but to lowest order in @xmath223 the spread in frequencies caused by this effect is @xmath224 , which is consistent with the observed narrowness of the kilohertz qpo peaks if @xmath225 .
the persistence of a clump orbiting at the sonic radius is limited by its decay as gas inspirals from it to the stellar surface and by its disruption by small - scale turbulence , pressure forces , and other effects . in the simple model considered here ,
the clump decay timescale is @xmath226 , which broadens the qpo peak by an amount @xmath227 , which is consistent with the observed narrowness of the kilohertz qpo peaks .
the dissipation of clumps by turbulent diffusion contributes a relative width @xmath228 , which is consistent with the observed narrowness of the keplerian frequency qpo peaks if @xmath229 or the effects that form the clumps in the first place , such as thermal instability , magnetic stresses , and radiation forces , continue to hold them together as they orbit at the sonic point .
this simple model shows that if a clump is too small , it will persist for such a short time that the power it generates will be spread over a broad frequency range , whereas if a clump is too large it will persist for such a long time that its orbital frequency will change appreciably during its lifetime and the power it generates will again be spread over a broad frequency range .
it is the clumps of intermediate size that generate relatively coherent oscillations .
in addition to frequency variation and lifetime broadening , other effects may increase the width of the keplerian frequency qpo peak .
for example , the radiation field inside the sonic radius is not perfectly axisymmetric , and hence the radiation forces acting on the inspiraling gas will vary slightly with azimuth . as a result ,
the radius of the sonic point and the density patterns produced by the inspiraling gas will be slightly different at different azimuths . a quantitative treatment of these effects is well beyond what is possible with current accretion disk models .
the simple model we have discussed shows the crucial role played by the sharp transition to rapid radial inflow at the sonic radius that was found in 3.2 .
in the absence of such a transition , density and magnetic field fluctuations in the inner disk would produce weak brightness variations over a wide range of frequencies rather than in the narrow range of frequencies needed to create a qpo peak . in the presence of such a transition , on the other hand , clumps orbiting in a narrow range of radii near the sonic radius produce strong x - ray brightness variations with a correspondingly narrow range of frequencies .
simple estimates of the broadening produced by the different orbital frequencies of these clumps and by their finite lifetimes appear consistent with the high observed coherence of the keplerian frequency qpos .
_ coherence of related qpos_.if there is a qpo at the beat frequency @xmath107 , we expect that its fwhm will in general be comparable to the fwhm of the qpo at @xmath104 . this is because the peak at @xmath107 is at the beat frequency of the sonic - point keplerian frequency with the stellar spin frequency , which is nearly coherent and therefore adds relatively little to the width of the qpo peak .
however , we do _ not _ expect the widths of the qpo peaks at @xmath107 and @xmath104 to be identical , because scattering by the moving gas in the central corona affects the oscillations at @xmath107 and @xmath104 differently .
the x - ray spectra of systems powered by accreting , weakly magnetic neutron stars , such as the kilohertz qpo sources , are formed primarily by comptonization , as described in 2.2 .
cyclotron and bremsstrahlung photons with energies 1 kev are produced near the neutron star surface and are then comptonized by hot electrons in the region where the accretion flow interacts with the stellar surface and in the hot central corona that surrounds the star and its magnetosphere , yielding the observed x - ray spectra . during a sonic - point keplerian or beat - frequency oscillation ,
the rotation of the patterns of denser gas spiraling inward from clumps orbiting at the sonic radius causes the optical depth along the line of sight from the neutron star surface to the observer to vary quasi - periodically .
the rate of production of soft photons and the physical properties of the comptonizing gas may also oscillate , especially during the sonic - point beat - frequency oscillation , which is mainly a luminosity ( accretion rate ) oscillation .
these oscillations cause the x - ray spectrum produced by the system to oscillate quasi - periodically with various frequencies .
a small fractional variation in the optical depth causes a much larger fractional variation in the number of photons at high photon energies , because of the characteristic way in which the spectrum produced by comptonization changes as the optical depth oscillates ( see miller & lamb 1992 ) . to the extent that the effects of comptonization dominate , the oscillations at @xmath230 will lag the oscillations at @xmath231 , because the photons above 10 kev have scattered more times in escaping from the source .
an accurate , quantitative treatment of the x - ray spectral oscillations produced by the inhomogeneous and time - dependent accretion flow in the kilohertz qpo sources would require a three - dimensional , time - dependent radiation hydrodynamic calculation in full general relativity .
this is beyond present computational abilities .
however , one can obtain a qualitative understanding of the dependence of qpo amplitudes on photon energy and the time lags to be expected using the following simplified model of the comptonization process .
assume that soft photons with a characteristic energy @xmath232 that is much smaller than the electron temperature @xmath45 are injected at the center of a static , uniform , spherical comptonizing region of radius @xmath233 , electron density @xmath133 , and scattering optical depth @xmath234 , where @xmath235 is the thomson scattering cross section .
assume further that @xmath236 , where @xmath46 is the electron rest mass , and that the electrons have a negligible bulk velocity .
these are good approximations for modeling formation of the time - averaged x - ray spectra of the atoll sources ( see psaltis & lamb 1998b ) .
the region of approximately radial inflow that develops when the mass accretion rate becomes comparable to @xmath237 ( see lamb 1989 , lamb 1991 , and
2 ) introduces additional effects on the spectra of the z sources that can not be treated in this way ; we do not consider these effects here ( but see psaltis & lamb 1998b ) .
we treat the effects of oscillations in the rate of soft - photon production and in the properties of the comptonizing region near the star by varying these quantities in the spherical model . in particular , we mimic the effects of the quasi - periodic variation of the optical depth along the line of sight by varying the optical depth of the model .
we assume that the oscillation period is longer than the mean time for photons to escape from the comptonizing region , an approximation that appears to be excellent for the kilohertz qpo sources .
_ photon - energy dependence of the sonic - point keplerian and beat - frequency oscillation amplitudes_.the x - ray number spectrum that emerges from the comptonizing region at photon energies @xmath238 can be approximated by ( see , e.g. , rybicki & lightman 1979 , pp .
221 - 222 ) @xmath239 , \;\ ; e\gg e_b \end{array } \right .
\label{eq : spectrum}\ ] ] where @xmath45 is the electron temperature , @xmath240 is a normalization constant that is typically a few kev , @xmath241 is the cutoff energy , and @xmath242 for small - amplitude oscillations in which the variation of the electron temperature in the comptonizing region is negligible ( see below ) , the relative amplitude @xmath243 of the oscillation is approximately ( miller & lamb 1992 ) @xmath244 \,\delta \tau\;. \label{eq : ampgeneral}\ ] ] here @xmath245 is the relative change in the rate at which soft photons are injected into the comptonizing region during an oscillation and @xmath246 is the change in the optical depth of the region . the relative amplitude @xmath243 is computed by considering a full oscillation period .
in general , the soft photon injection rate and the optical depth both vary during the oscillation .
if so , equations ( [ eq : spectrum])([eq : ampgeneral ] ) give @xmath247 where @xmath248 \left\ { \begin{array}{ll } \frac{\ln(e / e_a)}{\ln(e_b / e_a)}\ ; , & e\ll e_b\\ \\ \quad\ ; 1 \ ; , & e\gg e_b \end{array } \right .
\label{eq : gammatau } \end{aligned}\ ] ] in terms of the expected properties of the comptonizing region and we have neglected the weak dependence of the normalization constant @xmath240 on @xmath47 .
equations ( [ eq : ampabsgeneral ] ) and ( [ eq : gammatau ] ) show that in this case the relative amplitude of the qpo depends both on the relative change in the soft - photon injection rate , @xmath249 , and on the relative change in the optical depth , @xmath250 . in writing equation ( [ eq : ampgeneral ] )
we have assumed that the electron temperature in the comptonizing cloud remains constant during the oscillation .
if the electron temperature were allowed to vary , then the relative amplitude of the oscillation would not be constant at high photon energies , in disagreement with observations .
the quantity @xmath251 is negative for the spectrum ( [ eq : spectrum ] ) and is small at low photon energies .
if the oscillation in the injection rate of the soft photons is in phase with the oscillation in the optical depth , i.e. , if @xmath249 and @xmath252 have the same sign , and if @xmath253 is larger in magnitude than @xmath254 at high photon energies , then the relative amplitude of the oscillation in the photon number will not be a monotonic function of photon energy . for @xmath249 comparable in magnitude to @xmath250 and a comptonizing region with the properties used in scaling equation ( [ eq : gammatau ] ) , the relative amplitude of the oscillation has a minimum in the energy range @xmath255 .
the sonic - point beat frequency qpo is primarily a luminosity oscillation ( see 3.1 ) , so the photon production rate is likely to vary significantly at the beat frequency . _
the relative amplitude of the beat - frequency qpo may therefore have a minimum in the @xmath256 energy range_. if the relative change in the photon injection rate during an oscillation is very small , the first term on the right side of equation ( [ eq : ampabsgeneral ] ) is likely to be negligible compared to the second term . in this case
the relative amplitude of the oscillation will increase monotonically with increasing photon energy , for energies less than @xmath257 kev , but will become independent of photon energy for energies greater than @xmath258 .
we expect this to be the case for the oscillation at the sonic - point keplerian frequency , which is mainly a beaming oscillation ( see 3.1 ) and hence involves a relatively small variation in the photon production rate .
_ we therefore expect the relative amplitude of the keplerian - frequency qpo to increase steeply in the @xmath256 energy range and then flatten at high energies_. when the photon injection rate and spectrum of injected soft photons do not change appreciably during a qpo cycle , the amplitude of the oscillation in luminosity caused by the oscillation in optical depth is comparable to the amplitude of the optical depth oscillation and is given by @xmath259 equations ( [ eq : ampkepler ] ) and ( [ eq : lumosc ] ) demonstrate that _ a relatively weak oscillation in optical depth of only a few percent that is accompanied by a luminosity oscillation of moderate amplitude can produce naturally a much larger oscillation in the countrate at high photon energies_. equation ( [ eq : ampkepler ] ) also shows that the relative amplitude of the oscillation at a given photon energy is different for different electron temperatures and optical depths , unless the relative amplitude of the oscillation in optical depth @xmath250 varies in such a way as to compensate for this effect . we therefore expect that as the mass accretion rate onto the neutron star changes ( on timescales much longer than a beat - frequency period ) , the electron temperature and optical depth in the comptonizing region will change , causing the photon - energy dependence of the beat - frequency and keplerian - frequency oscillation amplitudes to change .
_ photon - energy dependence of the sonic - point keplerian and beat - frequency oscillation phases_.at photon energies 30 kev , the cross section and hence the electron scattering mean free path of the photons in the central corona is nearly independent of photon energy .
therefore , the average escape time of photons from the corona is also independent of their energy .
however , the photons that stay in the corona longer experience more scatterings on average , and therefore gain more energy by scattering off hot electrons and emerge from the medium with a larger energy . as a result
, the oscillation at high photon energies is expected to lag the oscillation at low photon energies .
if the injection rate of soft photons and the properties of the corona are constant in time , the magnitude of the time lag is determined mainly by the properties of the corona ( see , e.g. , wijers , van paradijs , & lewin 1987 ; bussard et al .
if instead the source of soft photons or the properties of the corona are time - dependent , then the time lag will depend on the details of this time - dependence . for the simplified model described earlier in this section , the average energy after @xmath260 scatterings of a photon injected at energy @xmath232 is @xmath261\;.\ ] ] therefore , for photons with energies smaller than @xmath262 , the ratio of the energies of two photons that have experienced a different number of scatterings is @xmath263 , where @xmath264 is the difference in the number of scatterings .
the photon mean - free time is @xmath265 and hence the time lag introduced between the oscillations at the two photon energies is @xmath266 when both @xmath267 and @xmath268 are smaller than @xmath269 . because of the diffusion in energy and the systematic downscattering of photons with energies @xmath270 , equation([eq : timelag ] ) is not valid at energies @xmath270 .
we emphasize that , for a variety of reasons , equation ( [ eq : timelag ] ) gives only an upper bound to the expected time lag at photon energies @xmath271 .
first , this equation assumes that all photons are injected at the center of a spherical medium .
however , if most of the photons are produced near the surface of the neutron star and escape a few kilometers above it , the time lag can be significantly smaller .
second , equation ( [ eq : timelag ] ) was derived under the assumption that photons slowly diffuse outward .
this is not a very good approximation for electron scattering optical depths @xmath272 ; if a fraction of the photons escape directly from the corona without interacting with the electrons , the average escape time could be significantly smaller . finally , equation ( [ eq : timelag ] ) was derived under the assumption of a uniform electron density .
if instead most of the optical depth is concentrated near the center of the corona , this will also reduce significantly the average escape time . _ in summary , to the extent that the effects of comptonization dominate , we expect the sonic - point keplerian- and beat - frequency oscillations at high photon energies should lag the corresponding oscillations at lower photon energies and the time lag between a few kev and @xmath13 kev should be a fraction of a millisecond _
, for the densities and temperatures of the electrons expected around the neutron stars in the z and atoll sources . in the sonic - point model ,
gas spirals inward from density inhomogeneities in the accretion flow at the sonic radius and collides with the stellar surface , producing a radiation pattern that rotates around the star with the keplerian frequency at or near the sonic point @xmath273 .
because the rate of mass accretion over the whole neutron star surface remains approximately constant in time as the radiation pattern rotates , the total luminosity emerging from the system also remains approximately constant in time .
for conciseness we call the oscillations produced by rotation of this radiation pattern a _ beaming _ oscillation , because it is caused by the angular variation of the radiation field , even though the radiation pattern is unlikely to be a narrow beam . in a pure beaming oscillation ,
the total luminosity emerging from the source is independent of time .
as described earlier , the accretion flow is also expected to produce a much weaker beaming oscillation at the stellar spin frequency @xmath83 .
if the magnetic field of the neutron star is strong enough to channel the accretion flow near the stellar surface , the sonic - point model predicts that the resulting radiation pattern , which rotates with the star , will modulate inflow from the inner disk at the beat frequency @xmath106 , producing a _ luminosity _ oscillation with this frequency . in a pure luminosity oscillation ,
the total luminosity of the system changes with time , but the angular pattern of the radiation field remains unchanged . in the sonic - point model ,
the only luminosity oscillation is the beat frequency oscillation .
when the luminosity oscillation is present , it is necessarily modulated by the motion of the emission regions around the stellar surface , which generates second - order beaming oscillations at the stellar spin frequency @xmath83 and the difference frequency @xmath274 ( see 3.3 and table 3 ) . in general , higher - order oscillations are also generated at other ( mostly much higher ) frequencies .
scattering generally attenuates beaming oscillations much more than luminosity oscillations , for the conditions relevant to lmxbs ( lamb 1988 ) .
the reason is that an anisotropic radiation pattern becomes approximately isotropic after only a few scatterings , whereas a luminosity oscillation is attenuated by scattering only if the time required for photons to escape from the scattering region is much larger than the period of the oscillation , which requires a large optical depth for the oscillation frequencies and coronal dimensions of interest to us .
the relative amplitude @xmath275 of a luminosity oscillation with angular frequency @xmath276 outside a spherical scattering cloud of radius @xmath233 and optical depth @xmath47 is related to the relative amplitude @xmath277 at the center of the cloud by the expression ( kylafis & phinney 1989 ) @xmath278 where @xmath279^{1/2}$ ] .
this expression is accurate to better than 6% for @xmath280 ( kylafis & phinney 1989 ) .
the first term in parenthesis on the right side of equation ( [ eq : lumatten ] ) describes the amplitude attenuation caused by the spread of escape times from the cloud whereas the second term ( which has been added in by hand ) describes the amplitude of the oscillation produced by the photons that escape from the cloud without scattering . for a beaming oscillation caused by rotation of a narrow pencil beam ,
the amplitude of the oscillation outside the cloud is ( brainerd & lamb 1987 ; kylafis & phinney 1989 ) @xmath281a_{0,{\rm beam}}\;. \label{eq : beamatten}\ ] ] the factor multiplying @xmath282 in equation ( [ eq : beamatten ] ) describes the tendency of scattering to isotropize the photon distribution .
the amplitudes of oscillations produced by the broad radiation patterns or the radiation patterns with more than one lobe that we are concerned with here are reduced even more by scattering ( brainerd & lamb 1987 ) .
of all the beaming oscillations that may be generated by the accretion flow , only the one at the sonic - point keplerian frequency is likely to have a large enough amplitude at the neutron star to produce an oscillation that is strong enough , after attenuation by passage through the central corona , to be observed easily with current instruments .
in contrast , the oscillation at the sonic - point beat frequency , which is the only luminosity oscillation produced by the flow , should be strong enough to be observed even if its amplitude at the neutron star surface is moderate . as a result
, we expect that from the collection of possible frequencies of oscillations , the only oscillations that will appear strong far from the star will be those at the sonic - point keplerian and beat frequencies .
in this section we describe the observational implications of the sonic - point model and compare them with observations of the kilohertz qpos .
we first list the important properties of these qpos that any model must explain .
these properties are ( see 1 ) : ( 1 ) high ( 3001200 hz ) frequencies , which can vary by several hundred hertz in a few hundred seconds ( see wijnands et al . 1998 and van der klis 1995 ) ; ( 2 ) similar frequency ranges in stars with significantly different magnetic field strengths and accretion rates ; ( 3 ) relatively high coherence ( @xmath283 up to @xmath102 ) ; ( 4 ) 260 kev rms amplitudes @xmath14% in the z sources but much higher ( up to 15% ) in the atoll sources ; ( 5 ) the common occurrence of two ( but never more than two ) simultaneous kilohertz qpos in a given source ; ( 6 ) a frequency separation @xmath6 between the two kilohertz qpos that is approximately constant in a given source ; ( 7 ) approximate consistency of @xmath6 with the stellar spin frequency inferred from burst oscillations ; ( 8) the similarity of most atoll and z source spin frequencies ; ( 9 ) the frequently similar fwhm of the two qpo peaks seen simultaneously ; ( 10 ) the increase of the frequencies of the kilohertz qpos with increasing inferred accretion rate that is observed in many sources ; and ( 11 ) a steep increase in the amplitude of the higher - frequency of the two kilohertz qpos with increasing photon energy in the 215 kev energy band in many sources . in 4.1 we show that the stellar magnetic field strengths and electron scattering optical depths required for the sonic - point mechanism to operate are consistent with the stellar magnetic field strengths and optical depths of the compact central corona inferred from previous modeling of the x - ray spectra and @xmath284 hz x - ray variability of the atoll and z sources . in 4.2 we compare the kilohertz qpo and neutron star spin frequencies expected in the sonic - point model with the kilohertz qpo and neutron star spin frequencies observed in the atoll and z sources .
we also discuss the expected and observed dependence of the qpo frequencies on accretion rate . in
4.3 describe the amplitude and coherence of the various qpos expected in the sonic - point model , comparing them with the observed amplitudes and coherence of the kilohertz qpos .
in particular , we show that the sonic - point model explains naturally why at most two kilohertz qpos have been detected in the power spectrum of each source .
we also discuss the relative coherence of qpos at different frequencies and the expected dependence of qpo amplitudes on accretion rate and photon energy . in 4.4 we describe the inverse correlation between qpo amplitude and stellar magnetic field expected in the sonic - point model and show that there is already substantial evidence for such a correlation .
the operation of the sonic - point mechanism for producing strong , coherent kilohertz qpos requires that a number of conditions be satisfied : ( 1 ) in order to produce the higher - frequency kilohertz qpos with the large amplitudes observed in some sources , the neutron star magnetic fields in these sources must be relatively weak , so that a substantial fraction of the accreting gas penetrates close to the star in a keplerian disk flow ( 3.1 ) .
( 2 ) in order to produce the lower - frequency kilohertz qpos , the neutron star magnetic field , although relatively weak , must be strong enough to channel some of the accreting gas near the stellar surface to produce bright spots that rotate with the star and modulate the accretion rate at the sonic - point beat frequency ( 3.1 ) .
( 3 ) in order to generate qpos with the high coherence observed in most sources , the disk flow at the sonic point must be geometrically thin ( 3.4 ) .
( 4 ) in order that only the qpos at the sonic - point keplerian and beat frequencies be detectable and that x - ray oscillations at the stellar spin frequency be very weak or undetectable at present sensitivities , electron scattering in the central corona must help to attenuate the weak beaming oscillations at other frequencies and at the stellar spin frequency .
the central corona must therefore have an electron scattering optical depth @xmath285 ( 3.6 ) .
as we now explain briefly , all of these conditions follow naturally from the unified model of weak - field accreting neutron stars described in
2.2 . the magnetic field strengths in the 4u atoll sources are thought to be low enough ( @xmath286 g ) that at their inferred accretion rates ( @xmath287@xmath288 ) the cylindrical radius @xmath63 at which the keplerian disk flow couples strongly to the stellar magnetic field is @xmath289 cm
. the magnetic field strengths of the neutron stars in the z sources are thought to be a few times larger than in the 4u atoll sources , but the accretion rates are much larger as well , so @xmath63 is comparable to its value in the atoll sources .
thus , in both the 4u atoll sources and the z sources , the gas in the keplerian disk penetrates close to the star before any of it couples strongly to the stellar magnetic field .
we therefore expect that a significant fraction of the accreting gas will remain in a keplerian disk flow down to the sonic point and will continue in a disk flow very close to the stellar surface , as required in the sonic - point model .
evidence that the magnetic fields of neutron stars in lmxbs , while typically weak , are nevertheless strong enough in many sources to channel the flow near the star and hence to produce a qpo at the sonic - point beat frequency comes from power spectra constructed from observations of ( van der klis et al.1996e ) , 1 ( van der klis 1996a , 1996b , 1996d , 1997b ) , ( wijnands et al .
1997c ) , and ( wijnands et al .
in addition to two simultaneous kilohertz qpos , these power spectra show horizontal branch oscillations , which appear to be magnetospheric beat - frequency oscillations ( see 2 and psaltis et al .
1998 ) .
these power spectra indicate that in the z sources , some of the gas in the disk couples strongly to the weak stellar magnetic field at two or three stellar radii and is funneled into hot spots that could modulate the mass flux from the sonic point at the sonic - point beat frequency by periodically irradiating the clumps at the sonic radius ( see 2 ) .
it is natural to expect that the magnetic fields of many of the atoll sources are only a little weaker than those in the z sources and are therefore still strong enough to produce a qpo at the sonic - point beat frequency .
this expectation is supported by the spectral modeling described in 2.2 .
accretion via a geometrically - thin keplerian disk flow , which is required to explain the high coherence of the kilohertz qpos , is expected in the atoll sources , because at their low accretion rates the inner part of the accretion disk is likely to be gas - pressure - dominated . in the z sources , which are thought to be accreting at close to the eddington critical rate , gas in the keplerian disk flow that is not channeled by the stellar magnetic field is likely to be pinched by the stellar field into a geometrically thin disk .
the final requirement listed above for the sonic - point model , namely that the optical depth of the hot central corona exceed @xmath290 , follows directly from the spectral modeling discussed in 2.2 . in summary ,
the physical picture of the atoll and z sources that was developed prior to the discovery of the kilohertz qpos , based on their 220 kev x - ray spectra and 1100 hz x - ray variability , give the magnetic field strengths , accretion flows , and electron scattering optical depths required by the sonic - point model . in the sonic - point model ,
the frequency of the higher - frequency kilohertz qpo ( when two are present ) is the keplerian orbital frequency @xmath104 at the point in the disk flow where the inward radial velocity increases rapidly within a small radial distance , whereas the frequency of the lower - frequency kilohertz qpo is approximately the difference between @xmath104 and the stellar spin frequency @xmath83 . in this section
we describe the range of sonic - point keplerian and neutron star spin frequencies and the variation of the sonic - point keplerian and beat frequencies with accretion rate expected in the model and then compare these expectations with the observations .
we discuss the expected amplitudes of oscillations at other frequencies in 4.3 .
_ similarity of the sonic - point keplerian qpo frequencies in different sources_.a key question is why the higher - frequency kilohertz qpos in different sources always have frequencies between @xmath0500 hz and , even though the neutron stars in these sources have time - averaged mass accretion rates that differ by more than a factor of 100 and magnetic fields that are thought to differ by more than a factor of 10 . in the sonic - point model
there are three main reasons why the keplerian - frequency qpo is restricted to this frequency range : \(1 ) as discussed in 3.2 , there is an upper bound on the radius @xmath291 of the sonic point , which is set by the maximum fraction of the angular momentum of the accreting gas that can be removed by radiation from the stellar surface and is @xmath292 , where @xmath98 is the radius of the marginally stable orbit .
there is also a lower bound on @xmath221 , which is @xmath94 if the radius @xmath54 of the neutron star is smaller than @xmath94 , or @xmath54 otherwise .
hence , the keplerian frequency at the sonic point is confined to a similar interval in the atoll and z sources , regardless of their accretion rates and magnetic field strengths . as a result ,
if the neutron stars in the kilohertz qpo sources have typically accreted a few tenths of a solar mass and therefore have masses @xmath293 , these constraints on the sonic - point keplerian frequency mean that the frequencies of their sonic - point keplerian qpos will fall between @xmath0400 hz and @xmath01300 hz .
\(2 ) modeling of the 2 - 20 kev x - ray spectra and 1100 hz power spectra of the atoll and z sources indicates that the magnetic fields of these neutron stars are positively correlated with their time - averaged accretion rates , i.e. , sources with higher accretion rates appear to have stronger magnetic fields ( see 2.2 and psaltis & lamb 1998a , 1998b , 1998c ) .
this has an important implication for the frequency range of the kilohertz qpos , because the stronger the stellar magnetic field , the _ larger _ the fraction of the gas in the disk that is likely to couple to the magnetic field at two or three stellar radii and be channeled out of the disk flow there . as a result ,
the stronger the magnetic field , the _ smaller _ the mass flux @xmath294 through the disk at the sonic point .
hence , @xmath294 differs by a much smaller factor in the z and atoll sources than does the total mass flux @xmath295 onto the neutron star .
the sonic - point keplerian frequency @xmath104 is governed more by @xmath296 than by @xmath295 , so the positive correlation between magnetic field and accretion rate means that the frequency ranges of the sonic - point qpos are likely to be more similar in a collection of neutron - star lmxbs than one would expect , based only on the different ranges of @xmath295 in different systems .
( 3 ) in addition , the inner disk is expected to be thicker if @xmath297 is large . as a result ,
the radial optical depth through the disk at a given accretion rate is smaller , which allows radiation from the neutron star surface to penetrate further out into the disk than one would expect based only on the different values of @xmath295 in the different systems .
this tendency also acts to make the range of sonic - point radii , and hence the frequency ranges of the keplerian - frequency qpos , more similar in different systems than consideration of the accretion rates alone would suggest ( see 3.2 ) . _
expected and inferred neutron star spin rates_.in the sonic - point model , the separation between the frequencies of the qpos in a pair is approximately equal to the stellar spin frequency .
the frequency separations observed in the kilohertz qpo sources so far ( see table [ table : spinfrequencies ] ) all indicate spin rates of a few hundred hertz , consistent with the spin rates expected in the magnetospheric beat - frequency model of the hbo ( see 2 , lamb et al .
1985 , and ghosh & lamb 1992 ) .
oscillations have been observed during thermonuclear x - ray bursts from five kilohertz qpo sources so far ( again see table [ table : spinfrequencies ] ) .
only a single oscillation has been observed from each source during a given burst , the oscillations in the tails of bursts appear to be highly coherent ( see , e.g. , smith et al .
1997 ) , and the frequencies are always the same in a given source . furthermore , comparison of burst oscillations from over about a year shows that the timescale for any variation in the oscillation frequency is @xmath298 yr ( see , e.g. , strohmayer 1997 ) .
the burst oscillations are thought to be caused by emission from one or two regions of brighter x - ray emission on the stellar surface ( see strohmayer et al .
1997b ) , producing oscillations at the stellar spin frequency or its first overtone , respectively .
the 363 hz frequency of the burst oscillation observed in is consistent with the separation frequency of its two simultaneous kilohertz qpos .
the 524 hz and 581 hz frequencies of the burst oscillations seen in and are probably twice the spin frequencies of these neutron stars ( see morgan & smith 1996 ; smith et al .
1997 ; wijnands & van der klis 1997 ; zhang et al . 1996 , 1997 ) .
if so , the spin rates of these stars are 262 hz and 290 hz , respectively
. the frequencies of the fairly coherent oscillations seen at 549 hz in 1 ( zhang et al .
1998 ) and at 589 hz in the so - far - unidentified source near the galactic center ( strohmayer et al .
1996d ) may also be twice the spin frequencies of these neutron stars .
these spin rates are all consistent with those expected in the sonic - point model
. _ similarity of the neutron - star spin frequencies in different sources_.if the neutron stars in the kilohertz qpo sources have spin frequencies comparable to their equilibrium spin frequencies , then we expect their spin frequencies to be a few hundred hertz . this can be seen from equation ( [ eq : mukepler ] ) in 2.2 , which can be solved for the equilibrium spin frequencies at which continued accretion at the given rate leaves the spin frequency unchanged , with the result @xmath299 here @xmath300 is the critical fastness ( ghosh & lamb 1979b ) .
hence , for @xmath301 , a @xmath302 atoll source with @xmath303 accreting from a gpd disk at a rate @xmath304 has an equilibrium spin frequency of 230 hz , which is similar to the 240 hz equilibrium spin frequency of a @xmath101 z source with @xmath305 accreting from an rpd disk at a rate @xmath306 . a recent analysis ( psaltis et al .
1998 ) of the properties of the hbos and kilohertz qpos in a collection of five z sources shows that they are consistent with many of the predictions of the magnetospheric beat - frequency model of the hbo and that the agreement is best if , as expected in this model , they are all in spin equilibrium ( see also ghosh & lamb 1992 ) . in particular , this analysis shows that _ the narrow range of spin frequencies observed in the z sources is to be expected if they are in spin equilibrium_. _ variation of kilohertz qpo frequencies with mass accretion rate_.in the sonic - point model , the frequencies of the kilohertz qpos are expected to rise as the accretion rate increases , within the bounds set by the orbital frequencies at the angular momentum loss radius and at the radius of the marginally stable orbit ( see above and 3.2 ) . in order to see why an increase is expected , suppose that the sonic radius is at @xmath307 for a given accretion rate .
an increase in the accretion rate will cause the radial optical depth from the stellar surface to @xmath307 to increase , if all other physical quantities remain constant , and hence the sonic point will move inward to the smaller radius @xmath308 at which the optical depth from the stellar surface is approximately the same as before , causing an increase in the orbital frequency at the sonic radius .
the x - ray luminosity between bursts is produced almost entirely by accretion and hence an increase in the accretion rate causes an increase in the luminosity .
we therefore expect a strong , positive correlation between the frequencies of the sonic - point keplerian- and beat - frequency qpos and the persistent x - ray luminosity .
these basic ideas are illustrated by the general relativistic accretion flow calculations of 3.2 .
as shown in figure [ fig : nukvsl ] , the sonic - point keplerian and beat frequencies given by these calculations increase steeply with increasing mass accretion rate until the sonic point reaches @xmath94 , at which point their frequencies stop changing .
this flattening of the two frequency vs. accretion luminosity relations is one possible signature of the existence of a marginally stable orbit ( see also 5.4 ) .
coherent keplerian- and beat - frequency qpos are unlikely to be produced if the sonic point moves inward , close to the stellar surface , because of the disruptive effect of the surface magnetic field and the strong viscous shear layer that is expected to develop if the keplerian flow interacts directly with the neutron star surface .
we emphasize that these calculations are intended to be illustrative rather than to reproduce the qpo frequency behavior of any particular source .
if the mass of the neutron star is @xmath302 , rather than @xmath101 as assumed in these calculations , the limiting frequency would be 1.3 rather than 1.6 ( see miller , lamb , & psaltis 1998 ) . again , if the accretion disk is thicker near the star than is assumed in these calculations , more of the inner disk will be illuminated and hence the angular momentum loss radius @xmath81 will be larger , causing the sonic - point keplerian frequency at a given accretion rate to be lower . if the disk thickness changes with accretion rate , then the variation of the sonic - point keplerian frequency with accretion rate will be different from the variation found in the calculations reported here , which treat the disk thickness as constant . despite these uncertainties ,
the calculations described here show that an increase in the mass accretion rate leads naturally to an increase in the sonic - point keplerian frequency .
we stress that caution must be used in comparing the predicted variation of qpo frequencies with accretion with the countrate measured by a given x - ray detector , because countrates are known to be poor indicators of the mass accretion rate and the accretion luminosity , at least for some sources at some times .
the most likely reason is that a particular instrument measures the photon number flux over only a restricted energy range and that a change in the mass accretion rate typically causes a change in the shape of the x - ray spectrum as well as its normalization .
hence the detector countrate typically is not proportional to the accretion rate . in the z sources ,
the mass accretion rate is known to be different when the countrate is the same but the source is on a different branch of the z track ( hasinger & van der klis 1989 ; lamb 1989 ; psaltis et al.1995 ) . moreover , in the `` cyg - like '' z sources , the normal / flaring branch vertex , which is thought to correspond to a mass accretion rate approximately equal to @xmath152 ( lamb 1989 ) , occurs at different countrates during different observations ( kuulkers 1995 ) .
the observations of ford et al .
( 1997a ) show that the properties of the accretion flow are not uniquely related to the x - ray countrate in the atoll sources . in these observations , the dependence of the kilohertz qpo frequencies on countrate seen in data on taken with in 1996 august differs , both in slope and in normalization , from the dependence of the qpo frequencies on countrate seen in data taken in 1996 april .
in contrast , when the qpo frequencies are plotted against the energy of the spectral peak , which may be a good indicator of the accretion rate ( psaltis & lamb 1998c ) , the slope and normalization of the curves are the same for both observations ( ford et al .
1997b ) .
clearly , bolometric and other corrections are typically important , so comparison of qpo frequency versus countrate data with the qpo frequency versus accretion rate predictions of theoretical models must be approached with caution . despite this caveat ,
it is worth emphasizing that there is a strong , positive correlation between the frequencies of the kilohertz qpos seen in , , , and ks 1731@xmath309260 , and the countrate measured by , as expected in the sonic - point model if the countrate increases with increasing accretion rate .
it is interesting to speculate about the dependence of kilohertz qpo frequencies on luminosity that is to be expected if such qpos are detected during a thermonuclear x - ray burst . in this case
the mass accretion rate and the radial optical depth are no longer tied to the luminosity , and hence the expected dependence of @xmath104 on luminosity is different from what is expected during the periods between type i x - ray bursts .
when radiation forces are extremely strong , such as at the peak of a type i x - ray burst that causes photospheric radius expansion , the radiation controls the accretion flow out to large radii and we therefore do not expect any qpo at the sonic - point keplerian frequency or at the sonic - point beat frequency .
however , qpos at these frequencies may reappear in the tail of the burst . while the burst luminosity is several times the persistent accretion luminosity
, we expect the sonic point to be farther from the star than when the luminosity is lower .
thus , we expect that if sonic - point qpos are detected in the tail of a thermonuclear burst , their frequencies will increase as the luminosity declines .
note , however , that during the decay phase of such a burst , the luminosity changes rapidly .
if the frequencies of the qpos track this change , the qpo peaks in power spectra will be smeared out , unless the spectra are constructed from short ( @xmath310 s ) segments of data .
so far , no kilohertz qpos have been observed near the maximum of type - i x - ray bursts , which is consistent with what is expected in the sonic - point model .
more sensitive searches will be required to determine if kilohertz qpos are present during the decay phase of bursts . _
amplitudes of oscillations at different frequencies_.a key question that any theory of the kilohertz qpos must answer is why at most two strong qpos have so far been seen simultaneously in the kilohertz qpo sources . in addressing this question , it is important to consider both the _ generation _ of oscillations near the neutron star and the effects of _ propagation _ of the radiation through the gas surrounding the neutron star . in order to be detectable far away
, an oscillation must be strong at the source or have a small attenuation , or both .
the qpos at the sonic - point keplerian and beat frequencies are the only qpos expected to have moderately high amplitudes far from the neutron star because the optical depth and luminosity oscillations that produce them generate relatively high amplitudes near the star , because these frequencies ( and their overtones ) are the _ only _ frequencies generated by the sonic - point mechanism in lowest order , and because scattering of photons by the gas surrounding the neutron star selectively suppresses the already weak higher - frequency oscillations generated at higher orders . as discussed in 3.3 , to lowest ( first ) order , the sonic - point mechanism generates oscillations only at the sonic - point keplerian and beat frequencies ( see table [ table : freqgen ] ) . as explained there ,
only very weak overtones of the sonic - point keplerian frequency are likely to be generated , because the angular distribution of the radiation from the bright impact footprints of the rotating density pattern produced by clumps is expected to be very broad . the radiation pattern that rotates with the star is also expected to be broad , but at the sonic radius , where it interacts with the orbiting gas to generate the luminosity oscillation at the sonic - point beat frequency , it is not likely to be perfectly sinusoidal
. hence overtones of the sonic - point beat frequency @xmath107 may be generated in the inward mass flux from the clumps at the sonic radius , producing overtones of @xmath107 in the x - ray flux from the star .
as radiation from the stellar surface propagates outward through the part of the central corona that extends beyond the sonic radius , the radiation pattern that rotates with the star is likely to be broadened further ( see below ) , reducing the amplitudes of oscillations at the stellar spin frequency and its overtones seen by a distant observer .
in addition to the qpos at the sonic - point keplerian and beat frequencies and their overtones , in second order the sonic - point mechanism also generates qpos at @xmath83 and @xmath311 and in third order at @xmath312 , @xmath313 , @xmath314 , and @xmath315 ( again see table [ table : freqgen ] ) .
these high - order oscillations are expected to be very weak , even near the neutron star .
the attenuation of an oscillation depends strongly on whether it is a _ beaming _ oscillation or a _ luminosity _ oscillation ( see 3.6 ) .
a beaming oscillation is an oscillation produced by rotation of an angular radiation pattern , like the beam of a lighthouse ; in a pure beaming oscillation , the luminosity remains constant as the radiation pattern rotates .
in contrast , a luminosity oscillation is one produced by an oscillation of the luminosity of the source ; in a pure luminosity oscillation , the radiation pattern remains static as the luminosity oscillates . in the sonic - point model ,
the sonic - point keplerian frequency is the only intrinsically strong beaming oscillation , and the sonic - point beat frequency is the only luminosity oscillation produced by the accretion flow . in passing through a scattering region ,
luminosity oscillations are attenuated only by time - of - flight smearing whereas beaming oscillations are also attenuated by the isotropization of the radiation pattern by the scattering . hence beaming
oscillations are weakened more by propagation through a scattering corona .
figure [ fig : attenfactor ] compares the attenuation of luminosity and beaming oscillations produced at the center of a uniform , spherical , scattering cloud of radius @xmath316 cm , as a function of the optical depth of the cloud .
the size of this cloud is comparable to the dimensions of the small scattering coronae with optical depths @xmath035 that are thought to surround the neutron stars in the atoll and z sources ( lamb 1989 , 1991 ; see also 2.2 ) .
the frequencies of the oscillations shown in figure [ fig : attenfactor ] have been chosen to represent a hypothetical accreting neutron star with @xmath317 hz and @xmath318 hz . the beat - frequency luminosity oscillation is therefore at 800 hz .
the attenuation factors have been calculated using equations ( [ eq : lumatten ] ) and ( [ eq : beamatten ] ) .
figure [ fig : attenfactor ] shows that the beaming oscillation at the sonic - point keplerian frequency must have a relatively high intrinsic amplitude in order to produce a strong , observable qpo for a distant observer , whereas the luminosity oscillation at the sonic - point beat frequency needs only to have a moderate amplitude .
the amplitudes of the oscillations at the overtones of either luminosity or beaming oscillations seen by a distant observer are expected to be far smaller than the amplitudes seen at the fundamental frequencies , for several reasons .
for example , beaming patterns with multiple lobes are much more easily isotropized by scattering than are beaming patterns with single lobes ( see , e.g. , brainerd & lamb [ 1987 ] ) . moreover , for either beaming or luminosity oscillations , overtones of an oscillation are at higher frequencies than the fundamental and are thus more susceptible to time - of - flight smearing . as a specific example
, we consider .
as interpreted within the sonic - point model , the spin frequency of this neutron star is 363 hz ( see table 3 ) . during one observation
, @xmath104 was 1045 hz . assuming that this neutron star is surrounded by a central corona with a radius @xmath319 cm and an optical depth @xmath320 , the amplitude of the luminosity oscillation at the beat frequency @xmath321 seen by a distant observer is about 85% of the amplitude at the neutron star .
for comparison , the beaming oscillation at @xmath322 , which is generated only in third order and is therefore likely to be very weak even near the neutron star , has an amplitude at infinity which is only about 20% of its amplitude at the star .
the amplitudes of oscillations at the overtones of the sonic - point keplerian frequency are all reduced by factors @xmath32320 . for these reasons ,
it is not surprising that the only strong qpos seen are at the sonic - point keplerian and beat frequencies .
although oscillations at frequencies other than the sonic - point keplerian and beat frequencies are likely to be very weak , their amplitudes are surely not zero . we therefore expect that , if neutron - star lmxbs are observed with sufficient sensitivity , oscillations at other frequencies will be detected . in particular , a careful search near the spin frequency , at the first overtone of the beat frequency , and near the sum of the keplerian and spin frequencies in sources with pairs of high - frequency qpo peaks may reveal very weak qpos at these frequencies .
the oscillation at the beat frequency is a _
oscillation , but in the sonic - point model it is created by interaction of the weakly _ beamed _ radiation pattern that is rotating at the stellar spin frequency with the clumps of gas that are orbiting at or near the sonic point .
it is therefore important to emphasize why the beaming oscillation at the stellar spin frequency is strong enough to modulate significantly the mass inflow rate from clumps at the sonic point , yet too weak to produce a significant peak in power spectra constructed by a distant observer .
previous modeling of the x - ray spectra ( lamb 1989 , 1991 ; psaltis et al .
1995 ; psaltis & lamb 1998a , 1998b , 1998c ) and the 1100 hz x - ray variability of the atoll and z sources ( lamb 1989 , 1991 ; miller & lamb 1992 ) strongly indicates that a hot , central , comptonizing corona surrounds the neutron star , extending several stellar radii from its surface ( see 2.2 ) .
the sonic radius is well inside this corona , so the scattering optical depth from the stellar surface to the sonic point is much less than the scattering optical depth from the stellar surface to infinity .
the finite size of the neutron star also diminishes the effect of attenuation on the amplitude of x - ray brightness oscillations for gas orbiting close to the star .
these factors make the brightness oscillation produced by the radiation pattern that rotates with the star much stronger at the sonic radius than it is far away .
_ dependence of qpo amplitudes on accretion rate_.in the sonic - point model , the ratio of the amplitude of the qpo at the sonic - point keplerian frequency to the amplitude of the qpo at the sonic - point beat frequency generally depends on the strength of the neutron star magnetic field .
if the magnetic field is too weak to channel gas even near the stellar surface , the amplitude of the oscillation at the beat frequency will be too small to be detected ( see 3.1 ) .
however , the ratio of the amplitudes of the qpos at different frequencies is also expected to depend on the mass accretion rate .
for example , as the mass accretion rate rises , the electron density and optical depth of the gas around the neutron star also rise .
hence , the amplitude at infinity of the beaming oscillation at the sonic - point keplerian frequency will fall faster relative to the amplitude of the luminosity oscillation at the sonic - point beat frequency , all else being equal ; indeed , the qpo at the keplerian frequency may become undetectable while the qpo at the beat frequency remains strong .
this expectation is consistent with the observed behavior of the kilohertz qpo pair in ( strohmayer et al .
1996c ) , assuming that the increasing mass accretion rate causes an increase in the countrate ( but see 4.2 ) .
in this source , the amplitude of the higher - frequency qpo decreases relative to the amplitude of the lower - frequency qpo as the countrate increases . alternatively ,
if the sonic point moves far enough away from the stellar surface , as may happen at low accretion rates ( and hence countrates ) , the optical depth through the flow from the stellar surface to the sonic radius may become large enough to reduce greatly the amplitude at the sonic point of the radiation pattern that rotates with the star , thereby suppressing the luminosity oscillation at @xmath107 .
this is consistent with an observation of in which the lower - frequency qpo is not observed when the frequency of the higher - frequency qpo is low ( strohmayer et al .
1996c ) .
this analysis shows that if only a _ single _ high - frequency qpo peak is observed , it could be either the sonic - point keplerian - frequency qpo or the sonic - point beat - frequency qpo , depending on the magnetic field of the neutron star and the mass accretion rate at the time . in this case , a secure identification can only be made by considering other properties of the qpo and comparing them with the properties of the qpos seen in other observations of the same source .
_ dependence of qpo amplitudes on photon energy_.in the sonic - point model , the relative amplitudes of the keplerian frequency and beat frequency qpos are expected to increase steeply with photon energy over the @xmath0510 kev energy range ( see 3.5 ) .
this is because the optical depth is expected to oscillate at both frequencies : the optical depth along the line of sight from the stellar surface oscillates at the keplerian frequency , while the total optical depth of the scattering region oscillates at the beat frequency as the density of the accreting gas falling on the stellar surface oscillates at this frequency . as shown in 3.5 ,
a modest oscillation in the optical depth produces a qpo with a large relative amplitude at high photon energies .
this is consistent with the steep increase of qpo amplitude with increasing photon energy observed in the higher - frequency of the two simultaneous qpos in ( zhang et al .
1996 ) , ( ford et al .
1997b ) , and ks 1731@xmath309260 ( wijnands & van der klis 1997 ) , in the lower - frequency of the two simultaneous qpos in ( berger et al .
1996 ; mndez et al .
1998 ) and ( strohmayer et al .
1996c ) , and in the kilohertz qpos in the z sources ( van der klis et al .
1996e ) , ( wijnands et al .
1997c ) , and ( wijnands et al .
1998 ) . in 3.5 we used a simple analytical model to derive an expression for the photon energy dependence of the rms amplitude of a qpo produced by oscillations in the optical depth and in the injection rate of soft photons . figure [ fig : ampvsenergy ] compares the results of a more detailed numerical calculation performed using the algorithm of miller & lamb ( 1992 ) with amplitude data for the lower - frequency kilohertz qpo observed in and the higher - frequency kilohertz qpo seen in . in performing this calculation
we made the same assumptions as in
3.5 and assumed further that the injection rate of soft photons is constant in time , that the spectrum of the injected photons is a blackbody at temperature @xmath324 kev , that the electron temperature in the central corona is @xmath325 kev , and that the optical depth varies from @xmath326 to @xmath327 during an oscillation .
the model matches the data well .
the geometry of the actual upscattering region is undoubtedly much more complicated than assumed in this calculation , but the excellent correspondence with the data and the ubiquity of the steep increase of amplitude with photon energy over the 510 kev energy range suggest that the model has many elements in common with the true physical situation . as a source moves in the x - ray color - color diagram and its spectrum changes ( implying a change in the average optical depth and electron temperature of the comptonizing region )
, we expect that the photon energy dependence of the rms amplitude of the oscillations will also change .
moreover , as discussed in 3.5 , we expect that , in general , the dependence of the beat - frequency qpo amplitude on photon energy will be different from the dependence of the keplerian - frequency qpo amplitude .
this is consistent with the observations of ( ford et al.1997b ) in which amplitude versus photon energy curve of the beat - frequency qpo has a minimum at @xmath010 kev , whereas the amplitude versus photon energy curve of the keplerian - frequency qpo increases monotonically from 2 kev to 20 kev . because the qpos amplitude depend strongly on photon energy , to be meaningful a comparison of qpo amplitudes measured for different sources or at different times for the same source must consider the same range of photon energies .
_ coherence of the kilohertz qpos_.in
3.4 we showed that the high coherence of the keplerian and beat - frequency qpos in the sonic - point model is primarily a consequence of the extremely sharp increase in the inward radial velocity near the sonic point , which maps a small range of orbital frequencies onto the stellar surface .
there we also considered the decoherence caused by destruction of clumps by turbulent dissipation within the disk flow , by advection to the stellar surface , and by decay as gas is stripped from them by the supersonic flow at the sonic point .
we concluded that the qpo peaks can be as narrow as observed ( @xmath328200 ) for reasonable conditions in the accretion flow .
the fwhm @xmath329 of the sonic - point beat frequency oscillation is expected to be comparable to the fwhm @xmath330 of the sonic - point keplerian frequency oscillation , because the beat - frequency qpo is produced by the beat of a nearly periodic signal at frequency @xmath83 against the keplerian frequency ( see 3.4 ) .
we do not expect the fwhm of the two kilohertz qpo peaks to be identical , however , because there are processes that can affect the fwhm of one but not the other .
for example , the fwhm of the beat - frequency peak depends in part on the range of radii over which radiation forces can affect the mass accretion rate significantly
. this range can be either less than or greater than the range of radii over which the inward radial velocity increases rapidly , and hence @xmath329 can be either less than or greater than @xmath330 .
this effect can also displace the centroid of the beat - frequency peak relative to the centroid of the keplerian frequency peak , so that the observed frequency difference @xmath331 is close to , but not exactly equal to , the stellar spin frequency . in the sonic - point model ,
the larger the magnetosphere , the smaller the fraction of the accreting gas that reaches the surface of the star without coupling to the magnetic field , and hence the smaller the amplitude of the kilohertz qpos .
therefore , we expect the rms amplitudes of the kilohertz qpos to be roughly anticorrelated with the strength of the stellar magnetic field , if all other physical quantities remain fixed . as discussed in 2.2 ,
the `` 4u '' atoll sources are thought to have the weakest magnetic fields , the `` gx '' atoll sources and the `` sco - like '' z sources are thought to have somewhat stronger fields , and the `` cyg - like '' z sources are thought to have the strongest fields ( psaltis & lamb 1998a , 1998b , 1998c ) .
hence , we expect kilohertz qpos to be common and strong in the `` 4u '' atoll sources , but weak or even undetectable with current instruments in the `` gx '' atoll sources and the z sources .. to date , kilohertz qpos have not been detected in any of the gx atoll sources .
] this is indeed the case , as is evident from table [ table : reportedqpos ] .
if , as indicated by the spectral modeling discussed in 2 , the magnetic field of cir x-1 is weak , we expect it to exhibit sonic - point qpos when it is in its low state .
we also expect that kilohertz qpos will be undetectable by current instruments in any sources that have strong magnetic fields and therefore produce strong , periodic oscillations at their spin frequencies .
the relation between kilohertz qpo amplitude and magnetic field strength may be made semi - quantitative using the spectral calculations ( see 2.2 ) of psaltis et al . ( 1995 ) and psaltis & lamb ( 1998a , 1998b , 1998c ) .
these calculations indicate that lmxbs containing neutron stars with weaker magnetic fields generally have larger hard x - ray colors ( e.g. , the ratio of the 720 kev countrate to the 57 kev countrate ) . based on this physical picture , we expect the rms amplitudes of the kilohertz qpos produced by the sonic - point mechanism to be higher in sources with larger hard colors .
this trend is evident in figure [ fig : ampvshardcolor ] , which shows the rms amplitudes versus the hard x - ray colors of four sources observed with the pca detector onboard the satellite .
if other sources with kilohertz qpos follow this same trend , this will be strong support for the sonic - point model . as discussed in 2.3 ,
the observed anticorrelation between kilohertz qpo amplitude and magnetic field strength is also strong evidence that the magnetospheric beat - frequency mechanism _ not _ the correct explanation of these oscillations .
in the sonic - point model , the higher - frequency qpo in a kilohertz qpo pair has a frequency equal to the orbital frequency of gas near the sonic point ; the relatively high coherence of this qpo is a consequence of the fact that the gas clumps that generate it are in nearly circular orbits . in this section
we show that the inferred existence of a nearly circular orbit around a neutron star with a frequency in the kilohertz range can be used to derive interesting new upper bounds on the mass and radius of the star and constraints on the equation of state of the dense matter in all neutron stars . for simplicity
we discuss first the case of a nonrotating star around which the radial component of the radiation force is negligible .
we then consider the changes in the mass and radius constraints caused by frame - dragging when the star is spinning and by the radial component of the radiation force .
we show that rotation at @xmath40 hz typically increases the bound on the mass by @xmath86% and the bound on the radius by a few percent . in contrast , the radiation force _ reduces _ the upper bounds on the masses and radii by a few percent or less in the atoll sources but perhaps by much larger percentages in the z sources .
observations of kilohertz qpos may be able to establish the existence of an innermost stable circular orbit around some neutron stars ( see 4.2 ) .
if this can be accomplished , it would be the first evidence concerning a prediction of general relativity in the strong - field regime .
if furthermore the frequency of a particular kilohertz qpo can be securely established as the orbital frequency at the radius of the innermost stable circular orbit and if the spin frequency of the neutron star can be determined , then the frequency of the qpo can be used to fix the mass of the neutron star for each assumed equation of state , tightening the constraints on the properties of dense matter and possibly ruling out many currently viable equations of state .
we discuss how this can be done and the evidence that would signal detection of a qpo with the orbital frequency of the marginally stable orbit .
suppose that the frequency of the higher - frequency qpo in a kilohertz pair is @xmath23 and that , as in the sonic - point model , @xmath23 is the orbital frequency of gas in a nearly circular keplerian orbit around the neutron star .
assume for now that the star is not rotating and is spherically symmetric .
then the exterior spacetime is the schwarzschild spacetime . in this spacetime ,
the orbital frequency ( measured at infinity ) of gas in a circular orbit at boyer - lindquist radius @xmath100 around a star of mass @xmath65 is ( see eq . [ [ eq : sonickeplerfreq ] ] ) @xmath332 here and below the superscript zero indicates that the relation is that for a nonrotating star .
if the mass of the star is known , equation ( [ eq : keplerfreqzero ] ) may be solved for the orbital radius @xmath333 where the keplerian frequency is @xmath23 , with the result @xmath334 conversely , if the orbital radius of the gas is known , equation ( [ eq : keplerfreqzero ] ) may be solved for the mass of the star that gives an orbital frequency equal to @xmath23 , with the result @xmath335 if , as is so far the case for the kilohertz qpo sources , neither @xmath65 nor @xmath333 are known , equations ( [ eq : orbitalradiuszero ] ) and ( [ eq : orbitalmasszero ] ) do not determine @xmath333 or @xmath65 but do establish a relation between them . in the radius - mass plane , this relation is a curve that begins at the origin and rises up and to the right . as a specific example ,
the dashed curve marked @xmath336 in figure [ fig : pieslice ] shows the relation given by equation ( [ eq : orbitalmasszero ] ) for @xmath337 , a value of @xmath23 observed in ( w. zhang , personal communication ) .
hence , if were not rotating , the mass of the neutron star in this source and the orbital radius of the gas clumps producing the qpo during this observation would have to correspond to some point along this dashed curve .
consider now the constraints on the neutron star mass and radius that follow from the _ frequency _ of the higher - frequency qpo . obviously , in order for gas to be in orbit , it must be outside the star .
this means that for an orbit of given radius @xmath333 , the mass @xmath65 of the star must be greater than @xmath338 ; alternatively , for a star of given mass @xmath65 , the radius @xmath54 of the star must be less than @xmath339 .
thus , the point that represents the mass and radius of the neutron star must lie above the curve @xmath338 in the radius - mass plane .
the larger the value of @xmath23 , the higher the curve , so the most stringent and hence the relevant constraint on the mass and radius of the neutron star in a particular source is given by the _ highest _ value of @xmath23 ever observed in that source , which we denote @xmath340 . for example
, 1220 hz is the highest value of @xmath23 seen so far in , so this is the current value of @xmath340 for this source ( in fact , this is the highest value of @xmath341 seen so far in _ any _ source ) .
consequently , if the neutron star in were not rotating , the point in the radius - mass plane that represents its mass and radius would have to lie above the curve , which is the dashed curve shown in figure [ fig : pieslice ] .
so far , the only information we have used to constrain the mass and radius of the star is the frequency of the higher - frequency qpo in a kilohertz pair .
however , we have the additional information that the _ coherence _ of the higher - frequency qpo is high ( ) in many sources .
let us assume that the higher - frequency qpo in a pair is generated by the same mechanism in all sources that show such pairs .
then the high coherence of these higher - frequency qpos imposes additional constraints on the neutron star s mass and radius because , in order to produce a qpo with such high coherence , the gas that generates the qpo must be in a nearly circular orbit .
hence @xmath342 must be greater than the radius @xmath343 of the innermost stable circular orbit , because gas inside @xmath94 spirals quickly inward to the stellar surface . for a nonrotating star ,
the radius of the innermost stable circular orbit is a function only of the mass of the star and is given by .
inverting this relation gives @xmath344 in the radius - mass plane , relation ( [ eq : msmasszero ] ) is a straight line of slope @xmath345 through the origin and is the dotted line marked @xmath346 in figure [ fig : pieslice ] .
this line intersects @xmath347 at the mass and radius values @xmath348 and @xmath349 if @xmath65 were larger than @xmath350 , the orbital radius @xmath351 of the gas generating the qpo would be less than @xmath352 , so @xmath353 is an upper bound on the mass of the star ; @xmath350 is inversely proportional to @xmath23 but independent of the star s radius and the orbital radius of the gas clumps that are producing the higher - frequency qpo .
similarly , @xmath354 is an upper bound on the radius of the star ; @xmath354 is also inversely proportional to @xmath23 but independent of the star s mass and the orbital radius of the gas clumps that are producing the higher - frequency qpo .
the heavy horizontal line plotted in figure [ fig : pieslice ] shows the upper bound on the mass of a nonrotating star for @xmath355 .
this upper bound is @xmath356 , so the mass of the neutron star in would have to be less than this if it were not rotating . for @xmath357 ,
@xmath358 is 16.0 km , so the radius of the neutron star in would have to be be less than this if the star were not rotating .
suppose now that the frequency @xmath359 of a particular qpo is securely identified as the orbital frequency of gas in the innermost stable circular orbit around a particular neutron star .
then @xmath360 for this qpo , so the representative point of the orbit is at the intersection of the diagonal @xmath346 line and the @xmath347 curve .
the mass of the star is therefore @xmath350 .
hence _ identification of a qpo frequency with the frequency of the innermost stable circular orbit immediately determines the mass of the star_. the radius of the star is not determined by such an identification , but must still be less than @xmath354 . as a specific example , suppose that the 1220 qpo observed in is securely identified as the orbital frequency of gas in the innermost stable circular orbit around this neutron star .
then the mass of this neutron star would be determined as 1.8 , if it were not rotating .
_ these arguments apply to stars with arbitrary spin rates as well as to nonrotating stars _ , although the expressions for @xmath361 and @xmath343 are different for a rotating star and depend on the star s spin rate as well as its mass .
they may be summarized as follows : ( 1 ) stellar radii are excluded , because there is no keplerian orbit with frequency @xmath340 outside a star with such a large radius ; the value of @xmath362 is irrelevant .
( 2 ) if @xmath363 but @xmath364@xmath361 , there is a keplerian orbit with frequency @xmath359 outside the star , but any oscillation produced by gas in this orbit would have a coherence much lower than that observed .
( 3 ) if @xmath363 and @xmath365@xmath361 , there is a keplerian orbit with frequency outside the star and the oscillation produced by gas in this orbit can have the required high coherence .
( 4 ) if the frequency @xmath340 of a particular qpo is securely identified as the orbital frequency of gas in the innermost stable circular orbit around a particular neutron star , the mass of the star is @xmath350 ; its radius must be less than @xmath358 . for a nonrotating star with
, the combinations of stellar mass and radius allowed by condition ( 3 ) are the points in the radius - mass plane above the curve but below .
this region is outlined in figure [ fig : pieslice ] by the heavy solid line .
the point representing the mass and radius of the neutron star in would have to lie in this region if the star were not rotating .
the allowed region collapses to the heavy solid horizontal line if the qpo frequency is identified as the orbital frequency of gas in the innermost stable circular orbit so that condition ( 4 ) applies .
figure [ fig : eosconstraints ] compares the mass - radius relations for nonrotating neutron stars given by five equations of state for neutron - star matter ranging from soft to hard with the regions of the radius - mass plane allowed if @xmath366 is a keplerian orbital frequency , for nonrotating stars and three values of @xmath367 . in order to make possible comparisons with previous studies of neutron star properties ( see ,
e.g , pethick & ravenhall 1995 ) , we show the mass - radius curve given by the early friedman - pandharipande - skyrme ( fps ) realistic equation of state ( friedman & pandharipande 1981 ; lorenz , ravenhall , & pethick 1993 ) . the fps equation of state uses a different approach , but is similar to the softest equations of state permitted by modern realistic models of the nucleon - nucleon interaction ( pandharipande , akmal , & ravenhall 1998 ) .
these equations of state all give maximum gravitational masses of about @xmath356 for nonrotating stars . as an example of the mass - radius curves given by later realistic equations of state , we show the mass - radius curve predicted by the uu equation of state ( wiringa , fiks , & fabrocini 1988 ) .
although it is based on older scattering data , the uu equation of state is similar to the recent @xmath368 equation of state ( akmal , pandharipande , & ravenhall 1998 ) , which is based on the most modern scattering data . like the @xmath369 equation of state ,
the uu equation of state gives a maximum mass of about @xmath370 for a nonrotating neutron star .
as an example of the mass - radius curves predicted by the relatively stiff equations of state typically given by mean field theories , we include the mass - radius curve for the mean - field equation of state of pandharipande & smith ( 1975b ; l in the arnett & bowers [ 1977 ] survey ) .
the maximum mass of a nonrotating star constructed using equation of state l is @xmath371 .
we also show the mass - radius curve given by the very early tensor interaction ( ti ) equation of state of pandharipande & smith ( 1975a ; m in the arnett & bowers [ 1977 ] survey ) and the reid soft - core equation of state of pandharipande ( 1971 ; a in the arnett & bowers [ 1977 ] survey ) .
the maximum masses of nonrotating stars constructed using equations of state m and a are @xmath356 and @xmath372 , respectively .
equations of state a , l , and m are no longer of interest to nuclear physicists and are included here primarily to facilitate comparison with previous work on mass - radius constraints . for a particular source , qpo frequency , and equation of state , the allowed portion of the mass - radius relation is the segment within the pie - slice shaped region analogous to the region in figure [ fig : pieslice ] outlined by the heavy line or , if the qpo frequency is identified as the frequency of the innermost stable circular orbit , simply the horizontal heavy line .
if the mass - radius relation given by a particular equation of state intersects the allowed region defined by the highest qpo frequency seen in a given source , that particular equation of state is viable .
if that equation of state is furthermore the correct equation of state , the mass and radius of the neutron star in the source must correspond to one of the points along the segment of the mass - radius relation that lies within the pie - slice shaped region , so the mass and radius of the star are bounded from above _ and _ from below ; for most equations of state , only a narrow range of radii is allowed . for example
, the highest - frequency qpo so far seen in any source is the 1220 hz qpo observed in ( w. zhang , personal communication ) .
hence , if equation of state m is the correct equation of state , then the mass and radius of the neutron star in would have to satisfy @xmath373 and @xmath374 if it were nonrotating .
if equation of state m is the correct equation of state and 1220 hz is identified as the frequency of the innermost stable circular orbit , then the mass and radius of the neutron star in would be determined as about @xmath356 and between 11 and 12 , if the star were not rotating .
an equation of state that gives a mass - radius relation that does _ not _ intersect the region allowed for a given source is ruled out for that source .
we stress that _ because the equation of state of the matter in neutron stars is expected to be essentially the same in all such stars , an equation of state that is inconsistent with the properties of any neutron star is excluded for all neutron stars_. for example , observation of a 1500 hz keplerian orbital frequency in any source would rule out equations of state l and m. so far our discussion of constraints on the neutron stars in the kilohertz qpo sources has assumed that they are not rotating .
as we discuss in the next subsection , the changes in these constraints caused by the spin of the star are likely to be small in most sources .
rotation affects the structure of the star for a given mass and equation of state and the spacetime exterior to the star .
hence the mass - radius relation , the orbital frequency at a given radius , and the radius @xmath94 of the innermost stable circular orbit are all affected . as a result ,
the bounds on the mass and radius of the star and the constraints on the equation of state implied by observation of a kilohertz qpo of a given frequency are affected .
obviously , the size of these effects depends on the spin rate of the star . treating these effects
accurately will be particularly important if the frequency of a kilohertz qpo is ever securely identified as the keplerian frequency at the marginally stable orbit around a neutron star ( see 5.4 ) , because the ability to rule out some equations of state may depend on it .
the parameter that characterizes the importance of rotational effects is the dimensionless quantity , where @xmath88 and @xmath65 are the angular momentum and gravitational mass of the star .
the value of @xmath90 that corresponds to a given observed spin frequency depends on the neutron star mass and equation of state , and is typically higher for lower masses and stiffer equations of state .
for example , at the 363 hz spin frequency inferred for ( strohmayer et al .
1996c ) , a @xmath101 neutron star with the softer equation of state a has , whereas a @xmath101 neutron star with the very stiff equation of state l has ( see table [ table : jvalues ] ) . as discussed in 5.1 ,
if the frequency of the higher - frequency qpo in a kilohertz qpo pair is an orbital frequency , then the region of the radius - mass plane allowed for the neutron star in that source is bounded below by the curve @xmath375 , which gives the orbital radius at which the keplerian frequency is equal to the highest observed frequency @xmath340 of the higher - frequency qpo .
given @xmath90 , it is straightforward to compute the orbital radius for any given orbital frequency as a function of @xmath65 .
however , it is not @xmath90 but instead the star s spin frequency @xmath83 that is determined by the observations .
hence , _ to obtain the relevant @xmath376 curve one must vary @xmath90 in such a way that @xmath195 is kept constant as @xmath65 is varied_. moreover , @xmath90 depends not only on @xmath195 and @xmath65 but also on the equation of state .
thus the relevant @xmath376 curve depends not only on the star s spin rate , but also on the equation of state assumed .
therefore , for rotating stars ( unlike static stars ) , one can not present a single @xmath375 curve that constrains all equations of state , even for a star with a given spin frequency . as also discussed in 5.1 ,
the region of the radius - mass plane allowed for the neutron star in a given source is bounded above by the horizontal line @xmath377 , which is the mass at which the curve @xmath376 intersects the curve @xmath343 ; the radius @xmath54 of the star is bounded above by @xmath378 .
the two curves @xmath362 and @xmath376 are both affected by rotation of the star , so @xmath379 depends on the stellar spin rate . determining the mass at which is equivalent to determining the mass at which @xmath380 is equal to @xmath340 . here
we focus on the latter condition .
the allowed mass of a rotating star is bounded above if , as @xmath65 is increased at constant @xmath83 , the orbital frequency @xmath381 crosses the qpo frequency @xmath340 from above , once and only once .
as we show below , this is the case for slowly rotating stars .
however , this may not be the case for some neutron star equations of state and spin rates .
( for a kerr black hole with fixed spin frequency , @xmath382 first decreases and then increases with increasing mass . ) fortunately , most of the neutron stars that exhibit kilohertz qpos appear to have spin frequencies in the range 250350 hz ( see 4.2 ) , which is low enough that a first - order treatment of rotational effects is adequate . to see this
, note that neutron stars with masses have moments of inertia , so stars with spin frequencies have @xmath90-values ( see table [ table : jvalues ] ) .
the lowest - order changes in the structure of a rotating star are @xmath383 so , to first order in @xmath90 , the mass - radius relation and moment of inertia of a rotating star are the same as for a nonrotating star of the same mass ( hartle & thorne 1968 ) .
thus the error made by neglecting higher - order terms is for spin frequencies .
we now compute the mass and radius constraints imposed on a slowly rotating neutron star by observation of a kilohertz orbital frequency .
the calculation is simplified by the fact that to first order in @xmath90 , the spacetime outside a uniformly and steadily rotating relativistic star is the same as the kerr spacetime for the same @xmath65 and @xmath90 ( hartle & thorne 1968 ) .
consider first the effect of stellar rotation on the motion of gas orbiting the star . to first order in @xmath90 , the orbital frequency ( measured at infinity ) of gas in a prograde keplerian orbit at a given boyer - lindquist radius @xmath100
is @xmath384\,\nu^\nonrot_k(r , m ) \label{eq : keplerfreqfirst}\ ] ] and the radius of the marginally stable orbit is @xmath385\,r^\nonrot_{\rm ms}(m)\ ; , \label{eq : msradiusfirst}\ ] ] where @xmath386 and @xmath387 are the keplerian frequency and radius of the marginally stable orbit for a nonrotating star .
expressions ( [ eq : keplerfreqfirst ] ) and ( [ eq : msradiusfirst ] ) are first - order expansions of the exact expressions for these quantities given by bardeen , press , & teukolsky ( 1972 ) for the kerr spacetime . hence , to first order in @xmath90 , the frequency of the prograde orbit at @xmath94 around a star of given mass @xmath65 and dimensionless angular momentum @xmath90 is ( cf .
kluniak , michelson , & wagoner 1990 ) @xmath388 \left [ 1+j(2/3)^{1/2 } \right]\nu^0_{\rm k , ms } \nonumber\\ \nonumber\\ & \approx & 2210\,(1 + 0.75j)(m_\odot / m)\,\hz\ ; , \end{aligned}\ ] ] where @xmath389 is the keplerian orbital frequency at the radius of the innermost stable circular orbit for a nonrotating star .
therefore , _ the net effect of the star s rotation is to increase the frequency of the prograde orbit at @xmath94_. as explained above , the region of the radius - mass plane allowed for a given star is not the region allowed for constant @xmath90 but is instead the region allowed for constant @xmath83 . to demonstrate that the mass and radius of the star are bounded from above , it is sufficient to show that @xmath381 decreases with increasing @xmath65 at fixed @xmath83 , or equivalently , that @xmath390 for all @xmath65 for the equation of state under consideration .
now to first order in @xmath90 , @xmath391_{\nu_{\rm spin } } = \quad \left\ { \left[\frac{\pd\nu_{\rm k , ms } } { \pd m}(m,0)\right]_{j } \right .
\nonumber\\ \nonumber\\ \left .
\mbox { } + \left[\frac{\pd \nu_{\rm k , ms } } { \pd j}(m,0)\right]_{m } \left(\frac{dj}{dm}\right ) \right\}_{\nu_{\rm spin}}\\ \nonumber\\ \approx -\ \frac{\nu^0_{\rm k , ms}(m)}{m } \left [ 1 + 0.75j\left(2 - \frac{d\ln i}{d\ln m}\right ) \right]_{\nu_{\rm spin } } . \end{aligned}\ ] ] note that the derivatives on the right side of equation ( [ slope ] ) are to be evaluated at @xmath392 . in order to show that @xmath393 is negative for slowly rotating stars
, it is sufficient ( but not necessary ) to show that @xmath394 is always less than 2 .
this is the case for all the equations of state tabulated by cook et al .
( 1994 ) and is the case even for incompressible matter , which is unphysically stiff .
( for a star made of incompressible matter , @xmath395 , so . )
thus , _ for slowly rotating stars @xmath381 decreases with increasing mass for constant @xmath83 , so the masses and radii of such stars are bounded above_. computation of the mass and radius constraints is straightforward but depends on the stellar spin rate and equation of state .
the upper bounds on the mass and radius are given implicitly by @xmath396m^\nonrot_{\rm max } \label{eq : firstmaxmass}\ ] ] and @xmath397r^\nonrot_{\rm max}\ ; , \label{eq : firstmaxradius}\ ] ] where @xmath398 is the value of @xmath90 for the observed stellar spin rate at the maximum allowed mass for the equation of state being considered and @xmath350 and @xmath354 are the maximum allowed mass and radius for a nonrotating star ( see eqs . [ [ eq : massboundzero ] ] and [ [ eq : radiusboundzero ] ] ) .
expressions ( [ eq : firstmaxmass ] ) and ( [ eq : firstmaxradius ] ) show that _ the bounds are always greater for a slowly rotating star than for a nonrotating star , regardless of the equation of state assumed_. figure [ fig : rotationeffects ] illustrates the effects of stellar rotation on the region of the radius - mass plane allowed for a given star for spin rates @xmath399 , like those inferred for the kilohertz qpo sources , and , the frequency of the highest - frequency qpo so far observed in , which is also the highest - frequency qpo so far observed in any source .
our calculations show that _ the mass of the neutron star in must be less than @xmath400 and its radius must be less than @xmath401_. as explained above , the precise upper bounds depend on the equation of state assumed . for rapidly rotating stars , @xmath90 is not small compared to unity and the structure of the star depends appreciably on its rotation rate .
derivation of bounds on the mass and radius of a given star for an assumed equation of state therefore requires construction of a sequence of stellar models and spacetimes for different masses using the assumed equation of state , with @xmath83 as measured at infinity held fixed .
the maximum and minimum possible masses and radii allowed by the observed qpo frequency can then be determined .
the luminosities of the z sources are typically @xmath402 , where @xmath29 is the eddington luminosity ( see 2 ) .
hence , in the z sources the outward acceleration caused by the radial component of the radiation force can be a substantial fraction of the inward acceleration caused by gravity .
therefore , in these sources the radial component of the radiation force must be included in computing the keplerian orbital frequency near the neutron star and taken into account when constraints on the mass and radius of the star are derived using the procedures discussed in 5.1 and 5.2 .
the radially outward component of the radiation force reduces the orbital frequency at a given radius .
for example , if the star is spherical and nonrotating and emits radiation uniformly and isotropically from its entire surface , the orbital frequency ( measured at infinity ) of a test particle at boyer - lindquist radius @xmath100 is @xmath403^{1/2}\ ; , \label{eq : radforcefreq}\ ] ] where @xmath404 is the luminosity of the star measured at infinity and @xmath405 is the keplerian frequency in the absence of radiation forces .
thus , the boyer - lindquist radius of a circular orbit with a given frequency is smaller in the presence of the radial radiation force and the constraints on the mass and radius of the star are therefore tightened . for the atoll sources , which have luminosities @xmath406 ,
the change in the keplerian frequency is at most @xmath05% . for the z sources , on the other hand , which have luminosities
, the change may be much larger , although the change in the sonic - point keplerian frequency may be smaller than would be suggested by a naive application of equation ( [ eq : radforcefreq ] ) , if a substantial fraction of the radiation produced near the star is scattered out of the disk plane before it reaches the sonic point .
a more detailed analysis of the effect of the radial radiation force on the constraints on the mass and radius of the star will be reported elsewhere . as explained in 4.2 ,
observations of kilohertz qpo sources may be able to demonstrate the existence of an innermost stable circular orbit around some of the neutron stars in these sources .
if so , this would be the first confirmation of a strong - field prediction of general relativity and an important step forward in our understanding of strong - field gravity .
if the frequency of a kilohertz qpo produced by an lmxb can be securely established as the orbital frequency at the radius of the innermost stable circular orbit and the spin frequency of the star can be determined , the mass of the neutron star in that source can be determined for each assumed equation of state .
depending on the range of allowed equations of state and masses , this could have profound consequences for our understanding of the equation of state of neutron stars .
for example , establishing a mass of @xmath407 for a slowly rotating neutron star would rule out eight of the twelve currently viable equations of state considered by cook et al .
( 1994 ) .
given the profound consequences that would follow from identifying the frequency of a kilohertz qpo with the orbital frequency of the innermost stable orbit in any source , it is very important to establish what would constitute strong , rather than merely suggestive , evidence of such a detection .
observations that would signal detection of the innermost stable orbit include the following : _ qpo frequency signature_.the strongest evidence that a qpo with the orbital frequency of the innermost stable orbit has been detected would be reproducible observation of a fairly coherent , kilohertz qpo with a frequency that first increases steeply with accretion rate but then , at a high frequency , becomes nearly constant as the accretion rate continues to increase ( see 3.3 , 4.2 , and fig .
the qpo frequency will approach a constant because the sonic point in the flow can not retreat closer to the star than the radius of the marginally stable orbit and coherent qpos are not expected from the rapidly inspiraling gas inside the sonic point .
although the count rate above which the frequency becomes constant may vary , the constant frequency itself should not vary , because the orbital frequency of the marginally stable orbit depends only on the mass and rotation rate of the neutron star , which remain almost constant over many years . in the sonic - point model ,
the frequency of the lower - frequency qpo in a kilohertz qpo pair is the beat frequency and therefore should also increase steeply with accretion rate at first but then become approximately constant at the same accretion rate at which the frequency of the higher - frequency qpo becomes constant .
_ qpo amplitude signature_.a second signature that the marginally stable orbit has been detected would be reproducible observation of a decrease in the amplitude of the lower - frequency qpo in a kilohertz qpo pair or a simultaneous decrease in the amplitudes of both qpos in a pair , at a qpo frequency that is always the same in a given source .
the amplitude of the qpo at the beat frequency is expected to decrease once the sonic point has moved inward to the marginally stable orbit because in the sonic - point model , the qpo at the beat frequency is generated by the drag force exerted by the radiation coming from the stellar surface that reaches the sonic point .
therefore , a strong qpo is expected at the sonic - point beat frequency only if the drag force exerted by the radiation is dynamically important at the sonic point .
if , however , the sonic point has moved inward to the innermost stable orbit and remains there as the accretion rate continues to increase , the optical depth from the stellar surface to the sonic point will continue to rise and radiation drag will become less and less important there , causing modulation of the inflow from the sonic point by the radiation force to weaken .
thus , a decrease in the amplitude of the qpo at the sonic - point beat frequency with increasing accretion rate would signal the approach of the sonic - point to the radius of the innermost stable orbit ( see 3.1 ) .
if radiation forces play an important role in creating or amplifying the clumps that produce the qpos ( see 3.1 ) , the amplitude of the keplerian - frequency qpo may also decrease with increasing accretion rate , once the radius of the sonic point has reached the radius of the innermost stable orbit ( see 3.2 ) .
_ possible qpo coherence signature_.a possible signature that the orbit of the gas that is generating the qpo has receded inside the radius of the marginally stable orbit would be a steep drop in the coherence of both qpos in a kilohertz qpo pair ( or in the coherence of the keplerian - frequency qpo , if the beat - frequency qpo is not visible ) , at a certain critical frequency , as the frequencies of the qpos increase steadily with accretion rate .
this would occur if radiation forces are able to generate clumping in the flow at a radius inside the sonic point .
if this occurs at all , clumps are likely to be produced at a range of radii and to last only a short time , so any oscillations that may be generated are likely to have low coherence ( see 3.4 )
. the critical frequency would be the orbital frequency at the marginally stable orbit and hence should always be the same in a given source .
approach of the sonic radius to the radius of the innermost stable orbit can be distinguished from approach of the sonic radius to the radius of the star because coherent keplerian- and beat - frequency qpos may continue in the fist case but are very unlikely in the second .
kilohertz qpos are unlikely to be generated if the sonic point moves close to the stellar surface because of the disruptive effect of the stellar magnetic field and the the viscous shear layer that is expected to develop if the keplerian flow interacts directly with the stellar surface ( see 4.2 ) .
the sonic - point model explains naturally the most important features of the kilohertz qpos observed in the atoll and z sources , including their frequencies , large amplitudes , and high coherence .
it also explains the frequent occurrence of two simultaneous qpos , the observed steep increase of kilohertz qpo amplitudes with increasing photon energy in the @xmath0510 kev energy range , and the anticorrelation of kilohertz qpo amplitudes with the strength of the stellar magnetic field inferred from spectral models .
an attractive feature of the sonic - point model is that the magnetic fields , accretion rates , and scattering optical depths that it requires are completely consistent with those inferred previously from observations and modeling of the x - ray spectra and lower - frequency x - ray variability of the kilohertz qpo sources .
( 1 ) _ kilohertz qpos with the properties seen in the atoll and z sources should not be observed in black - hole lmxbs _ , because collision of the accretion flow with the stellar surface plays an essential role in the sonic - point model .
( 2 ) _ the rms amplitudes of the kilohertz qpos should be anticorrelated with the strength of the neutron star s magnetic field_. from this it follows that kilohertz qpos should be very weak or undetectable with current instruments in the gx group of atoll sources and in the cyg - like z sources ( which include 2 , , and ) , and undetectable with current instruments in any sources that have strong magnetic fields and therefore produce strong , periodic oscillations at their spin frequencies .
( after this specific prediction was made in the original version of this paper , kilohertz qpos with very low amplitudes were detected in all six of the originally identified z sources ; see 4.4 . )
another consequence of expectation ( 2 ) is that the positive correlation of rms amplitude with hard color shown by the four sources plotted in figure [ fig : ampvshardcolor ] should be found to be general among sources showing kilohertz qpos .
( 4 ) _ weak oscillations should eventually be detected at overtones of the sonic - point beat and keplerian frequencies and perhaps at the stellar spin frequency , but oscillations at other frequencies , such as @xmath408 , should be extremely weak . _ at a lower level of certainty , the sonic - point model suggests that ( 1 ) the dependence of qpo amplitude on photon energy will change as the x - ray spectrum of the source changes , ( 2 ) the amplitude of the higher - frequency qpo in a kilohertz qpo pair will drop relative to the amplitude of the lower - frequency qpo at high luminosities , and ( 3 ) either the lower - frequency qpo or the higher - frequency qpo in a pair may be undetectable even when the other qpo is detectable . in closing , we emphasize that measurement of keplerian frequencies in the kilohertz range provides interesting new upper bounds on the masses and radii of the neutron stars in the kilohertz qpo sources and important new constraints on the equation of state of the matter in all neutron stars , as we have shown in 5 .
as we demonstrated there , if the neutron star in the atoll source has a spin frequency of @xmath409 hz , as indicated by the difference between the spin frequencies of its two high - frequency qpos , the 1220 hz qpo observed in this source constrains its mass to be less than about @xmath370 and its radius to be less than about 17 km ; the precise bounds depend on the equation of state assumed .
if at some future time we are able to demonstrate the existence of an innermost stable circular orbit around one or more neutron stars using observations of kilohertz qpos , this would be the first confirmation of a prediction of general relativity in the strong - field regime and a major advance in our understanding of strong - field gravity . if the frequency of a kilohertz qpo is securely established as the orbital frequency at the radius of the innermost stable circular orbit in a particular system by , for example , observing the signatures discussed in 5 , and
if , in addition , the spin frequency of the neutron star can be determined , then the frequency of the qpo will fix the mass of the neutron star for each assumed equation of state , providing a better understanding of the properties of dense matter and possibly ruling out many currently viable equations of state .
it is a pleasure to thank phil kaaret , vicky kalogera , ed morgan , michiel van der klis , tod strohmayer , tom baumgarte , stu shapiro , and tomek bulik for useful discussions .
we especially thank rudy wijnands for providing the pca response matrix and the x - ray spectral data used to compute the x - ray colors plotted in figures 11 and 12 , greg cook for providing the mass - radius relations for the equations of state plotted in figure [ fig : eosconstraints ] , and ron taam for discussions about black hole disk solutions . this work was supported in part by nsf grants ast 93 - 15133 and ast 96 - 18524 , nasa grant nag 5 - 2925 , and nasa rxte grants at the university of illinois , nasa grant nag 5 - 2868 at the university of chicago , and through the fellowship program , by nasa grant nas 5 - 2687 .
berger , m. , van der klis , m. , van paradijs , j. , lewin , w.h.g . , lamb , f.k . ,
vaughan , b. , kuulkers , e. , augusteijn , t. , zhang , w. , marshall , f.e . ,
swank , j.h . ,
lapidus , i. , lochner , j.c . , & strohmayer , t.e .
1996 , apj , 469 , l13 bradt , h. , shirey , r. , & levine , a. 1998 , in the active x - ray sky : results from bepposax and rossi - xte , nuclear phys .
b proceedings supplements , ed .
l. scarsi , h. bradt , p. giommi , & f. fiore , in press ( amsterdam : north holland ) pandharipande , v.r . , akmal , a. , & ravenhall , d.g .
1998 , in nuclear astrophysics , proc .
international workshop xxvi on gross properties of nuclei and nuclear excitations , ed .
m. buballa , n. nrenberg , j. wambach , & a. wirzba ( darmstadt : gsi ) , 11 4u 0614@xmath410091&atoll&400600&@xmath411&ford et al . 1996 , 1997a , 1997b
+ & & 5001145&&van der klis et al .
1996c + & & 327&&mndez et al .
1997 + & & 630 , 727&16.5 , 15.8 & + 4u 1608@xmath30952&atoll&650890&514&berger et al .
1996 + & & 9401125&&mndez et al .
1998 + & & 570800&&yu et al .
1997 + 4u 1636@xmath309536&atoll&840920&6.0 & zhang et al .
1996 , 1997a + & & 11501220&6.6 , 6.1&van der klis et al .
1996c + & & 580@xmath412&&wijnands et al .
1997b + 4u 1728@xmath30934&atoll&637716&5.26.9&strohmayer et al .
1996a , 1996b , + & & 5001100&5.58.1&1996c , 1997a + & & 363@xmath412 & 1.55 + ks 1731@xmath309260&atoll&524@xmath412&12&morgan & smith 1996 + & & 900 , 11701207&45&smith et al .
1997 + & & & & wijnands & van der klis 1997 + 4u 1735@xmath309444&atoll&1149&3.1&wijnands et al .
1996 + 4u 1820@xmath30930&atoll&546796&3.25.0&smale et al .
1996 , 1997 + & & 1066 & & + aql x@xmath3091&atoll&750830&&zhang et al . 1998a + & & 549@xmath412 & & + cyg x-2&z&730 1020 & 35 & wijnands et al .
1998 + & & 490530 & & + gx 5@xmath3091&z&567895&2.06.7&van der klis et al .
1996e + & & 325448 & & + gx 17@xmath4102&z&470780&35&van der klis et al .
1997a + & & 6451087&&wijnands et al .
1997c + gx 340@xmath41002&z&247 625&2.5&jonker et al . 1998 + & & 625 820&2 - 5 & + gx 349@xmath4102&z&712&1.2&zhang , strohmayer , & swank 1998b + & & 978&1.3 & + sco x-1&z&570830&0.91.2&van der klis et al . 1996a,1996b + & & 8701130&0.60.9&1996d , 1997b + unknown&unknown&589@xmath412&24&strohmayer et al . 1996d + @xmath413complete as of 1998 may 31 .
@xmath412burst oscillation .
@xmath107&@xmath416&& + @xmath104&@xmath417&& + @xmath418&&@xmath419& + @xmath200&&@xmath420& + @xmath83&&@xmath421& + @xmath311&&@xmath421& + @xmath312&&&@xmath422 + @xmath423&&&@xmath424 + @xmath425&&&@xmath422 + @xmath426&&&@xmath424 + 4u 0614@xmath410091&@xmath10330&&ford et al . 1997a , 1997b + 4u 1608@xmath30952&230290&&mndez et al .
1998 + 4u 1636@xmath309536&@xmath10290&@xmath10580&zhang et al . 1996 , 1997 + 4u 1728@xmath30934&@xmath10360&@xmath10363&strohmayer et al .
1996a , 1996b , 1996c + ks
1731@xmath309260&@xmath10260&@xmath10520&smith et al . 1997 , wijnands & van der klis 1997 + 4u 1735@xmath309444&&& + 4u 1820@xmath30930&@xmath10275&&smale et al .
1997 + aql x@xmath3091&&@xmath10550&zhang et al .
1998a + cyg x-2&@xmath10345&&wijnands et al .
1998 + gx 5@xmath3091&@xmath10325&&van der klis et al .
1996e + gx 17@xmath4102&@xmath10295&&wijnands et al .
1997c + gx 340@xmath4100&@xmath10325&&jonker et al . 1998 + gx
349@xmath4102&@xmath10266&&zhang et al .
1998 + sco x-1&@xmath10250300&&van der klis et al . 1996a , 1996b , 1996d , 1997b + unknown&&@xmath10590&strohmayer et al .
1997a + @xmath413complete as of 1998 may 31 . here
@xmath6 is the difference between the frequencies of the kilohertz qpos seen simultaneously in the persistent x - ray emission of the source and @xmath427 is the frequency of the brightness oscillation seen during type i ( thermonuclear ) x - ray bursts from the source .
|
quasi - periodic brightness oscillations ( qpos ) with frequencies ranging from @xmath0300 hz to @xmath01200 hz have been discovered in the x - ray emission from fourteen neutron stars in low - mass binary systems and from another neutron star in the direction of the galactic center .
these kilohertz qpos are very strong , with rms relative amplitudes ranging up to @xmath015% of the total x - ray countrate , and are remarkably coherent , with frequency to fwhm ratios as large as @xmath0200 .
two simultaneous kilohertz qpos differing in frequency by @xmath0250350 hz have been detected in twelve of the fifteen sources .
here we propose a model for these qpos . in this model
the x - ray source is a neutron star with a surface magnetic field @xmath1
g and a spin frequency of a few hundred hertz , accreting gas via a keplerian disk .
some of the accreting gas is channeled by the stellar magnetic field but some remains in a keplerian disk flow that penetrates to within a few kilometers of the stellar surface .
the frequency of the higher - frequency qpo in a kilohertz qpo pair is the keplerian frequency at a radius near the sonic point at the inner edge of the keplerian flow whereas the frequency of the lower - frequency qpo is approximately the difference between the keplerian frequency at a radius near the sonic point and the stellar spin frequency .
the difference between the frequencies of the pair of qpos is therefore close to ( but not necessarily exactly equal to ) the stellar spin frequency .
the amplitudes of the qpos at the sonic - point keplerian frequency and at the beat frequency depend on the strength of the neutron star s magnetic field and the accretion rate and hence one or both of these qpos may sometimes be undetectable .
oscillations at the stellar spin frequency and its overtones are expected to be weak but may sometimes be detectable .
this model is consistent with the magnetic field strengths , accretion rates , and scattering optical depths inferred from previous modeling of the x - ray spectra and rapid x - ray variability of the atoll and z sources .
it explains naturally the frequencies of the kilohertz qpos and the similarity of these frequencies in sources with different accretion rates and magnetic fields .
the model also explains the high coherence and large amplitudes of the kilohertz qpos , and the steep increase of qpo amplitude with photon energy .
the increase in qpo frequency with inferred accretion rate seen in many sources is also understandable in this model .
we show that if the frequency of the higher - frequency qpo in a pair is an orbital frequency , as in the sonic - point model , the frequencies of these qpos place interesting upper bounds on the masses and radii of the neutron stars in the kilohertz qpo sources and provide new constraints on the equation of state of matter at high densities .
further observations of these qpos may provide compelling evidence for the existence of a marginally stable orbit , confirming a key prediction of general relativity in the strong - field regime .
# 1 10^#1 @xmath2 # 1_#1 @xmath3 # 1^#1 @xmath4 * = " 002a # 1#1 ps . # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1
|
early astronomers looked to the sky and saw the moon as a habitable world covered in vast oceans , venus as a swampy marshland enshrouded in clouds , and mars with grand canals ( lowell 1895 ) .
not one of these worlds has maintained its promise of abundant life .
instead , the solar system , once thought to be teeming with life , may be barren , although hope remains for environments under the icy crust of europa ( marion et al .
2003 ) , in the tiger stripes of enceladus ( parkinson et al .
2007 ) , in water under the martian surface ( malin & edgett 2000 ) , or perhaps lurking somewhere as yet unidentified . with the discovery of more than 700 extrasolar planets since 1989 , the real estate market in our solar system is no longer the only place we might look for evidence of life beyond earth .
a likely place to search for planets harboring life is in the habitable zones of nearby stars .
the term `` habitable zone '' was first coined by huang ( 1959 ) as a region around a star where a planet could support life . since then , there have been many definitions of habitability , most based on the presence of liquid water on the surface of a planet .
these planets could be terrestrial in nature , which are known to exist ( borucki & the kepler team 2010 ) , or perhaps moons of gas - giant planets , which are suspected to exist ( weidner & horne 2010 ) .
examples of nearby stars hosting potentially habitable super - earths include gj 581 , an m dwarf with potentially two planets in its habitable zone ( mayor et al .
2009 ; vogt et al .
2010b ) and gj 667c , another m dwarf with one planet in its habitable zone ( anglada - escud et al .
2012 ) . kasting et al .
( 1993 ) did pioneering work in describing habitable zones ( hzs ) around main sequence stars . to approximate the location of the hz
, they introduced a one - dimensional climate model that yields the distances from main sequence stars where liquid water would be present , given an initial assumption of a @xmath5/@xmath6/@xmath7 atmosphere and an earth - sized planet .
they describe the inner boundary of their model as the point at which the atmosphere becomes saturated with @xmath6 , causing a loss of water via photolysis and hydrogen escape ; the outer boundary is marked by the formation of @xmath5 clouds that cool a planet s surface by increasing its albedo and lowering its convective lapse rate .
they give an equation for the distance from a star , _
d _ , of the hz in au based on the incident flux that a planet receives , @xmath8 , the star s luminosity relative to that of the sun , and @xmath9 , the ratio of outgoing ir flux to the incoming incident flux at the top of the planet s atmosphere : @xmath10 using more explicit terms , equation [ eqno1 ] can be used to show the distance from a star at which a planet would have a given temperature , based on energy balance ( kaltenegger et al .
equation [ eqno2 ] can then be used to show that the equilibrium temperature , @xmath11 , of a planet at a distance , @xmath12 , from its host star is a function of the stellar effective temperature , @xmath13 , stellar radius , @xmath14 , and the planet s bond albedo , @xmath15 .
thus , if the stellar @xmath13 and @xmath14 are known , we can calculate the range of distances where a planet with given albedo 0.3 for the hypothetical planets used in this paper . ] would have a surface temperature suitable for liquid water , as described by the planet s temperature , @xmath11 : @xmath16 in this paper , we apply this equation to the nearby population of stars whose stellar properties are well known to estimate the total habitable real estate that they provide .
we use the complete sample of all stars currently known to be within 5 pc of the sun from henry ( 2012 ) , and an extended 10 pc sample of afgk stars , as well as binary properties , photometric and astrometric data .
we present our derived @xmath13 and @xmath14 for each star in the sample and discuss the methods used to derive each star s empirical habitable zone ( ehz ) . for multiple star systems ,
we assess dynamical stability to eliminate stellar systems unsuitable for long - term planetary orbits .
our main goal is to determine , as a function of spectral type , the cumulative available ehz in the solar neighborhood .
we do not yet have a detailed understanding of the architectures of all types of planetary systems orbiting various types of stars , although there has been recent progress for planets in close to stars via the kepler mission ( e.g. , howard et al .
2012 for planets in orbital periods less than 50 days around fgkm stars ) . in this paper
we assume no bias in the final locations of planets around stars , including formation and migration ( i.e. , we assume the distribution of planets to be uniform in semi - major axis ) , to assess the integrated ehzs of various types of stars found in the solar neighborhood .
we focus on the presence of the _ first _ habitable planet around a given star , although wider ehzs may of course include more than one planet .
this is an important assessment to make given the limited telescope time , funding , and energy of astronomers , so here we focus on the question , `` what set of targets might be most appropriate to observe to improve the odds of detecting at least one habitable planet ? ''
this question is posed in wide - ranging arenas , from conversations between students and faculty when developing research projects , to discussions of programmatic directions such as those of nasa s exoplanet program analysis group and the nasa / nsf exoplanet task force . in this paper
we evaluate stars within 10 pc in a consistent fashion to assess what spectral types offer the greatest promise for nearby habitable planets .
the nearby sample contains stars that have been characterized carefully , and as a population provide the most accurate snapshot of the stellar content of the galaxy .
thus , this sample provides the foundation for a realistic understanding of the relative merits of examining different types of stars .
our results , coupled with the planet frequency statistics from the kepler mission , ( e.g. , borucki and koch 2011 , howard et al .
2012 , and dressing and charbonneau 2013 ) , can provide statistical measures for the of number of habitable planets within larger stellar populations , and particularly among volume - limited samples of nearby stars , as we build to a comprehensive sample of the nearest stars ( e.g. , henry et al .
2006 ; henry 2012 ) .
for example , howard et al . (
2012 ) show that the number of super - earth size planets increases with decreasing t@xmath17 for orbital periods less than 50 days . dressing and charbonneau ( 2013 )
show that for cool stars ( t@xmath17 @xmath18 4000k ) the occurrence rate for planets with 0.54 r@xmath19 is 0.9 planets per star for orbital periods less than 50 days .
they also calculate that the lower limit on the occurrence rate of earth - size planets in the hz of cool stars is 0.04 with a 95% confidence . using the population we describe here
, this implies that there could be two earth - size planets within the ehzs of m dwarfs within 5 pc of the sun . assuming a constant density of m dwarfs to 10 pc ( not all of which have yet been identified ) the number of earth - size planets jumps to 16 .
one of the primary motivations of this study is to determine , in aggregate , how the odds of finding such planets around m dwarfs compares to other spectral types .
the results of our habitable real estate calculations , as outlined in this paper , are particularly valuable in highlighting that the ubiquitous m dwarfs provide many locales meeting the canonical definition for habitability .
current searches for habitable worlds orbiting the nearest stars have yielded many jovian planets and a few terrestrial worlds , but most of the searches to date have been carried out using the radial velocity technique and have focused on bright fgk stars that provide many photons to spectrographs ( and which are in the sweet spot for kepler ) . for transit searches ,
m dwarfs provide higher contrast ratios for a given planet , whereas for radial velocity and astrometric searches they provide larger wobbles for a planet than do more massive fgk stars . knowing that the m dwarfs , as a group , are important stars to search for habitable worlds helps direct our focus back to the solar neighborhood , in which three - quarters of all stars are red dwarfs . because of their proximity , these stars hold great promise for the detailed characterization of exoplanets .
primary science goals of the research consortium on nearby stars ( recons ) and the center for high angular resolution astronomy ( chara ) , are to find and determine the fundamental properties of our nearest stellar neighbors .
this has led to the most complete assessment to date of stars within 5 pc of the sun . the modern sample of all known stars and brown dwarfs within 5 pc of the sun ( henry 2010 , 2012 ) , listed in table [ tab : parx ] ( assembled photometry in table [ tab : phot ] )
was first published in the _ observer s handbook 2010_. the sample , which is updated yearly , was created using the combination of several ground - based and space - based parallax programs , including the general catalogue of trigonometric stellar parallaxes ( van altena et al .
1995 ) , hipparcos ( perryman et al .
1997 ) , recons ( henry et al .
1997 , 2006 ; deacon et al . 2005 ; jao et al . 2005 ; costa et al .
2005 ) , hst ( benedict et al . 1999 , 2002 )
, gatewood ( 1989 , 1994 ) and gatewood et al .
( 1992 , 1993 ) . to be included in the sample ,
a system must have a weighted mean trigonometric parallax measurement of 200 mas or greater with an error of 10 mas or less . to create this sample
the parallaxes compiled were combined and weighted based on the individual measurement errors .
the spectral type , with reference , for each star or star system is given with the astrometric data , including ra , dec , the weighted mean trigonometric parallax , and the number of parallaxes included in the weighted mean , in table [ tab : parx ] . the closing date for the sample is 2012.0 .
the 5 pc sample contains 67 stars , including the sun , and 4 presumed brown dwarfs ( spectral types l or t ) in 50 systems .
of the 50 systems , 34 are single , 11 are binaries , and 5 are triples , giving a multiplicity fraction of 32% .
this fraction is consistent with previous volume limited surveys ( 35% ; reid & gizis 1997 ) .
the majority ( 85% ) of the sample are main sequence stars ; the exceptions include procyon ( gj 280a ) , which has a slightly evolved spectral type of f5iv - v , 5 white dwarfs , and the 4 l / t dwarfs .
the spectral type breakdown includes 1 a , 1 f , 3 g s ( including the sun ) , 7 k s , 50 m s , 1 l , 3 t s , and 5 white dwarfs .
we do not include white dwarfs and brown dwarfs in subsequent calculations of habitable real estate , as they are objects that are cooling continually , resulting in unsustainable hzs on long ( @xmath20gyr ) timescales . due to the sparse population of all but the m stars in the 5 pc sample , it is difficult to draw meaningful statistics to estimate which stellar spectral types possess the most habitable real estate .
therefore , we extend our sample to 10 pc for a , f , g and k stars . using the hipparcos catalog , which is complete to v=9 ( perryman et al .
1997 ) , we selected all objects with a parallax greater than 100 mas for inclusion into the sample .
we expect that the extended 10 pc sample is complete for spectral types a through k , given m@xmath21=9.0 for a m0.0v star ( henry et al .
however , rather than using spectral type as a selection criterion , we used a color cutoff of @xmath0 @xmath22 3.5 as the dividing line between k and m dwarfs ( kenyon & hartmann 1995 ) .
as with the 5 pc sample , weighted mean parallaxes from the general catalogue of trigonometric stellar parallaxes , the hipparcos catalog and other sources are listed with the astrometry data for the extended 10 pc sample in table [ tab:10pca ] , as well as spectral types and references .
available photometric data are listed in table [ tab:10pcphot ] .
we note that there are no o type or b type stars within 10 pc . because the m star population is not complete out to 10 pc ( henry et al .
2006 ) , we approximate the total m star population by scaling by a factor of eight from the 5 pc sample . for clarity , we refer to the additional afgk stars from 5 - 10 pc as the `` extended 10 pc sample '' and our estimates of all afgkm stars within 10 pc as the `` estimated 10 pc sample '' . in total , the stellar population of the estimated 10 pc sample consists of 66 afgk stars in 57 systems .
broken down by spectral type , the sample contains 4 a stars , 6 f stars , 21 g stars , 35 k stars and an estimated 400 m stars within 10 pc .
the @xmath23 photometry used in this paper , listed in tables [ tab : phot ] & [ tab:10pcphot ] , was extracted from sources in the literature with preference given to large surveys and measurements consistent with photometry obtained as part of the ctio parallax ( ctiopi ) program , which uses the johnson - kron - cousins system . @xmath24 and @xmath25 magnitudes from the usno - b1.0 catalog ( monet et al .
2003 ) are incorporated in the extended 10 pc sample and both are rounded to the nearest 0.1 mag .
@xmath24 magnitudes are averaged from the first and second epochs .
the majority of @xmath26 photometry is taken from the 2mass database , identified as those values with errors listed explicitly in tables 2 and 4 ( cutri et al .
2003 ) . in cases of close binaries with magnitude difference measurements in the literature ( e.g. , henry & mccarthy 1993 and henry et al .
1999 ) , the optical and 2mass photometry is used with the published magnitude differences to split the component fluxes into individual magnitudes .
note that stars brighter than @xmath205 mag are saturated in 2mass images and typically have relatively large photometric errors ( @xmath27 0.2 mag ) .
where possible , we use @xmath26 measurements in the literature for these stars ( johnson et al . 1966 ;
johnson et al . 1968 ; glass 1975 ; mould & hyland 1976 ) .
these magnitudes , listed in their unconverted form in tables [ tab : phot ] & [ tab:10pcphot ] and with specific references listed in the error columns , were then converted to 2mass magnitudes using color transformations from carpenter ( 2001 ) .
of the 16 multiple systems in the 5 pc sample , eight have spatially unresolved photometry in some or all passbands and are marked as `` joint '' in the notes section of table [ tab : phot ] .
similarly , 10 of the 19 multiple systems in the extended 10 pc sample have spatially unresolved photometry in some or all passbands and are listed as `` joint '' in the notes section of table [ tab:10pcphot ] .
the hz around a star is primarily a function of the total energy output of a star that reaches the surface of a planet .
this can be determined if the stellar temperature and radius are known , which we calculate for our sample of stars as described in section [ sed ] . however , a range of other factors , such as atmospheric pressure , composition , and cloud cover play roles in determining the surface temperature of a planet . to account for this , in section [ ehz ]
we follow previous studies and approximate these effects by using empirical temperature constraints provided by planets in the solar system .
although spectral types are often used to estimate stellar effective temperatures , the various methods for determining spectral types are inhomogeneous and often depend upon the spectral range used . as an alternative
, we fit synthetic stellar spectra to broad - band energy distributions to determine effective temperatures and radii .
in particular , we use photometric measurements spanning @xmath28 ( 0.3 to 2.4 @xmath29 m ) in conjunction with model spectra generated using the phoenix code ( hauschildt et al .
this prescription is only used for single stars and for stars in multiple systems in the 5 pc and extended 10 pc samples with at least four spatially resolved photometric measurements in different filter bandpasses .
estimates for unresolved multiples are discussed in section [ bhz ] .
these available models span the temperature ranges of @xmath13 of 10,000 k to 7000 k in 200 k increments and 6900 k to 2000 k in 100 k increments .
the model spectra have a resolution of 2 and range from 10 to 500 @xmath29 m .
the models were convolved with @xmath28 filter responses to create synthetic photometry . for the @xmath23 synthetic photometry
, we took zero points from bessell et al .
( 1998 ) and filter responses from ctio . @xmath26
filter responses and zero points were from 2mass .
the zero points adopted are listed in table [ tab : zp ] .
we adopt log @xmath30 values of 4.5 or 5.0 for all stars .
solar metallicity is adopted for all stars except gj 191 ( fe / h = @xmath310.98 ; woolf and wallerstein 2005 ) and gj 451 ( fe / h = @xmath311.16 ; valenti and fisher 2005 ) , for which we adopt metallicities of @xmath311.0 .
the assumption of solar metallicity for the remainder of our 10 pc sample is based on the 36 fgk stars that overlap with valenti and fisher ( 2005 ) .
these stars have an average metallicity of fe / h = @xmath310.048 with a standard deviation of 0.168 dex .
the flux values in the model spectra are given as a surface flux that must be scaled by a radius to a known distance for fitting with observed integrated flux values .
a stellar radius grid from 0.001 r@xmath32 to 3.00 r@xmath32 with a step size of 0.001 r@xmath32 is calculated for each model spectrum . each spectrum and radius combination is then convolved with filter response and the zero point data referenced above to derive consequent fluxes observed at earth .
these integrated fluxes are then fit via a @xmath33 minimization routine , written in idl and described in equation [ eqno5 ] , to compare the model flux with the observed flux across each filter bandpass . here
, @xmath34 is the observed integrated flux from the photometric measurements , @xmath35 is the estimated integrated flux from the model grids , @xmath36 is the number of photometric points ( degrees of freedom ) , and @xmath37 is the average error in the photometric measurement
. examples of fits for afgkm stars in the 5 pc sample are shown in figure [ fig : fits ] .
@xmath38 the output is an effective temperature from the model and a radius that best fits the measured photometric data .
we assume no interstellar extinction for our sample , and choose four photometric points as the minimum number needed to make a fit . when deriving radii , all stars in the sample are assumed to be spherical and radiate isotropically , which is not the case for rapidly rotating stars that may be oblate and experience gravity darkening ( e.g. , altair , see monnier et al . 2007 ;
vega , see aufdenberg et al .
this is most common among stars earlier than mid - f spectral type , as these stars are fully radiative and consequently do not possess an efficient rotational braking mechanism ( e.g. , wilson 1965 ) .
however , all but one early type star ( i.e. , earlier than spectral type f5 ) within 10 pc have projected rotational velocities less than 100 km / s , which correspond to projected oblateness values @xmath392% ( absil et al .
these apparently slow rotating stars include sirius a ( gj 244 , a1v , 16 km / s ; royer 2002 ) , procyon a ( gj 280 , f5v - iv , 4.9 km / s ; fekel 1997 ) , vega ( gj 721 , a0v , 24 km / s ; royer 2007 ) , and fomalhaut ( gj 881 , a4v , 93 km / s ; royer et al .
we note that despite having a small @xmath40sin@xmath2 value , vega is believed to be rapidly rotating with a nearly pole on orientation ( aufdenberg et al .
an additional rapidly rotating star is altair ( gj 768 , a7v , 217 km / s ; royer 2007 ) which has an oblateness of 18% determined from interferometric measurements by the chara array ( monnier et al .
2007 ) . as expected , our derived radii and effective temperatures for altair and vega are intermediate to the polar and equatorial values listed in monnier et al .
( 2007 ) and aufdenberg et al .
( 2006 ) , respectively .
because the ehz boundaries are a function of the square root of the total luminosity ( see section [ ehz ] ) , our adopted methodology should yield reliable estimates of the habitable real estate these stars provide , even if the stellar temperatures vary with stellar latitude . as a consequence of the limited temperature resolution of the phoenix grids used ( @xmath41t=200k from 10000k to 7000k and
@xmath41t=100k from 6900k to 2000k ) , the fitting routine can determine a slightly larger radius coupled with a cooler temperature , or vice versa .
systematic uncertainties in the phoenix models are such that a @xmath41 t of less than 100k are unreliable ( hauschildt , priv .
as the stefan - boltzmann law shows r@xmath42@xmath43t@xmath17@xmath44@xmath20l@xmath42 , the luminosity determined from the sed remains the same as long as the radii and effective temperatures move in opposite directions , and therefore the ehz , being a function of the square root of the luminosity , does not change . of key importance , the output radii and temperatures determined here allow us to compare our results directly to those found via interferometric techniques . of the stars investigated , 28 have spatially resolved angular diameters from long baseline interferometric instruments such as chara , pti , and vlti ( table [ tab : temp ] ) .
fifteen of these measurements are from the recent effort of boyajian et al .
( 2012 ) . because these measurements determine sizes to within a few percent , we use them to test the accuracy of the radii determined in our fitting process .
the published radii are on average 7.4% larger than our derived radii , which shows a systematic effect .
our derived model @xmath13 are on average 2.4% hotter than derived from interferometric measurements .
a comparison of our model versus published values is shown in figure [ fig : lumz ] .
the temperature uncertainties correspond to spectral type uncertainties of 1 - 2 spectral subclasses , similar to the error associated with most spectral classification methods . as the average values indicate , the overprediction of temperature leads to the expected under - prediction of radius , and the combination yields a more accurate luminosity , and thus consistent ehz . given this agreement ,
we adopt our model values for t@xmath17 and r / r@xmath32 in final calculations of the habitable real estate .
a few stars are worthy of note .
procyon is slightly evolved ( f5.0iv - v ) and has a highly constrained log @xmath30 of [email protected] ( fuhrmann et al .
1997 ) measured via high resolution spectroscopy in concert with masses determined using astrometry from the procyon - white dwarf orbit .
we adjusted the log @xmath30 to 4.0 for this star and recomputed the @xmath13 and @xmath24 .
this , nevertheless , yielded identical values within the model grid resolution as those of log @xmath30=4.5 .
the model spectra used in this work have a low temperature limit of 2000 k , which is the value derived for the three intrinsically faintest stars in this sample : scr 1845 - 6357a ( m8.5v ) , den 1048 - 3956 ( m8.5v ) , and lp 944 - 020 ( m9.0v ) . kasting et al .
( 1993 ) use a one - dimensional climate model to calculate hzs around single main sequence stars .
a one - dimensional climate model characterizes a planet s global temperature by dividing the planet into latitudinal bands , and treats the planet as uniform with respect to longitude .
these one - dimensional radiative - convective models are a good approximation of global temperature , but more complicated 3-d global climate models are needed to account for the complex physical interactions associated with oceans , clouds , and land surface processes .
these inputs add parameters such as land / ocean surface coverage and clouds , which can vary from planet to planet , complicating the overall goal to characterize the ehz of a star . here
we use the one - dimensional model to generalize the ehz , and adopt a modified version of the one - dimensional model based on the `` venus and early mars criterion '' from selsis et al .
( 2007 ) . in that work
, they argue that empirical evidence shows that venus has not had water on its surface for at least one billion years , and mars had water on its surface around 4 billion years ago .
the solar fluxes at those times were 8% and 28% lower , respectively ( baraffe et al .
venus ( 0.72 au today ) and mars ( 1.52 au today ) would need to be at distances of @xmath200.75 au and @xmath201.77 au , respectively , to receive these levels of solar flux today .
selsis et al .
( 2007 ) provide a method for estimating the inner and outer edges of hzs for stars with t@xmath17 = 3700k-7200k .
the stars in the sample discussed here range in temperatures from 2000k to 10000k . we have therefore chosen to derive new relations for
the hz boundaries that span the entire stellar temperature regime of our sample , and thereby provide a consistent methodology for all stars in the sample . in defining our ehz inner and outer boundaries , we assume a planet with an atmosphere , radius , mass , and bond albedo ( 0.3 ) that matches earth ( kasting 1996 ) .
this leads to the ehz equations [ eqno3 ] and [ eqno4 ] , used to determine the empirical surface temperature of an earth - like planet that satisfies the `` venus criterion '' ( we adopt 0.80 au on the suggestion of kasting , priv .
comm . ) and `` mars criterion '' ( 1.77 au ) .
the resulting inner and outer radii of the ehz correspond to equilibrium temperatures of 285 k to 195 k , respectively .
@xmath46 @xmath47 note that these only depend on r@xmath42 and t@xmath48 , or essentially the square root of the star s luminosity .
values for both samples are listed in tables [ tab : hz ] and [ tab:10pchz ] .
only stars used in the cumulative ehz calculations are listed , except for the cases of sirius ( gj 244a ) and procyon ( gj 280a ) , which provide useful benchmarks .
although stars in multiple systems are often avoided in planet searches , planets have been found in binary and multiple systems ( e.g. patience et al 2002 ; raghavan et al .
2006 ; eggenberger et al .
the potentially dynamically disruptive effects of any close stellar companion must be considered when assessing the possibility of formation , long - term dynamical stability , and ultimately the habitability of planets in multiple star systems . the @xmath49 centauri triple system , at a distance from the sun of 1.34 pc for the g2v - k0v pair ( gj 559 ab ) and 1.30 pc for the wide m5v companion ( gj 551 ) , includes the nearest set of sun - like stars , and provides a test case to study the habitability of multiple stars .
the g - k pair has an orbit with a semimajor axis of 17@xmath5057 and eccentricity of 0.518 ( pourbaix et al .
this gives periastron and apastron distances of 11.33 au and 26.67 au , respectively .
barbieri et al . (
2002 ) show for @xmath49 centauri a that planets can form in stable orbits on a timescale of 5 myr . using numerical simulations , they show that not only could planets form , but in some models they formed directly in the habitable zone .
quintana et al .
( 2007 ) similarly find that binary separations greater than 10 au did not inhibit the formation of terrestrial planets at 2 au .
consistent with this , wiegert & holman ( 1997 ) show that the orbit of a planet can be stable as long as the ratio of the semimajor axis of the binary to that of the planet is more than 5:1 .
the 36 multiple systems in the 5 pc and extended 10 pc samples consist of 27 binaries , 6 triples , and 3 quadruple systems ( gj 423 , gj 570 , and gj 695 ) , for a total of 84 stars / brown dwarfs .
however , nine are not main sequence stars , and excluded because of their evolutionary states ( e.g. , sub - giants , white dwarfs , and brown dwarfs ) .
additionally , 16 are m star companions in the extended 10 pc afgk sample . because the habitable real estate of m stars out to 10 pc
is estimated by scaling the 5 pc results , we do not consider these m stars in our ehz calculations .
this leaves 59 stars in multiple systems with potential habitable zones . for clarity
, we emphasize that the ehz of each star in a system is calculated separately . of the 59 stars , 43 ( 73% ) have at least 4 spatially resolved photometric measurements in different filter bandpasses .
for these stars , the same prescription used in section [ ehz ] is used to calculate the ehzs .
for the 16 stars that are not photometrically spatially resolved from their nearest companion(s ) , we estimate the ehz locations based on spectral type information for the components . using the assembled spectral types listed in tables [ tab : parx ] & [ tab:10pca ]
, we estimate the components @xmath0 color using the spectral type versus @xmath0 colors relations of kenyon & hartmann ( 1995 ) , and estimate the location of the ehzs using the relations described in section [ vmk ] .
to assess the dynamical stability of any planets in these ehzs , we compare the locations of the outer ehz boundaries to the binary separations listed in table [ tab : bin ] .
we consider a planet to be dynamically stable if the periastron distance of a star s nearest companion is greater than five times the outer ehz boundary . in cases where the periastron distance can not be calculated ( because its full orbit solution is not known ) , we use the projected separation . fortunately
, these exceptions are all large separation multiples , and thus minor errors in the adopted separation are likely irrelevant for dynamical stability considerations
. the ratios of periastron distances to outer ehz boundaries are illustrated in figure [ fig : sep ] .
for each star and its nearest companion , the ratios range from 0.001 to 164107 .
49 of the 59 stars have ratios greater than 5 to their nearest companions , and thus ehzs in which planets should be in dynamically stable orbits .
the remaining 10 stars ( gj 53a , gj 222a , gj 244a , gj 280a , gj 423a , gj 423b , gj 713a , gj 713b , gj 866a , and gj 866c ) are considered to have ehzs in which planets would not be in dynamically stable orbits , so are excluded in our estimate of total ehz real estate .
seven of these have ratios less than 1.0 , hinting at the possibility of circumbinary planets . however , the majority of the ehz outer radii are only a few times the binary separations , making it unlikely that these systems would have dynamically stable circumbinary planets in the ehz .
nonetheless , it is worth noting that a few massive planets have been reported in circumbinary orbits ( lee et al .
2009 ; qian et al 2010 ) .
the described hz calculations are used to assess the total habitable real estate in the solar neighborhood and to determine the amount of habitable real estate as a function of spectral type
. evolved stars , brown dwarfs , and multiple stars with separations detrimental to the orbital stability of a planet in the ehz , as described above , are excluded in these assessments . in the 5 pc sample ,
stars removed from the analysis due to close companions include gj 244a , gj 244b , gj 280a , gj 280b , gj 866a , and gj 866c .
the distribution by spectral type , after the removal of these stars in the 5 pc sample is as follows : 0 a , 0 f , 3 g , 7 k , and 48 m stars .
similarly , stars removed from the extended 10 pc sample analysis due to close companions include gj 53a , gj 53b , gj 222a , gj 222b , gj 423a , gj
423 b , gj 423c , gj 423d , gj 713a , and gj 713b . stars with evolved spectral types such as gj 150 , gj 695a , and gj 780 are also removed from subsequent calculations . by spectral type ,
the total stellar samples considered in the final ehz assessment are 3 a , 4 f , 14 g , 34 k , and an estimated 384 m stars . using our initial assumptions of a terrestrial `` earth - like '' planet as the basis for our ehz
, we estimate the habitable real estate using linear au separations from the central star , essentially the width of the ehz . in table
[ tab : hzsp ] we present the ehz width totals for each spectral type in the 5 pc and total 10 pc samples .
the cumulative ehz width for stars in the 5 pc subsample is 8.8 au , including 2.6 au for the 3 g stars ( including the sun ) and 2.9 au for the 7 k stars .
the 48 m dwarfs in the 5 pc sample _ en masse _ provide 3.3 linear au available for habitable planets , or 38% of the available ehz .
the dominant contribution of m dwarfs to the ehz width is demonstrated clearly using the estimated 10 pc sample .
the total ehz width for the estimated 10 pc sample is 71.5 au , including 13.2 au for the 3 a , 4.9 au for the 4 f stars , 11.9 au for the 14 g stars , and 15.4 au for the 34 k stars .
the estimated 384 m stars _ en masse _ provide 26.1 au of linear ehz .
this accounts for 36.5% of the total ehz .
thus , by spectral type , m stars _ en masse _ provide the largest ehz real estate . given the rapid pace of exoplanet discovery , it would be helpful to have a tool to easily and accurately predict the location of the ehz to determine whether or not a planet resides within it .
predicting the ehz of a star based on spectral classification can be problematic due to the inhomogeneity in classification and spectral types being determined over different wavelength ranges . as shown in henry et al .
( 2006 ) , the @xmath0 color is a useful temperature diagnostic for the a through m stars that dominate the solar neighborhood .
this relation can also be very helpful in determining a rough estimate of the ehz based on observable photometry , and may be easily scaled to larger populations .
we use the results from the 5 pc and extended 10 pc samples to derive a relation between @xmath0 color and the size and location of the ehz . as in the computation of total habitable real estate
, we removed binary stars with unresolved photometry , as well as stars known to be evolved .
we do , however , use stars such as sirius ( gj 244 ) and procyon ( gj 280 ) for which ehzs have been determined , even though their companions corrupt their ehzs .
their white dwarf companions do not significantly contribute to their luminosities or @xmath0 , and their inclusion improves the statistics of our fit .
we fit a second order polynomial , described by equation [ eqvmk ] , to the @xmath0 colors and computed ehz widths shown in figure [ fig:10pchz ] .
@xmath51 using the same method , a relationship for @xmath0 color and the inner and outer radii of the ehzs are described by equations [ eqvmkin ] and [ eqvmkout ] , respectively .
these relations are only valid for main sequence stars .
@xmath52 @xmath53 as a check on these relations , we use these to calculate the total ehz width by spectral type of the estimated 10 pc sample and compare these values to the direct calculations described in sections [ ehz ] & [ bhz ] .
we find the total ehz widths from the empirical relations differ on average from the calculated total widths by @xmath3112.7% , @xmath316.4% , 1.6% , 4.7% , and 0.3% for a , f , g , k , and m stars respectively ; negative values correspond to underpredictions of the ehz width totals .
the higher percentage differences for a type and f type totals are due to the relatively small populations within 10 pc and the effects of large 2mass photometric errors due to brightness .
the results imply that this relation is useful for quickly estimating the amount of habitable real estate for a population with known @xmath0 values .
we also test how well the predicted inner and outer boundaries from our relations agree for any given star . on average , these predictions yield values consistent to 3% with a dispersion of 22% for both inner and outer boundaries for afgkm stars in our samples .
these dispersions can be interpreted as the uncertainty in the locations of these boundaries from these relations . of the 14 confirmed planetary systems within 10 pc of the sun for which orbits have been determined ,
five contain multiple planets ( gj 139 , gj 506 , gj 581 , gj 667c , and gj 876 ) .
all 14 systems are listed in table [ tab : exop ] with published values for semimajor axis and eccentricity , as well as calculated inner and outer radii of the ehz from this work .
three of the systems , gj 581 , gj 667c , and gj 876 have planets in the ehz .
gj 581 , an m2.5v star at a distance of 6.25 pc , has six proposed planets , but the existence of gj 581 g and gj 581f are currently debated ( see : vogt et al . 2010 ; andrae et al . 2010 ; gregory 2011 ; anglada - escud 2010 ; tuomi 2011 ) . if real , gj 581 g , with a semimajor axis of 0.146 au , orbits in the ehz in a presumed circular orbit . using the chara array , von braun et al .
( 2011 ) recently measured the size of the star and derived hz boundaries of r@xmath54= 0.11 au and r@xmath55= 0.21 au .
our ehz is somewhat closer in to the star , r@xmath54= 0.083 au and r@xmath55= 0.179 au , but still places gj 581 g in the ehz .
the differences are due to the our calculated luminosity ( l= 0.11@xmath56 ) being 8% lower than von braun et al .
( 2011 ) , as they adopted an extinction of a@xmath57= 0.174 for gj 581 , which they note as unexpected for a star at this distance .
gj 581d , with semimajor axis of 0.22 au and eccentricity of 0.38 , also moves in and out of the ehz of gj 581 .
gj 667c , an m1.5v star at a distance of 7.23 pc , hosts two planets , and possibly four ( anglada - escud et al . 2012 ) .
although gj 667cb does not lie within the ehz , gj 667cc
( @xmath1 sin @xmath2 of 4.5 @xmath3 ) lies within the ehz for the majority of its orbit . with a semimajor axis of 0.123 au and eccentricity @xmath180.27 , it may lie completely within the ehz once the eccentricity is more highly constrained .
gj 876 , an m3.5v star at a distance of 4.66 pc , hosts four planets ( rivera et al .
our calculations show an ehz spanning 0.090 - 0.191 au .
rivera et al .
( 2010 ) report orbital fits for two planets near the ehz with semimajor axes of 0.13 au ( gj 876c ) and 0.21 au ( gj 876b ) , and eccentricities of 0.25 and 0.03 , respectively .
as shown in figure [ fig : gj876 ] , this allows for gj 876c to be in the ehz of its host star for the full duration of its orbit , while gj 876b lies just outside the ehz .
although these planets are not considered terrestrial ( @xmath1 sin @xmath2 of 0.56 and 1.89 @xmath58 ) , the possibility exists that they could have terrestrial - like moons that could be habitable .
the previous sections provide estimates of ehzs based on the requirement of liquid water on a planetary surface . of course
, a planet s location in the ehz of a host star does not guarantee its habitability .
a host of other factors , such as planet size , atmosphere , magnetic fields , and even plate tectonics , play vital roles in determining the habitability of a planet in the ehz . without a sufficiently thick atmosphere ,
biologically harmful radiation can penetrate to the surface of a planet . on earth ,
atmospheric @xmath5 levels are kept in check by the carbon - silicate cycle .
known to regulate climate temperatures through a negative feedback , this cycle allows for as much as 60 bars ( kasting 1996 ) of @xmath5 to be locked away in rock and sediments .
this slow carbon - silicate cycle requires that water be present , because without water , the atmospheric @xmath5 can not be sequestered as carbonate .
a magnetosphere on earth also plays an important role by deflecting harmful charged particles .
tectonic activity may be one of the key factors in keeping a planet habitable ( doyle et al .
1998 ) . without water ,
the lithosphere of a planet may become a stagnant lid , halting tectonic activity and the sequestration of @xmath5 .
there is a temporal constraint on habitability as well , as the hz may change considerably over the lifetime of a star ( kasting et al .
1993 ; kasting 1996 ; tarter et al . 2007 ;
selsis et al .
putting these complications aside , the requirement of liquid water on the surface of a planet is a good first order approximation to habitability .
we assess the sample of stars currently known to be within 5 pc of the sun for the purpose of determining the habitable real estate and its dependence on spectral type . because of the sparse population of high mass stars within 5 pc of the sun
, we expand this sample for afgk stars to 10 pc ; there are no o or b stars within 10 pc of the sun . after eliminating evolved stars , substellar objects , and close multiples in which planets in hzs
would be dynamically unstable , we use the final sample to estimate the ehzs for stars in both the 5 pc and extended 10 pc samples .
we do not consider circumbinary habitable zones in this work , but ehzs are calculated in the same fashion as single stars for each of the 49 components in multiple systems that satisfy our dynamical stability constraint
. using phoenix models convolved with filter response curves , we fit observed @xmath28 photometry for each object , assuming spherical , non - rapidly rotating stars with solar metallicity and log @xmath30 values of 4.0 to 5.0 , with the 2 exceptions being the metal poor stars gj 191 and gj 451 .
this fitting process allows us to determine a radius and @xmath13 for each star that is then used to determine its surrounding ehz , calculated using a modified `` venus and early mars criterion '' from selsis et al .
( 2007 ) .
we use estimates of linear au to map the ehz of each star and sum by spectral type _ en masse _ :
48 m dwarf stars used in the 5 pc sample provide more habitable real estate ( 3.3 au ) than the three g dwarfs stars ( 2.6 au ) and seven k stars ( 2.9 au ) found within 5 pc of the sun . even after extending the sample of afgk stars to 10 pc , the anticipated sample of m dwarfs within 10 pc ( not all have yet been identified ) possess more ehz real estate than any other spectral type , spanning @xmath2026 au compared to 13.2 au , 11.9 au , and 15.4 au for each of the a , g , and k types ( the f stars provide only 4.9 linear au of ehz ) .
the result is a natural consequence of the large relative numbers of m dwarfs , and the frequency of close companions that declines with mass .
as a population m dwarfs provide more options and more habitable real estate than their more massive counterparts .
furthermore , recent results from kepler show that for stars with t@xmath17>4000k within 5 pc of the sun , there is likely to be at least 2 earth - size planets in the hz .
that number increases to 16 within 10 pc ( dressing & charbonneau 2013 ) . using the 5 pc and extended 10 pc samples , we derive relations between @xmath0 color and ehz width and inner and outer limits .
comparisons of color predicted locations suggest they are comparable to uncertainties associated with habitability assumptions ( e.g. , section [ ehz ] ) .
thus , these color relations are practical tools for estimating the ehzs of stars using commonly available photometric measurements .
the relations for the inner and outer radii of the ehz are helpful for quickly estimating whether or not a known planet or disk is within the ehz .
the relation for ehz size is useful in predicting the habitable real estate available in a stellar population . in particular
, we consider the results for the 14 extrasolar planetary systems known within 10 pc of the sun .
the three systems with planets in the calculated ehzs gj 581 , gj 667c , and gj 876 are all m dwarfs . in total , as many as four planets circling these stars spend at least part of their time in the ehzs , providing an ideal set of targets for future efforts to detect biosignatures .
the authors would like to thank j. kasting for extremely helpful advice and guidance .
we would also like to thank the recons team for the generous amount of data and support provided .
data used in this paper were acquired via support from the national science foundation ( grants ast 05 - 07711 and ast 09 - 08402 ) and through the continuous cooperation of the smarts consortium .
this project was funded in part by nsf / aag grant no . 0908018 .
this research has made use of the simbad database , operated at cds , strasbourg , france , data from the two micron all sky survey , which is a joint project of the university of massachusetts and ipac , funded by nasa and nsf , and the sixth catalog of orbits of visual binary stars , operated by the us naval observatory .
absil , o. , di folco , e. , mrand , a. , et al .
2008 , , 487 , 1041 alonso , a. , arribas , s. , & martinez - roger , c. 1995 , , 297 , 197 andrae , r. , schulze - hartung , t. , & melchior , p. 2010
, arxiv:1012.3754 anglada - escud , g. , arriagada , p. , vogt , s. s. , et al . 2012 , , 751 , l16 anglada - escud , g. 2010 , arxiv:1011.0186 aufdenberg , j. p. , et al . 2006 , , 651 , 617 bailey , j. , butler , r. p. , tinney , c. g. , jones , h. r. a. , otoole , s. , carter , b. d. , & marcy , g. w. 2009 , , 690 , 743 baize , p. , & petit , m. 1989 , , 77 , 497 baize , p. 1950 , journal des observateurs , 33 , 1 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1998 , , 337 , 403 barbieri , m. , marzari , f. , & scholl , h. 2002 , , 396 , 219 barnes , r. , jackson , b. , greenberg , r. , & raymond , s. n. 2009 , , 700 , l30 batten , a. h. , fletcher , j. m. , & maccarthy , d. g. 1989 , publications of the dominion astrophysical observatory victoria , 17 , 1 benedict , g. f. , et al . 1999 , , 118 , 1086 benedict , g. f. , et al .
2002 , , 581 , l115 benedict , g. f. , et al .
2006 , , 132 , 2206 berger , d. h. , et al .
2006 , , 644 , 475 bessell , m. s. , castelli , f. , & plez , b. 1998 , , 333 , 231 bessel , m. s. 1990 , , 83 , 357 bessell , m. s. 1991 , , 101 , 662 biller , b. a. , kasper , m. , close , l. m. , brandner , w. , & kellner , s. 2006 , , 641 , l141 bonfils , x. , et al .
2007 , , 474 , 293 borucki , w. j. , & for the kepler team 2010 , arxiv:1006.2799 boyajian , t. s. , von braun , k. , van belle , g. , et al .
2012 , arxiv:1208.2431 breckinridge , j. b. , & kron , g. e. 1964 , , 76 , 139 burgasser , a. j. , et al .
2000 , , 531 , l57 butler , r. p. , et al . 2006 , ,
646 , 505 butler , r. p. , johnson , j. a. , marcy , g. w. , wright , j. t. , vogt , s. s. , & fischer , d. a. 2006 , , 118 , 1685 carpenter , j. m. 2001 , , 121 , 2851 chance , d. r. , & hershey , j. l. 1998 , , 110 , 425 cohen , m. , wheaton , w. a. , & megeath , s. t. 2003 , , 126 , 1090 costa , e. , mndez , r. a. , jao , w .- c . ,
henry , t. j. , subasavage , j. p. , brown , m. a. , ianna , p. a. , & bartlett , j. 2005 , , 130 , 337 costa , e. , mndez , r. a. , jao , w .- c . , henry , t. j. , subasavage , j. p. , & ianna , p. a. 2006 , , 132 , 1234 couteau , p. 1959
, bulletin astronomique , 23 , 127 cutri , r. m. , et al .
2003 , the irsa 2mass all - sky point source catalog , nasa / ipac infrared science archive.http://irsa.ipac.caltech.edu/applications/gator/ deacon , n. r. , & hambly , n. c. 2007 , , 468 , 163 deacon , n. r. , hambly , n. c. , henry , t. j. , subasavage , j. p. , brown , m. a. , & jao , w .- c .
2005 , , 129 , 409 delfosse , x. , forveille , t. , udry , s. , beuzit , j .- l . , mayor , m. , & perrier , c. 1999 , , 350 , l39 demory , b .- o . , et al .
2009 , , 505 , 205 di folco , e. , thvenin , f. , kervella , p. , domiciano de souza , a. , coud du foresto , v. , sgransan , d. , & morel , p. 2004
, , 426 , 601 doyle , l. r .. billingham , j. , devinvenzi , d. l. , 1998 , acta astronautica , 42 , 399 dressing , c. d. , & charbonneau , d. 2013 , , 767 , 95 drummond , j. d. , christou , j. c. , & fugate , r. q. 1995 , , 450 , 380 dumusque , x. , pepe , f. , lovis , c. , et al .
2012 , , 491 , 207 eggen , o. j. 1956 , , 61 , 405 eggen , o. j. 1965 , , 70 , 19 eggen , o. j. 1996 , , 111 , 466 eggenberger , a. , udry , s. , chauvin , g. , beuzit , j .- l . , lagrange , a .- m . , sgransan , d. , & mayor , m. 2007 , , 474 , 273 farrington , c. d. , et al .
2010 , , 139 , 2308 fekel , f. c. 1997 , , 109 , 514 forveille , t. , et al .
1999 , , 351 , 619 forveille , t. , et al .
2009 , , 493 , 645 fuhrmann , k. , pfeiffer , m. , frank , c. , reetz , j. , & gehren , t. 1997 , , 323 , 909 gatewood , g. d. , & gatewood , c. v. 1978 , , 225 , 191 gatewood , g. , stein , j. , de jonge , j. k. , persinger , t. , reiland , t. , & stephenson , b. 1992 , , 104 , 1237 gatewood , g. , de jonge , k. j. , & stephenson , b. 1993 , , 105 , 1101 gatewood , g. , coban , l. , & han , i. 2003 , , 125 , 1530 gatewood , g. 1989 , , 97 , 1189 gatewood , g. 1994 , , 106 , 138 geyer , d. w. , harrington , r. s. , & worley , c. e. 1988 , , 95 , 1841 glass , i. s. 1975 , , 171 , 19p gliese , w. , & jahrei , h. 1991 , on : the astronomical data center cd - rom : selected astronomical catalogs , vol .
i ; l.e .
brotzmann , s.e .
gesser ( eds . ) , nasa / astronomical data center , goddard space flight center , greenbelt , md , golimowski , d. a. , henry , t. j. , krist , j. e. , schroeder , d. j. , marcy , g. w. , fischer , d. a. , & butler , r. p. 2000 , , 120 , 2082 gould , a. 2003 , , 126 , 472 gray , r. o. , napier , m. g. , & winkler , l. i. 2001 , , 121 , 2148 gray , r. o. , corbally , c. j. , garrison , r. f. , mcfadden , m. t. , bubar , e. j. , mcgahee , c. e. , odonoghue , a. a. , & knox , e. r. 2006 , , 132 , 161 gregory , p. c. 2011 , arxiv:1101.0800 greissl , j. , meyer , m. r. , wilking , b. a. , fanetti , t. , schneider , g. , greene , t. p. , & young , e. 2007 , , 133 , 1321 hale , a. 1994 , , 107 , 306 han , i. , & gatewood , g. 2002 , , 114 , 224 hauschildt , p. h. , allard , f. , & baron , e. 1999 , , 512 , 377 heintz , w. d. 1974 , , 79 , 819 heintz , w. d. 1986 , , 92 , 446 heintz , w. d. 1987 , , 99 , 1084 heintz , w. d. 1994 , , 108 , 2338 heintz , w. d. 1996 , , 111 , 412 henry , t.j .
2010 , in royal astronomical society of canada observer s handbook 2010 , ed .
patrick kelly .
toronto : royal astronomical society of canada henry , t.j .
2012 , in royal astronomical society of canada observer s handbook 2012 , ed .
davidm.f .
chapman .
toronto : royal astronomical society of canada henry , t. j. , & mccarthy , d. w. , jr .
1993 , , 106 , 773 henry , t. j. , kirkpatrick , j. d. , & simons , d. a. 1994 , , 108 , 1437 henry , t. j. , ianna , p. a. , kirkpatrick , j. d. , & jahreiss , h. 1997 , , 114 , 388 henry , t. j. , walkowicz , l. m. , barto , t. c. , & golimowski , d. a. 2002 , , 123,2002 henry , t. j. , subasavage , j. p. , brown , m. a. , beaulieu , t. d. , jao , w .- c . , & hambly , n. c. 2004 , , 128 , 2460 henry , t. j. , jao , w .- c .
, subasavage , j. p. , beaulieu , t. d. , ianna , p. a. , costa , e. , & mndez , r. a. 2006 , , 132 , 2360 howard , a. w. , et al .
2010 , arxiv:1011.0414 howard , a. w. , marcy , g. w. , bryson , s. t. , et al .
2012 , , 201 , 15 huang , s .- s .
1959 , , 71 , 421 hg , e. , et al .
2000 , , 355 , l27 irwin , a. w. , fletcher , j. m. , yang , s. l. s. , walker , g. a. h. , & goodenough , c. 1992 , , 104 , 489 irwin , a. w. , yang , s. l. s. , & walker , g. a. h. 1996 , , 108 , 580 jao , w .- c . , henry , t. j. , subasavage , j. p. , brown , m. a. , ianna , p. a. , bartlett , j. l. , costa , e. , & mndez , r. a. 2005 , , 129 , 1954 jao , w .- c . , henry , t. j. , beaulieu , t. d. , & subasavage , j. p. 2008
, , 136 , 840 johnson , h. l. , & morgan , w. w. 1953 , , 117 , 313 johnson , h. l. , & morgan , w. w. 1953 , , 117 , 313 johnson , h. l. , iriarte , b. , mitchell , r. i. , & wisniewski , w. z. 1966 , communications of the lunar and planetary laboratory , 4 , 99 johnson , h. l. , macarthur , j. w. , & mitchell , r. i. 1968 , , 152 , 465 johnson , j. a. , butler , r. p. , marcy , g. w. , et al .
2007 , , 670 , 833 johnson , j. a. , et al .
2010 , , 122 , 149 kalas , p. , graham , j. r. , chiang , e. , et al .
2008 , science , 322 , 1345 kaltenegger , l. , fridlund , m. , & kasting , j. 2002 , esa sp-514 : earth - like planets and moons , 277 kaltenegger , l. 2010 , , 712 , l125 kasting , j. f. 1996 , , 241 , 3 kasting , j. f. , whitmire , d. p. , & reynolds , r. t. 1993 , icarus , 101 , 108 kenyon , s. j. , & hartmann , l. 1995 , , 101 , 117 kervella , p. , & fouqu , p. 2008
, , 491 , 855 kervella , p. , thvenin , f. , morel , p. , bord , p. , & di folco , e. 2003 , , 408 , 681 kervella , p. , thvenin , f. , morel , p. , berthomieu , g. , provost , j. , bord , p. , & sgransan , d. 2003 , solar and solar - like oscillations : insights and challenges for the sun and stars , 25th meeting of the iau , joint discussion 12 , 18 july 2003 , sydney , australia , 12 kervella , p. , et al .
2008 , , 488 , 667 kidder , k. m. , holberg , j. b. , & mason , p. a. 1991 , , 101 , 579 kirkpatrick , j. d. , dahn , c. c. , monet , d. g. , reid , i. n. , gizis , j. e. , liebert , j. , & burgasser , a. j. 2001 , , 121 , 3235 koch , d. g. , et al .
2004 , , 5487 , 1491 landolt , a. u. 1992 , , 104 , 372 lane , b. f. , boden , a. f. , & kulkarni , s. r. 2001 , , 551 , l81 lee , j. w. , kim , s .- l . , kim , c .- h . , koch , r. h. , lee , c .- u . ,
, & park , j .- h . 2009 , , 137 , 3181 leggett , s. k. 1992 , , 82 , 351 leinert , c. , haas , m. , allard , f. , et al . 1990 , , 236 , 399 lowell , p. 1895
, popular astronomy , 2 , 255 lutz , t. e. 1971 , , 83 , 488 luyten , w. j. 1957 , minneapolis , lund press , 1957 . , lpine , s. , & shara , m. m. 2005 , , 129 , 1483 malin , m. c. , & edgett , k. s. 2000 , science , 288 , 2330 marcy , g. w. , butler , r. p. , fischer , d. , vogt , s. s. , lissauer , j. j. , & rivera , e. j. 2001 , , 556 , 296 marion , g. m. , fritsen , c. h. , eicken , h. , & payne , m. c. 2003 , astrobiology , 3 , 785 mason , b. d. , & hartkopf , w. i. 2006 , journal of double star observations , 2 , 171 mason , b. d. , mcalister , h. a. , hartkopf , w. i. , & shara , m. m. 1995 , , 109 , 332 mayor , m. , et al .
2009 , , 507 , 487 mccaughrean , m. j. , close , l. m. , scholz , r .- d . ,
lenzen , r. , biller , b. , brandner , w. , hartung , m. , & lodieu , n. 2004 , , 413 , 1029 mermilliod , j .- c .
1986 , catalogue of eggen s ubv data . , 0 ( 1986 ) ,
0 monet , d. g. , et al .
2003 , , 125 , 984 monnier , j. d. , zhao , m. , pedretti , e. , et al .
2007 , science , 317 , 342 montes , d. , ramse y , l .
w. , & welty , a. d. 1999 , , 123 , 283 montes , d. , lpez - santiago , j. , glvez , m. c. , fernndez - figueroa , m. j. , de castro , e. , & cornide , m. 2001 , , 328 , 45 mould , j. r. , & hyland , a. r. 1976 , , 208 , 399 oja , t. 1984 , , 57 , 357 oja , t. 1993 , , 100 , 591 parkinson , c. d. , liang , m .- c . , hartman , h. , hansen , c. j. , tinetti , g. , meadows , v. , kirschvink , j. l. , & yung , y. l. 2007 , , 463 , 353 pasquini , l. , liu , q. , & pallavicini , r. 1994 , , 287 , 191 patience , j. , et al .
2002 , , 581 , 654 pepe , f. , lovis , c. , sgransan , d. , et al .
2011 , , 534 , a58 perryman , m. a. c. , lindegren , l. , kovalevsky , j. , et al .
1997 , , 323 , l49 pourbaix , d. , et al .
2002 , , 386 , 280 pourbaix , d. 2000 , , 145 , 215 poveda , a. , herrera , m. a. , allen , c. , cordero , g. , & lavalley , c. 1994 , , 28 , 43 qian , s .- b . ,
liao , w .- p . , zhu , l .- y . ,
& dai , z .- b .
2010 , , 708 , l66 quintana , e. v. , adams , f. c. , lissauer , j. j. , & chambers , j. e. 2007 , , 660 , 807 raghavan , d. , henry , t. j. , mason , b. d. , subasavage , j. p. , jao , w .- c . , beaulieu , t. d. , & hambly , n. c. 2006 , , 646 , 523 raghavan , d. , et al .
2010 , , 190 , 1 reid , i. n. , & gizis , j. e. 1997 , , 113 , 2246 reid , i. n. , hawley , s. l. , & gizis , j. e. 1995 , , 110 , 1838 reid , i. n. , et al .
2004 , , 128 , 463 rivera , e. j. , laughlin , g. , butler , r. p. , vogt , s. s. , haghighipour , n. , & meschiari , s. 2010 , , 719 , 890 royer , f. , gerbaldi , m. , faraggiana , r. , & gmez , a. e. 2002 , , 381 , 105 royer , f. , zorec , j. , & gmez , a. e. 2007 , , 463 , 671 ruck , m. j. , & smith , g. 1995 , , 304 , 449 schmidt , s. j. , cruz , k. l. , bongiorno , b. j. , liebert , j. , & reid , i. n. 2007 , , 133 , 2258 scholz , r .- d . , mccaughrean , m. j. , lodieu , n. , & kuhlbrodt , b. 2003 , , 398 , l29 selsis , f. , kasting , j. f. , levrard , b. , paillet , j. , ribas , i. , & delfosse , x. 2007 , , 476 , 1373 strand , k. a. 1969 , , 74 , 760 subasavage , j .
p. , jao , w .- c .
, henry , t. j. , bergeron , p. , dufour , p. , ianna , p. a. , costa , e. , & mndez , r. a. 2009 , , 137 , 4547 sderhjelm , s. 1999 , , 341 , 121 sderhjelm , s. 1999 , , 341 , 121 sgransan , d. , delfosse , x. , forveille , t. , beuzit , j .- l . ,
udry , s. , perrier , c. , & mayor , m. 2000 , , 364 , 665 sgransan , d. , kervella , p. , forveille , t. , & queloz , d. 2003 , , 397 , l5 tarter , j. c. , et al .
2007 , astrobiology , 7 , 30 tinney , c. g. , butler , r. p. , jones , h. r. a. , wittenmyer , r. a. , otoole , s. , bailey , j. , & carter , b. d. 2011 , , 727 , 103 tokovinin , a. a. 1997 , , 124 , 75 torres , g. , henry , t. j. , franz , o. g. , & wasserman , l. h. 1999 , , 117 , 562 torres , c. a. o. , quast , g. r. , da silva , l. , de la reza , r. , melo , c. h. f. , & sterzik , m. 2006 , , 460 , 695 tuomi , m. 2011 , , 528 , l5 turner , n. h. , ten brummelaar , t. a. , mcalister , h. a. , mason , b. d. , hartkopf , w. i. , & roberts , l. c. , jr .
2001 , , 121 , 3254 vogt , s. s. , et al .
2010 , , 708 , 1366 vogt , s. s. , butler , r. p. , rivera , e. j. , haghighipour , n. , henry , g. w. , & williamson , m. h. 2010 , , 723 , 954 weidner , c. , & horne , k. 2010 , , 521 , a76 weis , e. w. 1993 , , 105 , 1962 weis , e. w. 1996 , , 112 , 2300 wiegert , p. a. , & holman , m. j. 1997 , , 113 , 1445 wieth - knudsen , n. inf . circ . 13 , 1957 wilson , o. c. 1966 , , 144 , 695 woolf , v. m. , & wallerstein , g. 2005 , , 356 , 963 worley , c. e. , & behall , a. l. 1973 , , 78 , 650 van albada , g. b. 1957 , contributions from the bosscha observatory , 5 , 1 van altena , w. f. , lee , j. t. , & hoffleit , e. d. 1995 , the general catalogue of trigonometric stellar parallaxes , new haven , ct : yale university observatory , 1995 , 4th ed .
valenti , j. a. , & fischer , d. a. 2005 , vizier online data catalog , 215 , 90141 von braun , k. , boyajian , t. s. , kane , s. r. , et al .
2011 , , 729 , l26 values .
the solid line illustrates 1:1 agreement and the dashed lines represent 10% offsets .
error bars are 1@xmath37 for the model and given errors for the interferometrically derived values used to derive the published values . ] -@xmath59 value .
photometrically unresolved multiples are plotted as _ open squares _ , while resolved components are plotted as _ filled circles_. stars with companions that get closer than 5 times the outer ehz boundary ( _ dashed line _ ) are considered dynamically unstable planet hosts , and are not included in the total habitable real estate calculations .
gj 663ab and gj 663ba are plotted as the same point as their @xmath41v and @xmath41k = 0 .
the open square at @xmath0 = 8.88 , ratio = 173 , is scr 1845 - 6357 , an m dwarf with a brown dwarf companion in a highly uncertain orbit . ]
lrccccccc sun & & & & & & & g2.0v & + gj 1 & 1 & 00 05 24.4 & @xmath3137 21 27 & .23032 & .00090 & 2 & m1.5v & hen10 + gj 1002 & 2 & 00 06 43.2 & @xmath3107 32 17 & .21300 & .00360 & 1 & m5.0v & hen94 + gj 15 a & 3 & 00 18 22.9 & @xmath6044 01 23 & .27987 & .00060 & 3 & m1.5v & hen94 + gj 15 b & 4 & 00 18 22.9 & @xmath6044 01 23 & .27987 & .00060 & 3 & m3.5v & hen94 + gj 35 & 7 & 00 49 09.9 & @xmath6005 23 19 & .23270 & .00181 & 2 & dz7 & ypc95 + gj 54.1 & 138 & 01 12 30.6 & @xmath3116 59 57 & .26908 & .00299 & 2 & m4.5v & hen94 + gj 65 a & 9 & 01 39 01.3 & @xmath3117 57 01 & .37370 & .00270 & 1 & m5.5v & hen94 + gj 65 b & 10 & 01 39 01.3 & @xmath3117 57 01 & .37370 & .00270 & 1 & m6.0v & hen94 + gj 71 & 146 & 01 44 04.1 & @xmath3115 56 15 & .27397 & .00017 & 2 & g8.5v & gra06 + gj 83.1 & 11 & 02 00 13.2 & @xmath6013 03 08 & .22480 & .00290 & 1 & m4.5v & hen94 + so 0253@xmath601652 & & 02 53 00.9 & @xmath6016 52 53 & .25941 & .00089 & 3 & m7.0v & hen04 + den 0255 - 4700 & & 02 55 03.7 & @xmath3147 00 52 & .20137 & .00389 & 1 & l7.5v & cos06 + gj 144 & 1557 & 03 32 55.8 & @xmath3109 27 30 & .31122 & .00009 & 3 & k2.0v & gra06 + gj 1061 & 1565 & 03 36 00.0 & @xmath3144 30 46 & .27201 & .00130 & 2 & m5.0v & hen06 + lp 944 - 020 & & 03 39 35.2 & @xmath3135 25 41 & .20140 & .00421 & 1 & m9.0v & sch07 + gj 166 a & 23 & 04 15 16.3 & @xmath3107 39 10 & .20065 & .00023 & 2 & k0.5v & gra06 + gj 166 b & 24 & 04 15 22.0 & @xmath3107 39 35 & .20065 & .00023 & 2 & da4 & cns91 + gj 166 c & 25 & 04 15 22.0 & @xmath3107 39 35 & .20065 & .00023 & 2 & m4.5v & hen94 + gj 191 & 29 & 05 11 40.6 & @xmath3145 01 06 & .25567 & .00091 & 2 & m2.0vi & jao08 + gj 234 a & 1849 & 06 29 23.4 & @xmath3102 48 50 & .24444 & .00092 & 3 & m4.0v & mon01 + gj 234 b & 1850 & 06 29 23.4 & @xmath3102 48 50 & .24444 & .00092 & 3 & m5.5v & rei04 + gj 244 a & 219 & 06 45 08.9 & @xmath3116 42 58 & .38002 & .00128 & 2 & a1.0v & joh53 + gj 244 b & & 06 45 08.9 & @xmath3116 42 58 & .38002 & .00128 & 2 & da2 & cns91 + gj 273 & 33 & 07 27 24.5 & @xmath6005 13 33 & .26623 & .00066 & 3 & m3.5v & hen94 + gj 280 a & 233 & 07 39 18.1 & @xmath6005 13 30 & .28517 & .00064 & 4 & f5.0iv@xmath31v & gra01 + gj 280 b & & 07 39 18.1 & @xmath6005 13 30 & .28517 & .00064 & 4 & da & cns91 + gj 1111 & 248 & 08 29 49.5 & @xmath6026 46 37 & .27580 & .00300 & 1 & m6.0v & rei95 + gj 380 & 280 & 10 11 22.1 & @xmath6049 27 15 & .20553 & .00049 & 2 & k7.0v & hen94 + gj 388 & 5167 & 10 19 36.4 & @xmath6019 52 10 & .20460 & .00280 & 1 & m3.0v & hen94 + lhs 288 & 288 & 10 44 21.2 & @xmath3161 12 36 & .20970 & .00265 & 2 & m5.5v & hen04 + lhs 292 & 292 & 10 48 12.6 & @xmath3111 20 14 & .22030 & .00360 & 1 & m6.5v & rei95 + den 1048 - 3956 & & 10 48 14.6 & @xmath3139 56 07 & .24853 & .00118 & 3 & m8.5v & hen04 + gj 406 & 36 & 10 56 29.2 & @xmath6007 00 53 & .41910 & .00210 & 1 & m6.0v & hen94 + gj 411 & 37 & 11 03 20.2 & @xmath6035 58 12 & .39325 & .00057 & 2 & m2.0v & hen94 + gj 412 a & 38 & 11 05 28.6 & @xmath6043 31 36 & .20567 & .00093 & 2 & m1.0v & hen94 + gj 412 b & 39 & 11 05 30.4 & @xmath6043 31 18 & .20567 & .00093 & 2 & m5.5v & hen94 + gj 440 & 43 & 11 45 42.9 & @xmath3164 50 29 & .21612 & .00109 & 3 & dq6 & cns91 + gj 447 & 315 & 11 47 44.4 & @xmath6000 48 16 & .29814 & .00137 & 2 & m4.0v & hen94 + gj 473 a & 333 & 12 33 17.2 & @xmath6009 01 15 & .22790 & .00460 & 1 & m5.0v & hen10 + gj 473 b & & 12 33 17.2 & @xmath6009 01 15 & .22790 & .00460 & 1 & m7.0v & cns91 + gj 551 & 49 & 14 29 43.0 & @xmath3162 40 46 & .76885 & .00029 & 4 & m5.0v & cns91 + gj 559 a & 50 & 14 39 36.5 & @xmath3160 50 02 & .74723 & .00117 & 1 & g2.0v & gra06 + gj 559 b & 51 & 14 39 35.1 & @xmath3160 50 14 & .74723 & .00117 & 1 & k0.0v & cns91 + gj 628 & 419 & 16 30 18.1 & @xmath3112 39 45 & .23438 & .00150 & 2 & m3.0v & hen94 + gj 674 & 449 & 17 28 39.9 & @xmath3146 53 43 & .22011 & .00139 & 2 & m2.5v & mon01 + gj 687 & 450 & 17 36 25.9 & @xmath6068 20 21 & .22047 & .00083 & 2 & m3.0v & hen94 + gj 699 & 57 & 17 57 48.5 & @xmath6004 41 36 & .54551 & .00029 & 2 & m4.0v & hen94 + gj 725 a & 58 & 18 42 46.7 & @xmath6059 37 49 & .28383 & .00146 & 3 & m3.0v & hen94 + gj 725 b & 59 & 18 42 46.9 & @xmath6059 37 37 & .28383 & .00146 & 3 & m3.5v & hen94 + scr 1845 - 6357 a & & 18 45 02.6 & @xmath3163 57 52 & .25950 & .00111 & 2 & m8.5v & hen04 + scr 1845 - 6357 b & & 18 45 02.6 & @xmath3163 57 52 & .25950 & .00111 & 2 & t5.5v & bil06 + gj 729 & 3414 & 18 49 49.4 & @xmath3123 50 10 & .33722 & .00197 & 2 & m3.5v & hen10 + gj 1245 a & 3494 & 19 53 54.2 & @xmath6044 24 55 & .22020 & .00100 & 1 & m5.5v & hen94 + gj
1245 b & 3495 & 19 53 55.2 & @xmath6044 24 56 & .22020 & .00100 & 1 & m6.0v & hen94 + gj 1245 c & & 19 53 54.2 & @xmath6044 24 55 & .22020 & .00100 & 1 & m7.0v & rei04 + gj 820 a & 62 & 21 06 53.9 & @xmath6038 44 58 & .28608 & .00048 & 3 & k5.0v & hen94 + gj 820 b & 63 & 21 06 55.3 & @xmath6038 44 31 & .28608 & .00048 & 3 & k7.0v & hen94 + gj 825 & 66 & 21 17 15.3 & @xmath3138 52 03 & .25344 & .00080 & 2 & m0.0v & tor06 + gj 832 & 3685 & 21 33 34.0 & @xmath3149 00 32 & .20203 & .00100 & 2 & m1.5v & joh10 + gj 845 a & 67 & 22 03 21.7 & @xmath3156 47 10 & .27607 & .00028 & 2 & k4.0v & gra06 + gj 845 b & & 22 04 10.5 & @xmath3156 46 58 & .27607 & .00028 & 2 & t1.0v & mcc04 + gj 845 c & & 22 04 10.5 & @xmath3156 46 58 & .27607 & .00028 & 2 & t6.0v & mcc04 + gj 860 a & 3814 & 22 27 59.5 & @xmath6057 41 45 & .24806 & .00139 & 2 & m3.0v & hen94 + gj 860 b & 3815 & 22 27 59.5 & @xmath6057 41 45 & .24806 & .00139 & 2 & m4.0v & hen94 + gj 866 a & 68 & 22 38 33.4 & @xmath3115 18 07 & .28950 & .00440 & 1 & m5.0v & hen02 + gj 866 b & & 22 38 33.4 & @xmath3115 18 07 & .28950 & .00440 & 1 & m & + gj 866 c & & 22 38 33.4 & @xmath3115 18 07 & .28950 & .00440 & 1 & m & + gj 876 & 530 & 22 53 16.7 & @xmath3114 15 49 & .21447 & .00057 & 3 & m4.0v & mon01 + gj 887 & 70 & 23 05 52.0 & @xmath3135 51 11 & .30508 & .00070 & 2 & m2.0v & tor06 + gj 905 & 549 & 23 41 55.0 & @xmath6044 10 38 & .31637 & .00055 & 3 & m5.5v & hen94 + [ tab : parx ] llcccccccccccccccc sun & @xmath3125.97 & alo95 & @xmath3126.10 & alo95 & @xmath3126.75 & alo95 & @xmath3127.27 & alo95 & @xmath3127.56 & alo95 & @xmath3127.928 & alo95 & @xmath3128.211 & alo95 & @xmath3128.274 & alo95 & + gj 1 & & & 10.02 & bes90 & 8.54 & bes90 & 7.57 & bes90 & 6.41 & bes90 & 5.328 & 0.019 & 4.828 & 0.076 & 4.523 & 0.017 & + gj 1002 & 17.61 & leg92 & 15.73 & wei96 & 13.77 & bes91 & 12.16 & bes91 & 10.15 & bes91 & 8.323 & 0.019 & 7.792 & 0.034 & 7.439 & 0.021 & + gj 15 a & 10.87 & leg92 & 9.63 & leg92 & 8.08 & leg92 & & & 5.94 & leg92 & 5.252 & 0.264 & 4.476 & 0.200 & 4.018 & 0.020 & + gj 15 b & 14.26 & leg92 & 12.88 & wei96 & 11.06 & wei96 & 9.83 & wei96 & 8.24 & wei96 & 6.789 & 0.024 & 6.191 & 0.016 & 5.948 & 0.024 & + gj 35 & & & 12.94 & bes90 & 12.40 & bes90 & 12.14 & bes90 & 11.91 & bes90 & 11.688 & 0.022 & 11.572 & 0.024 & 11.498 & 0.025 & + gj 54.1 & 15.24 & leg92 & 13.95 & bes90 & 12.10 & bes90 & 10.73 & bes90 & 8.95 & bes90 & 7.258 & 0.020 & 6.749 & 0.033 & 6.420 & 0.017 & + gj 65 a & 14.96 & leg92 & 13.95 & bes90 & 12.61 * & hen99 & 10.40 & bes90 & 8.34 & bes90 & 6.86 * & hen93 & 6.30 * & hen93 & 5.91 * & hen93 & joint ubri + gj 65 b & & & & & 13.06 * & hen99 & & & & & 7.24 * & hen93 & 6.60 * & hen93 & 6.31 * & hen93 & + gj 71 & 4.42 & joh66 & 4.21 & bes90 & 3.49 & bes90 & 3.06 & bes90 & 2.67 & bes90 & 2.06 & joh66 & 1.800 & 0.234 & 1.68 & joh66 & + gj 83.1 & 15.47 & leg92 & 14.14 & bes90 & 12.31 & bes90 & 10.95 & bes90 & 9.21 & bes90 & 7.514 & 0.017 & 6.970 & 0.027 & 6.648 & 0.017 & + so 0253@xmath601652 & & & & & 15.13 & hen06 & 13.03 & hen06 & 10.65 & hen06 & 8.394 & 0.027 & 7.883 & 0.040 & 7.585 & 0.046 & + den 0255 - 4700 & & & & & 22.92 & cos06 & 19.90 & cos06 & 17.45 & cos06 & 13.246 & 0.027 & 12.204 & 0.024 & 11.558 & 0.024 & + gj 144 & & & 4.61 & bes90 & 3.73 & bes90 & 3.22 & bes90 & 2.79 & bes90 & 2.20 & gla75 & 1.75 & gla75 & 1.65 & gla75 & + gj 1061 & & & & & 13.09 & hen06 & 11.45 & hen06 & 9.46 & hen06 & 7.523 & 0.020 & 7.015 & 0.044 & 6.610 & 0.021 & + lp 944 - 020 & & & & & & & & & 13.29 & dea07 & 10.725 & 0.021 & 10.017 & 0.021 & 9.548 & 0.023 & + gj 191 & & & 10.41 & bes90 & 8.85 & bes90 & 7.90 & bes90 & 6.90 & bes90 & 5.821 & 0.025 & 5.316 & 0.027 & 5.049 & 0.021 & + gj 234 a & 14.03 & leg92 & 12.81 & bes90 & 11.18 * & hen99 & 9.78 & bes90 & 8.08 & bes90 & 6.57 * & hen93 & 5.97 * & hen93 & 5.73 * & hen93 & joint ubri + gj 234 b & & & & & 14.26 * & hen99 & & & & & 8.36 * & hen93 & 7.60 * & hen93 & 7.23 * & hen93 & + gj 244 ab & @xmath311.4 1 & joh66 & @xmath311.43 & bes90 & @xmath311.43 & bes90 & @xmath311.42 & bes90 & @xmath311.41 & bes90 & @xmath311.391 & 0.109 & @xmath311.391 & 0.184 & @xmath311.390 & 0.214 & joint + gj 273 & 12.59 & wei93 & 11.42 & bes90 & 9.85 & bes90 & 8.70 & bes90 & 7.16 & bes90 & 5.714 & 0.032 & 5.219 & 0.063 & 4.857 & 0.022 & + gj 280 ab & 0.82 & joh66 & 0.79 & bes90 & 0.37 & bes90 & 0.12 & bes90 & @xmath310.12 & bes90 & @xmath310.40 & gla75 & @xmath310.60 & gla75 & @xmath310.65 & gla75 & joint + gj 1111 & & & 16.95 & bes90 & 14.90 & bes90 & 12.90 & bes90 & 10.64 & bes90 & 8.235 & 0.021 & 7.617 & 0.018 & 7.260 & 0.024 & + gj 380 & 9.25 & leg92 & 7.97 & leg92 & 6.59 & leg92 & 5.74 & leg92 & 4.97 & leg92 & 3.98 & gla75 & 3.32 & gla75 & 3.19 & gla75 & + gj 388 & 11.91 & leg92 & 10.85 & leg92 & 9.32 & leg92 & 8.23 & leg92 & 6.81 & leg92 & 5.449 & 0.027 & 4.843 & 0.020 & 4.593 & 0.017 & + lhs 288 & & & & & 13.92 & bes91 & 12.33 & bes91 & 10.31 & bes91 & 8.492 & 0.021 & 8.054 & 0.044 & 7.728 & 0.027 & + lhs 292 & & & 17.70 & leg92 & 15.73 & bes91 & 13.67 & bes91 & 11.33 & bes91 & 8.857 & 0.021 & 8.263 & 0.036 & 7.926 & 0.033 & + den 1048 - 3956 & & & & & 17.39 & jao05 & 15.05 & jao05 & 12.55 & jao05 & 9.538 & 0.022 & 8.905 & 0.044 & 8.447 & 0.023 & + gj 406 & 17.03 & leg92 & 15.52 & bes90 & 13.53 & bes90 & 11.67 & bes90 & 9.50 & bes90 & 7.085 & 0.024 & 6.482 & 0.042 & 6.084 & 0.017 & + gj 411 & 10.12 & leg92 & 8.98 & leg92 & 7.47 & leg92 & 6.46 & leg92 & 5.32 & leg92 & 4.13 & gla75 & 3.56 & gla75 & 3.20 & gla75 & + gj 412 a & 11.48 & leg92 & 10.34 & wei96 & 8.77 & wei96 & 7.79 & wei96 & 6.70 & wei96 & 5.538 & 0.019 & 5.002 & 0.021 & 4.769 & 0.020 & + gj 412 b & & & 16.53 & bes90 & 14.44 & bes90 & 12.77 &
bes90 & 10.68 & bes90 & 8.742 & 0.025 & 8.177 & 0.024 & 7.839 & 0.026 & + gj 440 & 11.04 & lan92 & 11.68 & lan92 & 11.50 & sub09 & 11.34 & sub09 & 11.20 & sub09 & 11.188 & 0.024 & 11.130 & 0.025 & 11.104 & 0.026 & + gj 447 & 14.22 & leg92 & 12.92 & bes90 & 11.16 & bes90 & 9.85 & bes90 & 8.17 & bes90 & 6.505 & 0.023 & 5.945 & 0.024 & 5.654 & 0.024 & + gj 473 a & & & 15.06 * & tor99 & 13.25 * & tor99 & & & & & 7.69 * & tor99 & 7.06 * & tor99 & 6.59 * & tor99 & + gj 473 b & & & 15.11 * & tor99 & 13.24 * & tor99 & & & & & 7.82 * & tor99 & 7.26 * & tor99 & 7.03 * & tor99 & + gj 551 & 14.36 & leg92 & 12.88 & bes90 & 11.05 & bes90 & 9.43 & bes90 & 7.43 & bes90 & 5.357 & 0.023 & 4.835 & 0.057 & 4.384 & 0.033 & + gj 559 a & & & 0.64 & bes90 & 0.01 & bes90 & @xmath310.35 & bes90 & @xmath310.68 & bes90 & & & & & & & + gj 559 b & & & 2.18 & bes90 & 1.34 & bes90 & 0.87 & bes90 & 0.46 & bes90 & & & & & & & + gj 628 & 12.82 & wei93 & 11.68 & bes90 & 10.10 & bes90 & 8.94 & bes90 & 7.42 & bes90 & 5.950 & 0.024 & 5.373 & 0.040 & 5.075 & 0.024 & + gj 674 & & & 10.90 & bes90 & 9.37 & bes90 & 8.30 & bes90 & 6.97 & bes90 & 5.711 & 0.019 & 5.154 & 0.033 & 4.855 & 0.018 & + gj 687 & 11.76 & leg92 & 10.64 & wei96 & 9.17 & wei96 & 8.08 & wei96 & 6.68 & wei96 & 5.335 & 0.021 & 4.766 & 0.033 & 4.548 & 0.021 & + gj 699 & 12.54 & leg92 & 11.31 & bes90 & 9.57 & bes90 & 8.35 & bes90 & 6.79 & bes90 & 5.244 & 0.020 & 4.834 & 0.034 & 4.524 & 0.020 & + gj 725 a & 11.55 & leg92 & 10.42 & wei96 & 8.90 & wei96 & 7.83 & wei96 & 6.48 & wei96 & 5.189 & 0.017 & 4.741 & 0.036 & 4.432 & 0.020 & + gj 725 b & 12.41 & leg92 & 11.28 & wei96 & 9.69 & wei96 & 8.57 & wei96 & 7.13 & wei96 & 5.721 & 0.020 & 5.197 & 0.024 & 5.000 & 0.022 & + scr 1845 - 6357 ab & & & & & 17.39 & hen06 & 14.99 & hen06 & 12.46 & hen06 & 9.544 & 0.023 & 8.967 & 0.027 & 8.508 & 0.020 & joint + gj 729 & & & 12.18 & bes90 & 10.44 & bes90 & 9.21 & bes90 & 7.65 & bes90 & 6.222 & 0.018 & 5.655 & 0.034 & 5.370 & 0.016 & + gj 1245 a & & & 15.31 & leg92 & 13.46 * & hen99 & 11.81 & wei96 & 9.78 & wei96 & 8.09 * & hen93 & 7.53 * & hen93 & 7.21 * & hen93 & joint bri + gj 1245 c & & & & & 16.75 * & hen99 & & & & & 9.35 * & hen93 & 8.61 * & hen93 & 8.24 * & hen93 & + gj 1245 b & & & 15.98 & leg92 & 14.01 & wei96 & 12.36 & wei96 & 10.27 & wei96 & 8.275 & 0.025 & 7.728 & 0.031 & 7.387 & 0.018 & + gj 820 a & 8.63 & joh66 & 7.40 & joh66 & 6.03 & joh66 & 4.86 & joh66 & 4.03 & joh66 & 3.114 & 0.268 & 2.540 & 0.198 & 2.248 & 0.318 & + gj 820 b & 8.62 & leg92 & 7.40 & leg92 & 6.03 & leg92 & & & 4.41 & leg92 & 3.546 & 0.278 & 2.895 & 0.218 & 2.544 & 0.328 & + gj 825 & 9.29 & le192 & 8.09 & bes90 & 6.67 & bes90 & 5.77 & bes90 & 4.91 & bes90 & 3.915 & mou76 & 3.270 & mou76 & 3.075 & mou76 & + gj 832 & 11.36 & leg92 & 10.18 & bes90 & 8.66 & bes90 & 7.66 & bes90 & 6.47 & bes90 & 5.349 & 0.032 & 4.766 & 0.256 & 4.501 & 0.018 & + gj 845 a & 6.74 & joh66 & 5.73 & bes90 & 4.68 & bes90 & 4.06 & bes90 & 3.53 & bes90 & 2.894 & 0.292 & 2.349 & 0.214 & 2.237 & 0.240 & + gj 845 bc & & & & & & & & & & & 11.010 & 0.020 & 11.306 & 0.024 & 11.208 & 0.024 & joint + gj 860 a & & & & & 9.86 * & hen93 & & & & & 5.91 * & hen93 & 5.33 * & hen93 & 5.02 * & hen93 & + gj 860 b & & & & & 11.41 * & hen93 & & & & & 7.10 * & hen93 & 6.47 * & hen93 & 6.39 * & hen93 & + gj 866 ac & 15.83 & leg92 & 14.33 & bes90 & 12.94 * & hen99 & 10.70 & bes90 & 8.64 & bes90 & 7.06 * & lei90 & 6.46 * & lei90 & 6.05 * & lei90 & joint ubri + gj 866 b & & & & & 13.34 * & hen99 & & & & & 7.62 * & lei90 & 7.02 * & lei90 & 6.61 * & lei90 & + gj 876 & 12.90 & wei93 & 11.76 & bes90 & 10.18 & bes90 & 9.00 & bes90 & 7.43 & bes90 & 5.934 & 0.019 & 5.349 & 0.049 & 5.010 & 0.021 & + gj 887 & & & 8.84 & bes90 & 7.34 & bes90 & 6.37 & bes90 & 5.32 & bes90 & 4.338 & 0.258 & 3.608 & 0.230 & 3.465 & 0.200 & + gj 905 & 15.65 & leg92 & 14.20 & wei96 & 12.29 & wei96 & 10.77 & wei96 & 8.82 & wei96 & 6.884 & 0.025 & 6.247 & 0.027 & 5.929 & 0.020 & + gj 166 a & & & & & 4.43 & bes90 & 3.96 & bes90 & 3.54 & bes90 & 3.013 & 0.238 & 2.594 & 0.198 & 2.498 & 0.236 & + gj 166 b & 8.88 & kid91 & 9.56 & kid91 & 9.53 & kid91 & & & & & 9.849 & 0.029 & 9.986 & 0.039 & 9.861 & 0.071 & + gj 166 c & & & 12.84 & rei04 & 11.17 & rei04 & & & & & 6.747 & 0.020 & 6.278 & 0.040 & 5.962 & 0.026 & + [ tab : phot ] lrccccccc gj 17 & 5 & 00 20 04.2 & @xmath3164 52 29 & .11647 & .00016 & 2 & f9.5v & gra06 + gj 19 & 6 & 00 25 45.1 & @xmath3177 15 15 & .13407 & .00011 & 2 & g0.0v & gra06 + gj 33 & 121 & 00 48 23.0 & @xmath6005 16 50 & .13426 & .00049 & 2 & k2.5v & gra06 + gj 34 a & 123 & 00 49 06.3 & @xmath6057 48 55 & .16823 & .00046 & 2 & g3.0v & mon01 + gj 53 a & 8 & 01 08 16.4 & @xmath6054 55 13 & .13267 & .00074 & 2 & g5.0v & joh53 + gj 66 a & & 01 39 47.6 & @xmath3156 11 47 & .12999 & .00208 & 2 & k5.0v & mon01 + gj 66 b & & 01 39 47.6 & @xmath3156 11 36 & .12999 & .00208 & 2 & k5.0v & mon01 + gj 68 & 1287 & 01 42 29.8 & @xmath6020 16 07 & .13275 & .00049 & 2 & k1.0v & gra06 + gj 105 a & 15 & 02 36 04.9 & @xmath6006 53 13 & .13906 & .00044 & 2 & k3.0v & gra06 + gj 139 & 19 & 03 19 55.7 & @xmath3143 04 11 & .16547 & .00019 & 2 & g8.0v & pas94 + gj 150 & 1581 & 03 43 14.9 & @xmath3109 45 48 & .11063 & .00022 & 2 & k1.0iii - iv & gra06 + gj 178 & & 04 49 50.4 & @xmath6006 57 41 & .12393 & .00017 & 2 & f6.0v & mon01 + gj 183 & 200 & 05 00 49.0 & @xmath3105 45 13 & .11477 & .00048 & 2 & k3.0v & cns91 + gj 216 a & & 05 44 27.8 & @xmath3122 26 54 & .11204 & .00018 & 2 & f6.0v & mon01 + gj 216 b & & 05 44 26.5 & @xmath3122 25 19 & .11204 & .00018 & 2 & k2.0v & cns91 + gj 222 a & & 05 54 23.0 & @xmath6020 16 34 & .11522 & .00025 & 3 & g0.0v & cns91 + gj 250 a & 1875 & 06 52 18.1 & @xmath3105 10 25 & .11465 & .00043 & 2 & k3.0v & cns91 + gj 423 a & 2390 & 11 18 10.9 & @xmath6031 31 45 & .11951 & .00079 & 2 & g0.0v & bat89 + gj 423 b & 2391 & 11 18 11.0 & @xmath6031 31 46 & .11951 & .00079 & 2 & g5.0v & bat89 + gj 432 a & 308 & 11 34 29.5 & @xmath3132 49 53 & .10461 & .00037 & 2 & k0.0v & cns91 + gj 434 & & 11 41 03.0 & @xmath6034 12 06 & .10416 & .00026 & 2 & g8.0v & cns91 + gj 442 a & 311 & 11 46 31.1 & @xmath3140 30 01 & .10844 & .00022 & 2 & g2.0v & gra06 + gj 451 & 44 & 11 52 58.8 & @xmath6037 43 07 & .11013 & .00040 & 2 & g8.0v & joh53 + gj 475 & 2579 & 12 33 44.5 & @xmath6041 21 27 & .11848 & .00020 & 2 & g0.0v & cns91 + gj 502 & 348 & 13 11 52.4 & @xmath6027 52 41 & .10952 & .00017 & 2 & g0.0v & cns91 + gj 506 & 349 & 13 18 24.3 & @xmath3118 18 40 & .11690 & .00022 & 2 & g7.0v & gra06 + gj 566 a & & 14 51 23.4 & @xmath6019 06 02 & .14757 & .00072 & 2 & g8.0v & ruc95 + gj 566 b & & 14 51 23.4 & @xmath6019 06 02 & .14757 & .00072 & 2 & k4.0v & mon99 + gj 570 a & 387 & 14 57 28.0 & @xmath3121 24 56 & .17062 & .00067 & 3 & k4.0v & gra06 + gj 631 & 3224 & 16 36 21.4 & @xmath3102 19 29 & .10249 & .00040 & 2 & k2.0v & mon01 + gj 638 & & 16 45 06.4 & @xmath6033 30 33 & .10195 & .00070 & 2 & k7.0v & cns91 + gj 663 a & 437 & 17 15 20.9 & @xmath3126 36 09 & .16812 & .00040 & 4 & k1.0v & cns91 + gj 663 b & 438 & 17 15 21.0 & @xmath3126 36 10 & .16812 & .00040 & 4 & k1.0v & cns91 + gj 664(c ) * & 439 & 17 16 13.4 & @xmath3126 32 46 & .16812 & .00040 & 4 & k5.0v & cns91 + gj 667 a & 442 & 17 18 57.2 & @xmath3134 59 23 & .13822 & .00070 & 2 & k3.0v & cns91 + gj 667 b & 443 & 17 19 01.9 & @xmath3134 59 33 & .13822 & .00070 & 2 & k5.0v & cns91 + gj 666 a & 444 & 17 19 03.8 & @xmath3146 38 10 & .11371 & .00069 & 2 & g8.0v & cns91 + gj 673 & 447 & 17 25 45.2 & @xmath6002 06 41 & .12987 & .00071 & 2 & k7.0v & cns91 + gj 695 a & 3326 & 17 46 27.5 & @xmath6027 43 14 & .12032 & .00016 & 2 & g5.0iv & cns91 + gj 702 a & 458 & 18 05 27.4 & @xmath6002 29 59 & .19596 & .00087 & 2 & k0.0v & cns91 + gj 702 b & 459 & 18 05 27.4 & @xmath6002 29 56 & .19596 & .00087 & 2 & k5.0v & cns91 + gj 713 a & 3379 & 18 21 03.4 & @xmath6072 43 58 & .12343 & .00044 & 3 & f7.0v & cns91 + gj 713 b & & 18 21 03.4 & @xmath6072 43 58 & .12343 & .00044 & 3 & g8.0v & far10 + gj 721 & & 18 36 56.3 & @xmath6038 47 01 & .12985 & .00032 & 3 & a0.0v & cns91 + gj 764 & 447 & 19 32 21.6 & @xmath6069 39 40 & .17379 & .00018 & 2 & k0.0v & cns91 + gj 768 & 3490 & 19 50 47.0 & @xmath6008 52 06 & .19540 & .00046 & 3 & a7.0v & cns91 + gj 780 & 485 & 20 08 43.6 & @xmath3166 10 55 & .16371 & .00017 & 2 & g8.0iv & gra06 + gj 783 a & 486 & 20 11 11.9 & @xmath3136 06 04 & .16626 & .00027 & 2 & k2.5v & gra06 + gj 785 & 488 & 20 15 17.4 & @xmath3127 01 59 & .11222 & .00030 & 2 & k2.0v & gra06 + gj 827 & 3674 & 21 26 26.6 & @xmath3165 21 58 & .10797 & .00019 & 2 & f9.0v & gra06 + gj 881(a ) & & 22 57 39.0 & @xmath3129 37 20 & .13042 & .00037 & 4 & a4.0v & gra06 + gj 879(b ) * & & 22 56 24.1 & @xmath3131 33 56 & .13042 & .00037 & 4 & k5.0v & cns91 + gj 884 & 3885 & 23 00 16.1 & @xmath3122 31 28 & .12175 & .00069 & 2 & k7.0v & gra06 + gj 892 & 71 & 23 13 17.0 & @xmath6057 10 06 & .15284 & .00028 & 2 & k3.0v & cns91 + [ tab:10pca ] llcccccccccccccccc gj 17 & 4.82 & joh66 & 4.80 & bes90 & 4.22 & bes90 & 3.89 & bes90 & 3.57 & bes90 & 3.17 & gla75 & 2.87 & gla75 & 2.78 & gla75 & + gj 19 & 3.53 & joh66 & 3.42 & bes90 & 2.80 & bes90 & 2.45 & bes90 & 2.12 & bes90 & 1.72 & gla75 & 1.40 & gla75 & 1.32 & gla75 & + gj 33 & & & 6.60 & bes90 & 5.72 & bes90 & 5.21 & bes90 & 4.76 & bes90 & 4.24 & joh68 & 3.72 & joh68 & 3.48 & joh68 & + gj 34 a & 4.04 & joh66 & 4.02 & joh66 & 3.44 & joh66 & 3.1 & usnob & 2.8 & usnob & 2.109 & 0.570 & 2.086 & 0.504 & 1.988 & & + gj 53 ab & 5.96 & joh66 & 5.87 & joh66 & 5.18 & joh66 & 4.7 & usnob & 4.4 & usnob & 3.86 & joh68 & 3.39 & joh68 & 3.36 & joh68 & joint + gj 66 a & & & 6.69 & hog00 & 5.80 & hog00 & & & & & 4.043 & 0.378 & & & 3.510 & 0.282 & + gj 66 b & & & 6.84 & mer86 & 5.96 & mer86 & & & & & 3.573 & & & & 3.558 & 0.270 & + gj 68 & 6.57 & joh66 & 6.08 & joh66 & 5.24 & joh66 & 4.7 & usnob & 4.3 & usnob & 3.855 & 0.24 & 3.391 & 0.226 & 3.285 & 0.266 & + gj 105 ac & & & 6.78 & bes90 & 5.81 & bes90 & 5.24 & bes90 & 4.74 & bes90 & 4.07 & gla75 & 3.52 & gla75 & 3.45 & gla75 & joint + gj 139 & 5.19 & joh66 & 4.97 & bes90 & 4.26 & bes90 & 3.85 & bes90 & 3.47 & bes90 & 2.95 & gla75 & 2.59 & gla75 & 2.52 & gla75 & + gj 150 & & & 4.45 & bes90 & 3.53 & bes90 & 3.03 & bes90 & 2.59 & bes90 & 1.99 & gla75 & 1.53 & gla75 & 1.45 & gla75 & + gj 178 & 3.64 & joh66 & 3.65 & bes90 & 3.19 & bes90 & 2.92 & bes90 & 2.67 & bes90 & 2.35 & gla75 & 2.15 & gla75 & 2.07 & gla75 & + gj 183 & & & 7.29 & bes90 & 6.23 & bes90 & 5.44 & bes90 & 4.69 & bes90 & 4.389 & 0.244 & 3.797 & 0.214 & 3.706 & 0.228 & + gj 216 a & 4.07 & joh66 & 4.06 & bes90 & 3.59 & bes90 & 3.30 & bes90 & 3.02 & bes90 & 2.804 & 0.276 & 2.606 & 0.236 & 2.508 & 0.228 & + gj 216 b & & & 7.12 & bes90 & 6.18 & bes90 & 5.63 & bes90 & 5.17 & bes90 & 4.845 & 0.198 & 4.158 & 0.202 & 4.131 & 0.264 & + gj 222 ab & 5.04 & joh66 & 5.00 & joh66 & 4.41 & joh66 & 4.0 & usnob & 3.8 & usnob & 3.34 & joh68 & 3.04 & joh68 & 2.97 & joh68 & joint + gj 250 a & & & 7.64 & bes90 & 6.59 & bes90 & 5.98 & bes90 & 5.45 & bes90 & 5.013 & 0.252 & 4.294 & 0.258 & 4.107 & 0.036 & + gj 423 ac & & & 4.78 & lep05 & 4.27 & lep05 & 3.9 & usnob & 3.7 & usnob & 2.462 & 0.294 & 2.231 & 0.204 & 2.142 & 0.230 & joint + gj 423 bd & & & 5.36 & lep05 & 4.74 & lep05 & 4.4 & usnob & 4.0 & usnob & & & & & & & joint + gj 432 a & & & 6.78 & bes90 & 5.97 & bes90 & 5.52 & bes90 & 5.10 & bes90 & 4.784 & 0.228 & 4.138 & 0.214 & 4.022 & 0.036 & + gj 434 & 6.28 & joh66 & 6.05 & bes90 & 5.33 & bes90 & 4.9 & usnob & 4.5 & usnob & 3.99 & joh68 & 3.61 & joh68 & 3.60 & joh68 & + gj 442 a & & & 5.56 & bes90 & 4.90 & bes90 & 4.5 & usnob & 4.2 & usnob & 3.931 & 0.276 & 3.490 & 0.238 & 3.489 & 0.278 & + gj 451 & 7.37 & joh66 & 7.20 & joh66 & 6.45 & joh66 & 6.0 & usnob & 5.6 & usnob & 4.89 & gla75 & 4.43 & gla75 & 4.37 & gla75 & + gj 475 & 4.91 & joh66 & 4.86 & joh66 & 4.27 & joh66 & 3.9 & usnob & 3.6 & usnob & 3.23 & joh66 & 2.905 & 0.198 & 2.84 & joh66 & + gj 502 & 4.92 & joh66 & 4.84 & joh66 & 4.26 & joh66 & 3.9 & usnob & 3.6 & usnob & 3.24 & gla75 & 2.90 & gla75 & 2.87 & gla75 & + gj 506 & 5.69 & joh66 & 5.43 & bes90 & 4.72 & bes90 & 4.33 & bes90 & 3.97 & bes90 & 3.334 & 0.200 & 2.974 & 0.176 & 2.956 & 0.236 & + gj 566 a & 5.68 & lut71 & 5.44 & lut71 & 4.72 & lut71 & 4.52 & bre64 & 4.24 & bre64 & 2.660 & 0.448 & 2.253 & 0.698 & 1.971 & 0.600 & joint jhk + gj 566 b & 9.29 & lut71 & 8.14 & lut71 & 6.97 & lut71 & 6.30 & bre64 & 5.86 & bre64 & & & & & & & + gj 570 a & 7.88 & joh66 & 6.82 & joh66 & 5.71 & joh66 & & & & & 3.663 & 0.258 & 3.085 & 0.196 & 3.048 & 0.224 & + gj 631 & & & 6.57 & bes90 & 5.76 & bes90 & 5.31 & bes90 & 4.9 & bes90 & 4.33 & joh66 & 4.053 & 0.208 & 3.87 & joh66 & + gj 638 & & & 9.48 & joh53 & 8.11 & joh65 & 7.3 & usnob & 6.6 & usnob & 5.48 & 0.023 & 4.878 & 0.018 & 4.712 & 0.021 & + gj 663 ab & & & 5.93 & tor06 & 5.08 & tor06 & & & & & & & & & & & joint + gj 664(c ) * & & & 7.48 & bes90 & 6.32 & bes90 & 5.62 & bes90 & 5.04 & bes90 & 4.155 & 0.25 & & & 3.466 & 0.256 & + gj 667 ab & 7.77 & joh66 & 6.95 & joh66 & 5.91 & joh66 & 4.97 & joh66 & 4.38 & joh66 & 3.903 & 0.262 & 3.230 & 0.206 & 3.123 & 0.278 & joint + gj 666 a & & & 6.35 & bes90 & 5.47 & bes90 & 5.00 & bes90 & 4.54 & bes90 & 4.077 & 0.996 & 3.146 & 0.664 & 3.421 & 0.282 & + gj 673 & & & 8.89 & bes90 & 7.53 & bes90 & 6.69 & bes90 & 5.94 & bes90 & 4.934 & 0.024 & 4.341 & 0.044 & 4.14 & gla75 & + gj 695 ad & 4.56 & joh66 & 4.17 & joh66 & 3.42 & joh66 & 2.9 & usnob & 2.6 & usnob & 2.13 & joh66 & 1.559 & 0.184 & 1.77 & joh66 & joint + gj 702 a & 5.31 & egg65 & 4.98 & egg65 & 4.20 & egg65 & 3.87 & bre64 & 3.61 & bre64 & 2.343 & 0.296 & 1.876 & 0.244 & 1.791 & 0.304 & joint jhk + gj 702 b & & & 7.15 & egg65 & 6.00 & egg65 & 5.26 & bre64 & 4.82 & bre64 & & & & & & & + gj 713 ab & 4.01 & joh66 & 4.07 & joh66 & 3.58 & joh66 & 3.3 & usnob & 3.0 & usnob & 2.588 & 0.260 & 2.372 & 0.188 & 2.216 & 0.252 &
joint + gj 721 & 0.02 & joh66 & 0.02 & bes90 & 0.03 & bes90 & 0.04 & bes90 & 0.04 & bes90 & 0.02 & joh66 & @xmath310.029 & 0.146 & 0.02 & joh66 & + gj 764 & 5.86 & oja84 & 5.46 & oja93 & 4.68 & oja93 & 4.2 & usnob & 3.8 & usnob & 3.32 & joh66 & 3.039 & 0.214 & 2.78 & joh66 & + gj 768 & 1.07 & joh66 & 0.99 & bes90 & 0.77 & bes90 & 0.64 & bes90 & 0.50 & bes90 & 0.39 & joh66 & 0.102 & 0.220 & 0.26 & joh66 & + gj 780 & 4.76 & joh66 & 4.31 & bes90 & 3.55 & bes90 & 3.14 & bes90 & 2.79 & bes90 & 2.35 & gla75 & 2.03 & gla75 & 1.93 & gla75 & + gj 783 a & & & 6.19 & joh66 & 5.32 & joh66 & & & & & 3.518 & 0.300 & 2.999 & 0.422 & 3.008 & 0.602 & + gj 785 & & & 6.61 & bes90 & 5.73 & bes90 & 5.23 & bes90 & 4.81 & bes90 & 4.112 & 0.294 & 3.582 & 0.266 & 3.501 & 0.232 & + gj 827 & 4.58 & joh66 & 4.71 & bes90 & 4.22 & bes90 & 3.92 & bes90 & 3.61 & bes90 & 3.27 & gla75 & 3.00 & gla75 & 2.90 & gla75 & + gj 881(a ) & 1.30 & joh66 & 1.24 & bes90 & 1.15 & bes90 & 1.10 & bes90 & 1.07 & bes90 & 1.02 & gla75 & 1.03 & gla75 & 0.99 & gla75 & + gj 879(b ) * & & & 7.56 & bes90 & 6.46 & bes90 & 5.8 & bes90 & 5.78 & bes90 & 4.533 & 0.037 & 3.804 & 0.210 & 3.805 & 0.240 & + gj 884 & & & 9.25 & bes90 & 7.86 & bes90 & 7.00 & bes90 & 6.23 & bes90 & 5.346 & 0.021 & 4.696 & 0.076 & 4.478 & 0.016 & + gj 892 & 7.46 & oja84 & 6.56 & oja93 & 5.57 & oja93 & 4.9 & usnob & 4.5 & usnob & 3.80 & gla75 & 3.27 & gla75 & 3.18 & gla75 & + [ tab:10pcphot ] .
lrrc u & 4.18e-9 & 3660 & bessell et al.(1998 ) + b & 6.32e-9 & 4380 & bessell et al.(1998 ) + v & 3.63e-9 & 5450 & bessell et al.(1998 ) + r & 2.18e-9 & 6410 & bessell et al.(1998 ) + i & 1.13e-9 & 7980 & bessell et al.(1998 ) + j & 3.14e-10 & 12350 & cohen et al.(2003 ) + h & 1.11e-10 & 16620 & cohen et al.(2003 ) + k & 4.29e-11 & 21590 & cohen et al.(2003 ) + [ tab : zp ] llrcccclc gj 244 a & a1.0v & @xmath310.040 & 10000 & 9600 & 9900@xmath45200 & 1.645 & 1.711@xmath45 0.013 & kervella et al ( 2003a ) + gj 280 a & f5.0iv - v & 1.028 & 6700 & 6500 & 6524 & 1.898 & 2.048@xmath45 0.025 & kervella
et al ( 2003b ) + gj 559 a & g2.0v & 1.490 & 6000 & 5700 & 5810@xmath4550 & 1.103 & 1.224@xmath45 0.003 & kervella et al ( 2003b ) + gj 71 & g8.5v & 1.696 & 5700 & 5400 & 5264@xmath45100 & 0.708 & 0.816@xmath45 0.013 & di folco et al.(2004 ) + gj 631 & k2.0v & 1.890 & 5500 & 5300 & 5337@xmath4541 & 0.710 & 0.759@xmath45 0.012 & boyajian et al.(2012 ) + gj 166 a & k0.5v & 1.932 & 5200 & 5100 & 5143@xmath4514 & 0.735 & 0.806@xmath45 0.004 & boyajian et al.(2012 ) + gj 559 b & k0.0v & 1.940 & 5400 & 5100 & 5260@xmath4550 & 0.775 & 0.863@xmath45 0.005 & kervella et al ( 2003b ) + gj 144 & k2.0v & 1.954 & 5200 & 5000 & 5135@xmath45100 & 0.689 & 0.743@xmath45 0.005 & di folco et al.(2004 ) + gj 33 & k2.5v & 2.240 & 5200 & 5000 & 4950@xmath4514 & 0.643 & 0.695@xmath45 0.004 & boyajian et al.(2012 ) + gj 105 a & k3.0v & 2.360 & 5000 & 4700 & 4662@xmath4517 & 0.677 & 0.795@xmath45 0.006 & boyajian et al.(2012 ) + gj 892 & k3.0v & 2.390 & 5000 & 4800 & 4699@xmath4516 & 0.698 & 0.778@xmath45 0.005 & boyajian et al.(2012 ) + gj 702 a & k0.0v & 2.409 & 5500 & 5700 & 5407@xmath4552 & 0.754 & 0.831@xmath45 0.004 & boyajian et al.(2012 ) + gj 845 a & k5.0v & 2.443 & 4900 & 4600 & 5468@xmath4559 & 0.608 & 0.732@xmath45 0.007 & demory et al.(2009 ) + gj 570 a & k4.0v & 2.662 & 4700 & 4700 & 4507@xmath4558 & 0.734 & 0.739@xmath45 0.019 & boyajian et al.(2012 ) + gj 820
b & k7.0v & 3.486 & 4300 & 4200 & 4040@xmath4580 & 0.530 & 0.595@xmath45 0.008 & kervella et al ( 2008 ) + gj 380 & k7.0v & 3.628 & 4200 & 4200 & 4081@xmath4515 & 0.599 & 0.642@xmath45 0.005 & boyajian et al.(2012 ) + gj 820 a & k5.0v & 3.782 & 4000 & 4100 & 4400@xmath45100 & 0.709 & 0.665@xmath45 0.005 & kervella et al ( 2008 ) + gj 702 b * & k5.0v & & & & 4393@xmath45149 & & 0.670@xmath45 0.009 & boyajian et al.(2012 ) + gj 191 & m2.0vi & 3.801 & 3800 & 3600 & 3570@xmath45156 & 0.249 & 0.291@xmath45 0.025 & sgransan et al.(2003 ) + gj 887 & m2.0v & 3.875 & 3800 & 3700 & 3797@xmath4545 & 0.414 & 0.459@xmath45 0.011 & demory et al.(2009 ) + gj 411 & m2.0v & 3.969 & 3700 & 3500 & 3465@xmath4517 & 0.338 & 0.392@xmath45 0.004 & boyajian et al.(2012 ) + gj 412 a & m1.0v & 4.001 & 3700 & 3600 & 3497@xmath4539 & 0.353 & 0.398@xmath45 0.009 & boyajian et al.(2012 ) + gj 15 a & m1.5v & 4.062 & 3800 & 3700 & 3730@xmath4549 & 0.323 & 0.379@xmath45 0.006 & berger et al .
( 2006 ) + gj 725 a & m3.0v & 4.468 & 3400 & 3400 & 3407@xmath4515 & 0.344 & 0.356@xmath45 0.004 & boyajian et al.(2012 ) + gj 687 & m3.0v & 4.622 & 3400 & 3200 & 3413@xmath4528 & 0.406 & 0.418@xmath45 0.007 & boyajian et al.(2012 ) + gj 725 b & m3.5v & 4.690 & 3300 & 3200 & 3104@xmath4528 & 0.275 & 0.323@xmath45 0.006 & boyajian et al.(2012 ) + gj 699 & m4.0v & 5.046 & 3100 & 3100 & 3224@xmath4510 & 0.198 & 0.187@xmath45 0.001 & boyajian et al.(2012 ) + gj 551 & m5.0v & 6.666 & 2700 & 2900 & 3098@xmath4556 & 0.167 & 0.141@xmath45 0.007 & demory et al.(2009 ) + [ tab : temp ] lllccccccccc sun & g2.0v & 0.917 & 6000 & 0.789 & 1.686 & 0.897 & 1.00 & 5800 & 0.804 & 1.718 & 0.914 + den
1048 - 3956 & m8.5v & 0.134 & 2000 & 0.013 & 0.027 & 0.014 & & & & & + gj 1 & m1.5v & 0.349 & 3700 & 0.114 & 0.244 & 0.130 & & & & & + gj 15 a & m1.5v & 0.323 & 3800 & 0.112 & 0.239 & 0.127 & 0.387 & 3700 & 0.127 & 0.271 & 0.144 + gj 15 b & m3.5v & 0.197 & 3200 & 0.048 & 0.103 & 0.055 & & & & & + gj 54.1 & m4.5v & 0.181 & 2900 & 0.036 & 0.078 & 0.042 & & & & & + gj 65 a & m5.5v & 0.248 & 2600 & 0.040 & 0.086 & 0.046 & & & & & + gj 65 b & m6.0v & 0.225 & 2700 & 0.039 & 0.084 & 0.045 & & & & & + gj 71 & g8.0v & 0.700 & 5700 & 0.545 & 1.164 & 0.619 & 0.816 & 5400 & 0.570 & 1.218 & 0.648 + gj 83.1 & m4.5v & 0.195 & 2900 & 0.039 & 0.084 & 0.045 & & & & & + gj 144 & k2.0v & 0.630 & 5400 & 0.440 & 0.940 & 0.500 & 0.735 & 5000 & 0.440 & 0.941 & 0.500 + gj 166 a & k0.5v & 0.735 & 5200 & 0.476 & 1.017 & 0.541 & 0.806 & 5000 & 0.483 & 1.031 & 0.549 +
gj 166 c & m4.5v & 0.249 & 3200 & 0.061 & 0.130 & 0.069 & & & & & + gj 191 & m1.5vi & 0.240 & 3800 & 0.083 & 0.177 & 0.094 & 0.291 & 3500 & 0.085 & 0.182 & 0.097 + gj 234 a & m4.0v & 0.287 & 3000 & 0.062 & 0.132 & 0.070 & & & & &
+ gj 234 b & m5.5v & 0.155 & 2700 & 0.027 & 0.058 & 0.031 & & & & & + gj 244 a * & a1.0v & 1.645 & 10000 & 3.942 & 8.420 & 4.478 & 1.711 & 9600 & 3.779 & 8.071 & 4.293 + gj 273 & m3.5v & 0.316 & 3200 & 0.077 & 0.165 & 0.088 & & & & & + gj 280 a * & f5.0iv - v & 1.906 & 6700 & 2.050 & 4.380 & 2.329 & 2.048 & 6500 & 2.073 & 4.429 & 2.356 + gj 380 & k7.0v & 0.648 & 4100 & 0.261 & 0.558 & 0.297 & 0.642 & 4100 & 0.259 & 0.552 & 0.294 + gj 388 & m3.0v & 0.452 & 3300 & 0.118 & 0.251 & 0.133 & & & & & + gj 406 & m6.0v & 0.162 & 2500 & 0.024 & 0.052 & 0.028 & & & & & + gj 411 & m2.0v & 0.338 & 3700 & 0.111 & 0.237 & 0.126 & 0.392 & 3500 & 0.115 & 0.246 & 0.130 + gj 412 a & m1.0v & 0.353 & 3700 & 0.116 & 0.247 & 0.131 & 0.398 & 3600 & 0.124 & 0.264 & 0.140 + gj 412 b & m5.5v & 0.152 & 2600 & 0.025 & 0.053 & 0.028 & & & & & + gj 447 & m4.0v & 0.213 & 3000 & 0.046 & 0.098 & 0.052 & & & & & + gj 473 a & m5.0v & 0.212 & 2700 & 0.037 & 0.079 & 0.042 & & & & & + gj 473 b & m7.0v & 0.177 & 2900 & 0.036 & 0.076 & 0.040 & & & & & + gj 551 & m5.5v & 0.167 & 2700 & 0.029 & 0.062 & 0.033 & 0.141 & 2900 & 0.028 & 0.061 & 0.032 + gj 559 a & g2.0v & 0.884 & 6700 & 0.951 & 2.031 & 1.080 & 1.224 & 5700 & 0.953 & 2.036 & 1.083 +
gj 559 b & k0.0v & 0.774 & 5400 & 0.541 & 1.155 & 0.614 & 0.863 & 5200 & 0.559 & 1.194 & 0.635 + gj 628 & m3.0v & 0.321 & 3200 & 0.079 & 0.168 & 0.089 & & & & & + gj 674 & m2.5v & 0.355 & 3400 & 0.098 & 0.210 & 0.112 & & & & & + gj 687 & m3.0v & 0.406 & 3400 & 0.112 & 0.240 & 0.128 & 0.418 & 3400 & 0.116 & 0.247 & 0.132 + gj 699 & m4.0v & 0.198 & 3100 & 0.046 & 0.097 & 0.051 & 0.187 & 3100 & 0.043 & 0.092 & 0.048 + gj 725 a & m3.0v & 0.344 & 3400 & 0.095 & 0.204 & 0.109 & 0.356 & 3400 & 0.099 & 0.211 & 0.112 +
gj 725 b & m3.5v & 0.275 & 3300 & 0.072 & 0.153 & 0.081 & 0.323 & 3200 & 0.079 & 0.169 & 0.090 + gj 729 & m3.5v & 0.214 & 3100 & 0.049 & 0.105 & 0.056 & & & & & + gj 820 a & k5.0v & 0.709 & 4000 & 0.272 & 0.581 & 0.309 & 0.665 & 4100 & 0.268 & 0.572 & 0.304 +
gj 820 b & k7.0v & 0.530 & 4300 & 0.235 & 0.502 & 0.267 & 0.595 & 4200 & 0.252 & 0.537 & 0.286 + gj 825 & m0.0v & 0.516 & 4100 & 0.208 & 0.444 & 0.236 & & & & & + gj 832 & m1.5v & 0.423 & 3600 & 0.131 & 0.280 & 0.149 & & & & & + gj 845 a & k5.0v & 0.608 & 4900 & 0.350 & 0.747 & 0.397 & 0.732 & 4600 & 0.371 & 0.793 & 0.422 + gj 860 a & m3.0v & 0.328 & 3300 & 0.085 & 0.183 & 0.098 & & & & & + gj 860 b & m4.0v & 0.194 & 3200 & 0.048 & 0.102 & 0.054 & & & & & + gj 866 b & mv & 0.215 & 2700 & 0.037 & 0.080 & 0.043 & & & & & + gj 876 a & m4.0v & 0.390 & 3100 & 0.090 & 0.191 & 0.101 & & & & & + gj 887 & m1.5v & 0.414 & 3800 & 0.143 & 0.306 & 0.163 & 0.491 & 3700 & 0.161 & 0.344 & 0.183 + gj 905 & m5.5v & 0.216 & 2700 & 0.038 & 0.080 & 0.042 & & & & & + gj 1002 & m5.0v & 0.142 & 2900 & 0.029 & 0.061 & 0.032 & & & & & + gj 1061 & m5.0v & 0.178 & 2700 & 0.031 & 0.066 & 0.035 & & & & & + gj 1111 & m6.0v & 0.162 & 2400 & 0.022 & 0.048 & 0.026 & & & & & + gj 1245 a & m5.5v & 0.194 & 2700 & 0.034 & 0.072 & 0.038 & & & & & + gj 1245 b & m6.0v & 0.150 & 2700 & 0.026 & 0.056 & 0.030 & & & & & + gj 1245 c & m7.0v & 0.171 & 2100 & 0.018 & 0.039 & 0.021 & & & & & + lhs 288 & m5.5v & 0.150 & 2700 & 0.026 & 0.056 & 0.030 & & & & & + lhs 292 & m6.5v & 0.146 & 2400 & 0.020 & 0.043 & 0.023 & & & & & + lp 944 - 020 & m9.0v & 0.091 & 2000 & 0.009 & 0.019 & 0.010 & & & & & + scr 1845 - 6357 a & m8.5v & 0.129 & 2000 & 0.012 & 0.026 & 0.014 & & & & & +
so 0253@xmath601652 & m7.0v & 0.156 & 2400 & 0.022 & 0.046 & 0.024 & & & & & + [ tab : hz ] lllccccccccc gj 17 & f9.5v & 0.989 & 6100 & 0.882 & 1.884 & 1.002 & & & & & + gj 19 & g0.0v & 1.709 & 6000 & 1.474 & 3.149 & 1.675 & & & & & + gj 33 & k2.5v & 0.643 & 5200 & 0.417 & 0.890 & 0.473 & 0.695 & 5000 & 0.416 & 0.889 & 0.473 + gj 34 a & g3.0v & 0.989 & 6100 & 0.882 & 1.884 & 1.002 & & & & & + gj 66 a & k5.0v & 0.740 & 5100 & 0.461 & 0.985 & 0.524 & & & & & + gj 66 b & k5.0v & 0.794 & 4900 & 0.457 & 0.976 & 0.519 & & & & & + gj 68 & k1.0v & 0.739 & 5400 & 0.516 & 1.103 & 0.587 & & & & & + gj 105 a & k3.0v & 0.677 & 5000 & 0.406 & 0.866 & 0.460 & 0.795 & 4700 & 0.421 & 0.899 & 0.478 + gj 139 & g8.0v & 0.812 & 5700 & 0.632 & 1.350 & 0.718 & & & & & + gj 178 & f6.0v & 1.244 & 6600 & 1.299 & 2.774 & 1.475 & & & & & + gj 183 & k3.0v & 0.791 & 4800 & 0.437 & 0.933 & 0.496 & & & & & + gj 216 a & f6.0v & 1.137 & 6600 & 1.187 & 2.534 & 1.347 & & & & & + gj 216 b & k2.0v & 0.593 & 5300 & 0.399 & 0.853 & 0.454 & & & & & + gj 250 a & k3.0v & 0.597 & 4900 & 0.344 & 0.734 & 0.390 & & & & & + gj 432 a & k0.0v & 0.627 & 5500 & 0.455 & 0.971 & 0.516 & & & & & + gj 434 & g8.0v & 0.793 & 5700 & 0.617 & 1.319 & 0.702 & & & & & + gj 442 a & g2.0v & 0.792 & 6000 & 0.683 & 1.459 & 0.776 & & & & & + gj 451 & g8.0v & 0.542 & 5400 & 0.378 & 0.807 & 0.429 & & & & & + gj 475 & g0.0v & 0.952 & 6100 & 0.849 & 1.813 & 0.964 & & & & & + gj 502 & g0.0v & 1.029 & 6100 & 0.918 & 1.960 & 1.042 & & & & & + gj 506 & g7.0v & 0.927 & 5700 & 0.722 & 1.542 & 0.820 & & & & & + gj 566 a & g8.0v & 0.534 & 6200 & 0.491 & 1.049 & 0.558 & & & & & + gj 566 b & k4.0v & 0.376 & 5000 & 0.225 & 0.481 & 0.256 & & & & & + gj 570 a & k4.0v & 0.734 & 4700 & 0.389 & 0.830 & 0.441 & 0.739 & 4700 & 0.391 & 0.836 & 0.444 + gj 631 & k2.0v & 0.710 & 5500 & 0.515 & 1.099 & 0.584 & 0.759 & 5300 & 0.511 & 1.091 &
0.580 + gj 638 & k7.0v & 0.620 & 4100 & 0.250 & 0.534 & 0.284 & & & & & + gj 663 a & k1.0v & & & 0.523 & 1.118 & 0.595 & & & & & + gj 663 b & k1.0v & & & 0.523 & 1.118 & 0.595 & & & & & + gj 664 ( c ) * & k5.0v & 0.581 & 4600 & 0.295 & 0.629 & 0.334 & & & & & + gj 666 a & g8.0v & 0.841 & 5200 & 0.545 & 1.164 & 0.619 & & & & & + gj 667 a & k3.0v & & & 0.399 & 0.853 & 0.454 & & & & & + gj 667 b & k5.0v & & & 0.285 & 0.607 & 0.322 & & & & & + gj 673 & k7.0v & 0.644 & 4100 & 0.259 & 0.554 & 0.295 & & & & & + gj 702 a & k0.0v & 0.754 & 5500 & 0.547 & 1.168 & 0.621 & 0.670 & 5700 & 0.522 & 1.114 & 0.593 + gj 702 b & k5.0v & 0.411 & 5100 & 0.256 & 0.735 & 0.479 & & & & & + gj 721 & a0.0v & 2.543 & 9800 & 5.852 & 12.50 & 6.648 & & & & & + gj 764 & k0.0v & 0.694 & 5500 & 0.503 & 1.075 & 0.572 & & & & & + gj 768 & a7.0v & 1.701 & 7800 & 2.480 & 5.297 & 2.817 & & & & & + gj 783 a & k2.5v & 0.751 & 5000 & 0.450 & 0.961 & 0.511 & & & & & + gj 785 & k2.0v & 0.802 & 5100 & 0.450 & 1.068 & 0.618 & & & & & + gj 827 & f9.0v & 0.970 & 6400 & 0.952 & 2.034 & 1.082 & & & & & + gj 879 ( b ) * & k5.0v & 0.592 & 4700 & 0.313 & 0.669 & 0.356 & & & & & + gj 881 ( a ) & a4.0v & 1.635 & 9200 & 3.316 & 7.084 & 3.768 & & & & & + gj 884 & k7.0v & 0.553 & 4200 & 0.234 & 0.499 & 0.265 & & & & & + gj 892 & k3.0v & 0.698 & 5000 & 0.418 & 0.893 & 0.475 & 0.778 & 4800 & 0.430 & 0.918 & 0.488 + [ tab:10pchz ] lrrrrrrrrc gj 15 ab & 40@xmath61 & & & & & & & & eggen ( 1996 ) + gj 34 ab & & 11@xmath5099 & 0.497 & 34.76 & 98.43 & 88.59 & 480 & 1889.6 & strand ( 1969 ) + gj 53 ab & & 1@xmath5001 & 0.561 & 106.8 & 47.3 & 152.7 & 21.75 & 1975.74 & drummond et al .
( 1995 ) + gj 65 ab & & 2@xmath5006 & 0.615 & 127.3 & 150.5 & 285.4 & 26.52 & 1971.88 & worley & behall ( 1973 ) + gj 66 ab & & 7@xmath5082 & 0.534 & 142.8 & 13.1 & 18.37 & 483.66 & 1813.5 & van albada ( 1957 ) + gj 105 ac - b & 165@xmath61 & & & & & & & & golimowski et al . ( 2000 ) + gj 105 ac & 3@xmath503 & & & & & & & & golimowski et al .
( 2000 ) + gj 166 a - bc & 83@xmath61 & & & & & & & & baize & petit ( 1989 ) + gj 166 bc & 6@xmath5094 & & 0.410 & 108.9 & 150.9 & 327.8 & 252.1 & 1849.6 & heintz ( 1974 ) + gj 216 ab & 95@xmath61 & & & & & & & & eggen ( 1956 ) + gj 222 ab & & 0@xmath50688 & 0.451 & 95.94 & 126.36 & 111.57 & 14.11 & 1999.9 & han & gatewood ( 2002 ) + gj 234 ab & & 1@xmath5004 & 0.371 & 51.8 & 30.7 & 223 & 16.12 & 1999.38 & sgransan et al .
( 2000 ) + gj 244 ab & & 7@xmath5056 & 0.592 & 136.5 & 55.57 & 147.27 & 50.09 & 1894.13 & gatewood & gatewood ( 1978 ) + gj 250 ab & 58@xmath61 & & & & & & & & eggen ( 1956 ) + gj 280 ab & & 4@xmath50496 & 0.365 & 31.9 & 284.8 & 88.8 & 40.38 & 1967.86 & irwin et al .
( 1992 ) + gj 412 ab & 28@xmath61 & & & & & & & & gould ( 2003 ) + gj 423 ac - bd & & 2@xmath50533 & 0.421 & 112.1 & 101.3 & 127.3 & 59.84 & 1995.05 & heintz ( 1996 ) + gj 423 ac & & 0@xmath50056 & 0.53 & 94.9 & 263.5 & 143.0 & 1.832 & 1986.50 & mason et al .
( 1995 ) + gj 423 bd & * 274,000 km & & & & & & 0.015 & & mason et al .
( 1995 ) + gj 432 ab & 17@xmath61 & & & & & & & & henry et al .
( 2002 ) + gj 442 ab & 25@xmath504 & & & & & & & & poveda et al .
( 1994 ) + gj 473 ab & & 0@xmath50926 & 0.295 & 103.00 & 143.48 & 347.2 & 15.64 & 1992.30 & torres et al .
( 1999 ) + gj 559 ab & & 17@xmath5057 & 0.518 & 79.2 & 204.85 & 231.65 & 79.91 & 1875.66 & pourbaix et al .
( 2002 ) + gj 551-gj 559 ab & 2@xmath6218 & & & & & & & & poveda et al .
( 1994 ) + gj 566 ab & & 4@xmath5094 & 0.51 & 139 & 347 & 203 & 151.6 & 1909.3 & sderhjelm ( 1999 ) + gj 570 a - bc & & 32@xmath5034 & 0.20 & 72.53 & 317.31 & 252.1 & 2130 & 1689 & hale ( 1994 ) + gj 570 bc & & 0@xmath50151 & 0.756 & 107.6 & 195.9 & 127.56 & 0.846 & 1996.51 & forveille et al .
( 1999 ) + gj 570 abc - d & 258@xmath503 & & & & & & & & burgasser et al .
( 2000 ) + gj 663 ab & & 13@xmath500 & 0.916 & 99.79 & -85.8 & -90.2 & 470.9 & 1677.9 & irwin et al .
( 1996 ) + gj 663 ab - gj 664 & 74@xmath502 & & & & & & & & luyten ( 1957 ) + gj 666 ab & & 10@xmath5042 & 0.779 & 35.64 & 131.78 & 333.44 & 693.24 & 1907.2 & wieth - knudsen ( 1957 ) + gj 667 ab & & 1@xmath5081 & 0.58 & 128 & 313 & 247 & 42.15 & 1975.9 & sderhjelm ( 1999 ) + gj 695 ad & 1@xmath5043 & & 0.32 & 68.0 & 81.8 & 92 & 65 & 1951.0 & turner et al .
( 2001 ) ; heintz ( 1994 ) + gj 695 ad - bc & 34@xmath61 & & & & & & & & raghavan et al .
( 2010 ) + gj 695 bc & & 1@xmath5036 & 0.178 & 66.2 & 60.7 & 174.0 & 43.2 & 1965.4 & couteau ( 1959 ) + gj 702 ab & & 4@xmath5055 & 0.499 & 121.16 & 302.12 & 14.0 & 88.38 & 1895.94 & pourbaix ( 2000 ) + gj 713 ab & & 0@xmath50124 & 0.428 & 74.42 & 230.30 & 119.3 & 0.768 & 1984.83 & farrington et al .
( 2010 ) + gj 725 ab & & 13@xmath5088 & 0.53 & 66.0 & 136.9 & 234.6 & 408 & 1775.0 & heintz ( 1987 ) + gj 783 ab & 7@xmath501 & & & & & & & & poveda et al .
( 1994 ) + gj 820 ab & & 24@xmath5043 & 0.414 & 52.7 & 173.4 & 153.17 & 691.61 & 1689.14 & baize ( 1950 ) + gj 845
a - bc & 402@xmath503 & & & & & & & & scholz et al .
( 2003 ) + gj 845 bc & 0@xmath50732 & & & & & & & & mccaughrean et al . ( 2004 ) + gj 860 ab & & 2@xmath50383 & 0.410 & 167.2 & 154.5 & 211.0 & 44.67 & 1970.22 & heintz ( 1986 ) + gj 866 ac & & @xmath200@xmath5001 & 0.000 & @xmath20117 & & & & 1991.71 & delfosse et al .
( 1999 ) + gj 866 ac - b & & 0@xmath50346 & 0.446 & 112 & 161.5 & 337.6 & 2.25 & 1997.53 & delfosse et al . ( 1999 ) + gj 881-gj 879 & 1@xmath6297 & & & & & & & & poveda et al .
( 1994 ) + gj 1245 ac - b & 7@xmath50969 & & 0.320 & 135.0 & 80.0 & 38.0 & 15.22 & 1983.1 & tokovinin ( 1997 ) + gj 1245 ac & 0@xmath50800 & & & & & & & & tokovinin ( 1997 ) + scr 1845 - 6357 ab & 1@xmath50170 & & & & & & & & biller et al .
( 2006 ) + [ tab : bin ] lcccccc gj 139 & b & [email protected] & 0.0 & pepe et al .
( 2011 ) & 0.632 & 1.350 + gj 139 & c & [email protected] & 0.0 & pepe et al .
( 2011 ) & 0.632 & 1.350 + gj 139 & d & [email protected] & 0.0 & pepe et al .
( 2011 ) & 0.632 & 1.350 + gj 144 & b & [email protected] & [email protected] & benedict et al .
( 2006 ) & 0.439 & 0.938 + gj 176 & b & 0.066 & 0 & forveille et al .
( 2009 ) & 0.136 & 0.290 + gj 442 a & b & [email protected] & [email protected] & tinney et al .
( 2011 ) & 0.689 & 1.472 + gj 506 & b & [email protected] & [email protected] & vogt et al .
( 2010a ) & 0.689 & 1.472 + gj 506 & c & [email protected] & [email protected] & vogt et al .
( 2010a ) & 0.719 & 1.537 + gj 506 & d & [email protected] & [email protected] & vogt et al .
( 2010a ) & 0.719 & 1.537 + gj 559 b & b * & [email protected] & 0 & dumusque et al .
( 2012 ) & 0.541 & 1.155 + gj 581 & b & 0.04 & 0 & mayor et al .
( 2009 ) & 0.083 & 0.179 + gj 581 & c & 0.07 & [email protected] & mayor et al .
( 2009 ) & 0.083 & 0.179 + * gj 581 * & * d * & * 0.22 * & * [email protected] * & * mayor et al .
( 2009 ) * & * 0.083 * & * 0.179 * + gj 581 & e & 0.03 & 0 & mayor et al .
( 2009 ) & 0.083 & 0.179 + gj 581 & f * & [email protected] & & vogt et al .
( 2010b ) & 0.083 & 0.179 + * gj 581 * & * g * * & * [email protected] * & & * vogt et al . ( 2010b ) * & * 0.083 * & * 0.179 * + gj 667 c & b & 0.049 & [email protected] & anglada - escud et al . ( 2012 ) & 0.096 & 0.205 + * gj 667 c * & * c * & * [email protected] * & * @xmath18 0.27 * & * anglada - escud et al .
( 2012 ) * & * 0.096 * & * 0.205 * + gj 674 & b & 0.039 & [email protected] & bonfils et al .
( 2007 ) & 0.098 & 0.210 + gj 785 & b & [email protected] & [email protected] & howard et al .
( 2010 ) & 0.500 & 1.068 + gj 832 & b & [email protected] & [email protected] & bailey et al .
( 2009 ) & 0.131 & 0.280 + gj 849 & b & 2.35 & [email protected] & butler et al .
( 2006 ) & 0.136 & 0.290 + gj 876 & b & [email protected] & [email protected] & rivera et al .
( 2010 ) & 0.090 & 0.191 + * gj 876 * & * c * & * [email protected] * & * [email protected] * & * rivera et al . ( 2010 ) * & * 0.090 * & * 0.191 * + gj 876 & d & [email protected] & [email protected] & rivera et al . (
2010 ) & 0.090 & 0.191 + gj 876 & e & [email protected] & [email protected] & rivera et al .
( 2010 ) & 0.090 & 0.191 + gj 881 & b & 115 & 0.11 & kalas et al .
( 2008 ) & 3.316 & 7.084 + [ tab : exop ]
|
we use the sample of known stars and brown dwarfs within 5 pc of the sun , supplemented with afgk stars within 10 pc , to determine which stellar spectral types provide the most habitable real estate defined to be locations where liquid water could be present on earth - like planets .
stellar temperatures and radii are determined by fitting model spectra to spatially resolved broad - band photometric energy distributions for stars in the sample . using these values ,
the locations of the habitable zones are calculated using an empirical formula for planetary surface temperature and assuming the condition of liquid water , called here the empirical habitable zone , or ehz .
systems that have dynamically disruptive companions , assuming a 5:1 separation ratios for primary / secondary pairs and either object and a planet , are considered not habitable .
we use the results of these calculations to derive a simple formula to predict the location of the ehz for main sequence stars based on @xmath0 color .
we consider ehz widths as more useful measures of the habitable real estate around stars than areas because multiple planets are not expected to orbit stars at identical stellar distances .
this ehz provides a qualitative guide on where to expect the largest population of planets in the habitable zone of main sequence stars . because of their large numbers and lower frequency of short - period companions , m stars provide more ehz real estate than other spectral types , possessing 36.5% of the habitable real estate _ en masse_. k stars are second with 21.5% , while a , f , and g stars offer 18.5% , 6.9% and 16.6% , respectively .
our calculations show that three m dwarfs within 10 pc harbor planets in their ehzs gj 581 may have two planets ( d with @xmath1sin@xmath2 = 6.1 @xmath3 ; g with @xmath1sin@xmath2 = 3.1 @xmath3 ) , gj 667 c has one ( c with @xmath1sin@xmath2 = 4.5 @xmath3 ) , and gj 876 has two ( b with @xmath1sin@xmath2 = 1.89 @xmath4 and c with @xmath1sin@xmath2 = 0.56 @xmath4 ) .
if earth - like planets are as common around low mass stars as recent kepler results suggest , m stars are the most likely place to find earth - like planets in habitable zones .
|
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.